Eng-1

Handbook on
Monitoring and
Evaluation of
Human Resources
for Health
with special applications for

low- and middle-income countries
Handbook on
Monitoring and Evaluation of
Human Resources for Health
Edited by Mario R Dal Poz, Neeru Gupta,

Estelle Quain and Agnes LB Soucat
WHO Library Cataloguing-in-Publication Data
Handbook on monitoring and evaluation of human resources for health: with special applications for low- and
middle-income countries / edited by Mario R Dal Poz … [et al].
1.Health manpower. 2.Health personnel – administration and organization. 3.Health personnel – statistics and
numerical data. 4.Personnel management. 5.Qualitative analysis. 6.Developing countries. I.Dal Poz, Mario R
II.Gupta, Neeru. III.Quain, Estelle IV.Soucat, Agnes LB V.World Health Organization. VI.World Bank. VII.United
States. Agency for International Development.
ISBN 978 92 4 154770 3 (NLM classification: W 76)
© World Health Organization 2009
All rights reserved. Publications of the World Health Organization can be obtained from WHO Press, World
Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: +41 22 791 3264; fax: +41 22 791
4857; e-mail: bookorders@who.int). Requests for permission to reproduce or translate WHO publications –
whether for sale or for noncommercial distribution – should be addressed to WHO Press, at the above address
(fax: +41 22 791 4806; e-mail: permissions@who.int).
The designations employed and the presentation of the material in this publication do not imply the expression
of any opinion whatsoever on the part of the World Health Organization, World Bank or United States Agency
for International Development concerning the legal status of any country, territory, city or area or of its authori-
ties, or concerning the delimitation of its frontiers or boundaries. Dotted lines on maps represent approximate
border lines for which there may not yet be full agreement.
The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed
or recommended by the World Health Organization, World Bank or United States Agency for International
Development in preference to others of a similar nature that are not mentioned. Errors and omissions excepted,
the names of proprietary products are distinguished by initial capital letters.
All reasonable precautions have been taken by the World Health Organization, World Bank and United States
Agency for International Development to verify the information contained in this publication. However, the pub-
lished material is being distributed without warranty of any kind, either expressed or implied. The responsibility
for the interpretation and use of the material lies with the reader. In no event shall the World Health Organization,
World Bank or United States Agency for International Development be liable for damages arising from its use.
The named authors of each chapter alone are responsible for the views expressed in this publication.
Photos: Front cover, left; back cover; pages 23 and 157: Curt Carnemark, World Bank. Front cover, center;
and page 1: Ray Witlin, World Bank. Front cover, right; and page 61: Tran Thi Hoa, World Bank.
Printed in India
Editing, design and indexing by Inís Communication: www.inis.ie
Contents
Contents
List of tables, figures and boxes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .vi
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xi
List of contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii
Acronyms and abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
Part I: OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1 Monitoring and evaluation of human resources for health: challenges and opportunities . . . . . . . 3
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Global initiatives on HRH and information systems . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Key issues and challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4 Framework for health workforce monitoring: the working lifespan approach . . . . . . . . . . . . . 8
1.5 Road map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.6 Further information and comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2 Boundaries of the health workforce: definition and classifications of health workers. . . . . . . . . 13
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2 Who are health workers? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.3 Health workforce classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4 Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Part II: MONITORING THE STAGES OF THE WORKING LIFESPAN. . . . . . . . . . . . . . . . . 23
3 Monitoring the active health workforce: indicators, data sources and illustrative analysis . . . . . . 25
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.2 Core indicators for HRH analysis: what needs to be monitored? . . . . . . . . . . . . . . . . . . 26
3.3 Overview of potential data sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.4 Illustrative analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.5 Putting it all together: governance and use of HRH information sources. . . . . . . . . . . . . . 33
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
iii
Handbook on monitoring and evaluation of human resources for health
4 Framework and measurement issues for monitoring entry into the health workforce . . . . . . . . . 37
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.2 Framework for monitoring entry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

4.3 Measurement issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
5 Monitoring health workforce transitions and exits . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
5.2 Transitions within and exits from the health workforce: a framework for analysis. . . . . . . . . . 49
5.3 Indicators and measurement strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
5.4 Illustrative analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
5.5 Concluding remarks: implications for policy and planning . . . . . . . . . . . . . . . . . . . . . 57
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Part III: MEASUREMENT STRATEGIES AND CASE STUDIES . . . . . . . . . . . . . . . . . . . . 61
6 Measuring expenditure on the health workforce: concepts, data sources and methods . . . . . . . 63
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
6.2 What should be measured . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
6.3 Approaches to measuring HRH expenditure . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
6.4 Measurement frameworks and applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
6.5 Summary, conclusions and further developments . . . . . . . . . . . . . . . . . . . . . . . . . 76
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
7 Use of facility-based assessments in health workforce analysis . . . . . . . . . . . . . . . . . . . . 79
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
7.2 How facility-based assessments can be used for health workforce monitoring . . . . . . . . . . 80
7.3 Overview of key HFA methodologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
7.4 Some limitations of HFA methodologies for HRH data . . . . . . . . . . . . . . . . . . . . . . . 85
7.5 Empirical examples based on HFA data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
8 Use of population census data for gender analysis of the health workforce . . . . . . . . . . . . . 103
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
8.2 Importance of gender considerations in health workforce analysis . . . . . . . . . . . . . . . 103
8.3 Using census data for health workforce analysis . . . . . . . . . . . . . . . . . . . . . . . . . 104
8.4 Empirical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
8.5 Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .110
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
iv
Contents
9 Use of administrative data sources for health workforce analysis: multicountry experience in
implementation of human resources information systems. . . . . . . . . . . . . . . . . . . . . . . .113
9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .113
9.2 Recommended first steps to develop a human resources information system. . . . . . . . . . .114
9.3 Country case studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .119
9.4 Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
10 Understanding health workforce issues: a selective guide to the use of qualitative methods . . . 129
10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
10.2 Qualitative methods: their value and potential. . . . . . . . . . . . . . . . . . . . . . . . . . 130
10.3 Issues in the design and implementation of a qualitative study . . . . . . . . . . . . . . . . . 132
10.4 Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
11 Analysis and synthesis of information on human resources for health from multiple sources:
selected case studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .147
11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .147

11.2 Identification of potential information sources and their use to estimate indicators of health labour
market participation in Mexico. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .147
11.3 Using multiple sources of information to produce best estimates of India’s health workforce . 150
11.4 Triangulation of data from two different sources for monitoring health worker absenteeism and
ghost workers in Zambia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
11.5 Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Part IV: DATA DISSEMINATION AND USE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
12 Getting information and evidence into policy-making and practice: strategies and mechanisms. 159
12.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159

12.2 Strategies to get evidence into policy and practice . . . . . . . . . . . . . . . . . . . . . . . 160
12.3 A mechanism to harness the HRH agenda: health workforce observatories . . . . . . . . . . 162
12.4 Opportunities and directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .170
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .173
About the publishing agencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .176
v
List of tables,
figures and boxes
TABLES
Table 2.1 Framework for defining the health workforce. . . . . . . . . . . . . . . . . . . . . . . . . . . . .14

Table 2.2 Relevant levels of education and training for health occupations according to the International
Standard Classification of Education (ISCED-1997) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Table 2.3 Fields of vocational training related to health according to Fields of training manual . . . . . . . .17
Table 2.4 Occupational titles related to health according to the International Standard Classification of
Occupations (ISCO), 1988 and 2008 revisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18
Table 2.5 Economic sectors related to health activities on the basis of the International Standard Industrial
Classification of All Economic Activities (ISIC) Revision 4 . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Table 3.1 Selected key indicators for monitoring and evaluation of human resources for health . . . . . . 28
Table 3.2 Potential data sources for monitoring the health workforce . . . . . . . . . . . . . . . . . . . . 30
Table 3.3 Selected indicators and means of verification for monitoring implementation and use of the
national HRH information and monitoring system to support decision-making. . . . . . . . . . . . . . . . 35
Table 4.1 Key indicators and means of verification for measuring entry into the health workforce. . . . . . 44
Table 5.1 Potential sources of data on health workforce transitions and exits . . . . . . . . . . . . . . . . .51
Table 5.2 Annual numbers of overseas-trained nurses obtaining national licensure to practise in the
United Kingdom, 1998–2007 (main countries of origin outside the European Economic Area) . . . . . . . 54
Table 5.3 Estimates of annual losses due to mortality under age 60 among health workers in selected
countries of the WHO Africa Region, based on life table analysis . . . . . . . . . . . . . . . . . . . . . . 55
Table 6.1 Percentage distribution of expenditure by type of health-care agent, Mexico health accounts,
1995 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Table 6.2 Percentage distribution of expenditure by main providers, Peru health accounts, 2000 . . . . . 73
Table 6.3 Remuneration components in the “use” table of the income account, System of National
Accounts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .74
Table 6.4 National accounts supply and use table, South Africa, 2002 (millions Rand, partial display) . . . 75
Table 7.1 Estimated number of new graduates entering the public sector facility-based health workforce,
by cadre, Nigeria 2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Table 7.2 Number and density of facility-based health workers, by cadre, Zambia 2006 HFC . . . . . . . 88
Table 7.3 Mean number of health workers by type of facility, according to cadre, Kenya 2004 . . . . . . . 89
Table 7.4 Number of health workers currently in post, number recommended by staffing norm, and number
requested by facility managers to meet the norm, by cadre, Kenya 2004 . . . . . . . . . . . . . . . . . . 90
Table 7.5 Percentage of facilities reporting having at least one seconded health worker, by cadre,
according to management authority of the facility, Kenya 2004 . . . . . . . . . . . . . . . . . . . . . . . 91
Table 7.6 Skills mix of facility-based health workers, by type of facility, Kenya 2004. . . . . . . . . . . . . 92
Table 7.7 Percentage distribution of health workers in post at health facilities by managing authority,
according to cadre, Kenya 2004 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Table 7.8 Percentage distribution of facility-based health workers and of the total population by
geographical region, Kenya 2004 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Table 7.9 Median number of years of service in current position among facility-based health workers, by
type and management authority of facility, Kenya 2004 . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
vi
List of tables, figures and boxes
Table 7.10 Percentage of facility-based health workers with written job descriptions, perceived promotion
opportunities and other non-monetary incentives, by cadre and type of facility, Kenya 2004 . . . . . . . . 96
Table 7.11 Percentage distribution of facility-based health workers by number of hours normally worked
per week, according to cadre, Kenya 2004. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Table 7.12 Comparison of the increase in stock of the public sector health workforce from new graduates
with the attrition rate, Nigeria 2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Table 8.1 Countries and sources of census data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
Table 8.2 Percentage distribution of the health workforce by sex, by occupation, according to census data
for 13 countries (around 2000) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Table 8.3 Ratio of women to men by level of educational attainment, health workforce and total population,
according to census data for selected countries (around 2000) . . . . . . . . . . . . . . . . . . . . . . . 109
Table 8.4 Estimates of intercensal health workforce attrition by sex, Thailand, 1990 and 2000 censuses . 110
Table 9.1 Selected indicators and benchmarks from the health workforce information and monitoring
system in Sudan, 2006–2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
Table 9.2 Selected indicators and corresponding criteria used for evaluating the performance of the
HRH information and management system in Brazil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Table 9.3 Selected indicators and results from the evaluation of the institutional impact of the HRH
information and management system in Brazil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Table 10.1 Contrasting qualitative and quantitative approaches . . . . . . . . . . . . . . . . . . . . . . . 131
Table 10.2 Illustrations of qualitative research on health workers . . . . . . . . . . . . . . . . . . . . . . 133
Table 10.3 Group discussions or individual interviews? . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Table 11.1 Questions on education and labour activity included in the national population census and
intercensal counts, Mexico, 1995–2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
Table 11.2 Stock and distribution of the physician and nursing workforce by labour force status, based
on census and survey data, Mexico, 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Table 12.1 Main stakeholders and their role in support of the national HRH observatory in Sudan . . . . . 166
Table 12.2 Influences of selected achievements of the Brazilian Human Resources for Health Observatory
on policy processes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
FIGURES
Figure 1.1 Working lifespan approach to the dynamics of the health workforce . . . . . . . . . . . . . . . . 9
Figure 3.1 Stocks and flows of the health workforce . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Figure 3.2 Geographical distribution of the stock of health workers (per 100 000 inhabitants), Viet Nam
and Mexico censuses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Figure 3.3 Trends in the distribution of workers in health services by main occupational group, Namibia
labour force surveys, 1997–2004 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Figure 3.4 Distribution of health workers’ salaries by source, according to facility ownership, Rwanda
health facility assessment, 2006 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Figure 3.5 Gender and skills mix of the nursing workforce in Kenya, Nursing Council of Kenya database,
1960–2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Figure 4.1 Framework for monitoring entry into the health workforce . . . . . . . . . . . . . . . . . . . . 38
Figure 5.1 Transitions within and exits from the health workforce: a framework for analysis . . . . . . . . . 50
Figure 5.2 Foreign-born nurses and doctors enumerated in 24 OECD countries by main countries of origin
(population census data, around 2000) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
Figure 5.3 Estimated lifetime emigration rate of physicians born in selected non-OECD countries
and working in OECD countries at the time of the census . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Figure 5.4 Retirement rate among physicians by age group according to the National Medical Association
registry, Canada, 2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
vii
Figure 5.5 Proportion of survey respondents reporting a health occupation but not working in the health
services industry at the time of interview, selected countries. . . . . . . . . . . . . . . . . . . . . . . . . 56
Figure 6.1 Overview of the estimation process for measuring labour expenditure in the System of National
Accounts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Figure 6.2 Information on the labour market in the Netherlands’ national accounts . . . . . . . . . . . . . 68
Figure 6.3 Process to estimate hours worked in the Canadian national accounts . . . . . . . . . . . . . . 69
Figure 6.4 Calculation square . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Figure 6.5 Resource flows in a health system: an accounting representation . . . . . . . . . . . . . . . . 72
Figure 7.1 Ratio of facility-based health workers aged under 30 to those aged over 45, by cadre,
Kenya 2004 SPA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Figure 7.2 Number and percentage distribution of staff currently in post at health facilities, by cadre,
Kenya 2004 SPA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Figure 7.3 Ratio of health workers per inpatient bed, by facility size, Kenya 2004 SPA . . . . . . . . . . . 89
Figure 7.4 Percentage of health workers requested by facility managers to meet staffing requirements
who are currently in post, by facility type, Kenya 2004 SPA . . . . . . . . . . . . . . . . . . . . . . . . . 91
Figure 7.5 Percentage of health workers and of the total population located in urban areas, Zambia
2006 HFC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
Figure 7.6 Median number of years of education and training among health workers, by cadre, Kenya
2004 SPA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
Figure 7.7 Percentage of health workers who received in-service training in the past 12 months, and who
received personal supervision in the last 6 months, by type of facility, Kenya 2004 SPA . . . . . . . . . . 94
Figure 7.8 Percentage of assigned health workers present on the day of the assessment, by cadre,
Kenya 2005 SAM (selected districts) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Figure 7.9 Attrition rates for health workers in public sector facilities, by cadre, Nigeria 2005 . . . . . . . 98
Figure 7.10 Percentage distribution of outgoing health workers by reason for leaving the workforce,
Nigeria 2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
Figure 8.1 Sex distribution of the physician workforce, according to census data for selected countries
(around 2000) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
Figure 8.2 Sex distribution of the personal care workforce, according to census data for selected
countries (around 2000). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
Figure 8.3 Sex distribution of health systems support staff, by occupation, Uganda, 2002 census. . . . . 108
Figure 8.4 Relationship between sex ratio in tertiary-level educational attainment and health professional
work activity, according to census data for selected countries (around 2000) . . . . . . . . . . . . . . . . 108
Figure 9.1 Framework for institutionalizing a human resources information system . . . . . . . . . . . . . 115
Figure 9.2 Number of students entering nursing training programmes (leading to qualification as
registered or enrolled nurses) in Uganda, 1980–2004 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Figure 9.3 Number of student nurses who entered training between 1980 and 2004, passed the
professional licensing exam, and qualified, registered and became licensed with the Uganda Nurses
and Midwives Council. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Figure 9.4 Number and percentage of student nurses who passed the professional licensing exam and regis-
tered with the Uganda Nurses and Midwives Council, by school district (entrants between 1980 and 2001) 121
Figure 11.1 Density of the health workforce by cadre, according to data source, India 2005 . . . . . . . . 151
Figure 11.2 Density of the health workforce (per 10 000 population) based on census versus survey data,
by state, India 2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
Figure 11.3 Percentage of facility-based health workers on the duty roster but not accounted for on the
day of the assessment, by cadre, Zambia 2006 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Figure 11.4 Percentage of health workers recorded on the Ministry of Health payroll but not registered at
the facility level, Zambia 2006 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Figure 12.1 Basis for HRH policies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
viii
List of tables, figures and boxes
BOXES
Box 1.1 Defining monitoring and evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Box 1.2 Toolkit for monitoring health systems strengthening . . . . . . . . . . . . . . . . . . . . . . . . . 4
Box 1.3 Health-related Millennium Development Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Box 1.4 Assessment and strengthening of HRH information systems . . . . . . . . . . . . . . . . . . . . . 6
Box 1.5 Financial resource needs for strengthening HRH information systems . . . . . . . . . . . . . . . . 7
Box 1.6 Human resource needs for strengthening HRH information systems . . . . . . . . . . . . . . . . . 7
Box 1.7 Technological resource needs for strengthening HRH information systems . . . . . . . . . . . . . 7
Box 4.1 Illustrative example of the establishment of health worker education and training goals and
targets: region of the Americas, 2007–2015 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Box 4.2 Some notes on data collection, processing and use . . . . . . . . . . . . . . . . . . . . . . . . 43
Box 4.3 Illustrative example of data on trends in medical education institutional capacity, recruitment and
output: Rwanda, 2002–2006 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Box 6.1 Proposed minimum set of indicators for monitoring expenditure on human resources for health . 64
Box 6.2 Expenditure components under a health accounting approach through classes of the
International Standard Industrial Classification of All Economic Activities (fourth revision). . . . . . . . . . 66
Box 6.3 Labour data estimation criteria in the national accounts of the Russian Federation . . . . . . . . 69
Box 6.4 Defining remuneration of employees and self-employment income . . . . . . . . . . . . . . . . 70
Box 7.1 Core health workforce indicators potentially measurable with HFA data . . . . . . . . . . . . . . 80
Box 8.1 Questions used in the population census to determine main occupation, selected countries . . . 105
Box 9.1 Confidentiality and security of HRH information. . . . . . . . . . . . . . . . . . . . . . . . . . . 118
Box 9.2 Implementation and use of the SIGRHS information system for HRH management in
Guinea Bissau. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
Box 10.1 Three important techniques in qualitative research . . . . . . . . . . . . . . . . . . . . . . . . 130
Box 10.2 Qualitative research to inform quantitative work: absenteeism in Ethiopia and Rwanda . . . . . 132
Box 10.3 Types of qualitative interviews according to their degree of structure . . . . . . . . . . . . . . . 136
Box 10.4 Extract of semi-structured interview script of qualitative health worker study in Ethiopia and
Rwanda . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
Box 10.5 Selected quotes from qualitative interviews in Ethiopia and Rwanda . . . . . . . . . . . . . . . 138
Box 10.6 Selection criteria for focus groups with eight participants on the performance and career
choice of nurses and midwives in Ethiopia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
Box 10.7 Extract from the introduction to focus group participants in Rwanda . . . . . . . . . . . . . . . 140
Box 10.8 Matrix-based analysis of data from focus group discussions in Ethiopia . . . . . . . . . . . . . 142
Box 10.9 Recording, transcription, coding and analysis: an illustration from a health worker study in
Rwanda . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
Box 12.1 Evidence-Informed Policy Network: EVIPNet . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
Box 12.2 Africa Health Workforce Observatory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
Box 12.3 Brazil Human Resources for Health Observatory . . . . . . . . . . . . . . . . . . . . . . . . . 166
Box 12.4 Ghana Health Workforce Observatory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
Box 12.5 Andean Human Resources for Health Observatory Network . . . . . . . . . . . . . . . . . . . 168
ix
Preface
H E L AT E S T M E D I C I N E and the newest technologies can have little impact on human
T health unless there are systems in place to deliver them. The reality today, however, is that
health systems all over the world are suffering from years of neglect. One of the most obvious
manifestations of that neglect is a crippling lack of trained health workers.
In many countries, lack of personnel is one of the most important constraints to strengthening
the delivery of primary and other health services, including curative, promotional, preventive
and rehabilitative services. In sub-Saharan Africa, the health workforce crisis is so great that
36 countries are considered to have a critical shortage of health care professionals to provide
minimum coverage of even the most basic services in maternal, newborn and child health. In
many of the poorest countries of the world, the situation is worsened by the continual loss of
health personnel seeking better opportunities elsewhere.
The effects of poor workforce planning and development are felt everywhere. In Asia and the
Pacific, many countries have a shortage of health workers capable of treating chronic and
emerging diseases. In Europe, the countries of the newly independent states of the former
Soviet Union inherited a workforce that was especially ill-suited to the demands facing mod-
ern health care systems.
Clearly, if countries are to get anywhere near meeting their health system objectives, includ-
ing the health-related Millennium Development Goals, they need to be able to provide better
access to appropriately trained health workers.
The first step is to work out where the gaps are. Yet many countries currently lack the technical
capacity to accurately monitor their own health workforce: data are often unreliable and out-
of-date, common definitions and proven analytical tools are absent, skills and experience for
assessing crucial policy issues are lacking.
This Handbook aims to increase that technical capacity. It offers health managers, researchers
and policy makers a comprehensive, standardized and user-friendly reference for monitoring
and evaluating human resources for health. It brings together an analytical framework with
strategy options for improving the health workforce information and evidence base, as well as
country experiences to highlight approaches that have worked.
We gratefully acknowledge here the support provided by the three collaborating partners: the
United States Agency for International Development, the World Bank and the World Health
Organization. All three will continue to support countries in their application of the Handbook
to national contexts, as one more way to accelerate country action towards building a trained,
sufficient and motivated health workforce.
Carissa F. Etienne Yaw Ansu Gloria D. Steele

Assistant Director-General, Director, Human Development, Acting Assistant Administrator,
Health Systems and Services Africa Region Global Health
World Health Organization The World Bank United States Agency for
International Development
x
Acknowledgements
Acknowledgements
The present publication is the result of a collaborative effort between the United States Agency
for International Development, the World Bank and the World Health Organization. The vol-
ume was written by an international group of authors and edited by Mario R Dal Poz, Neeru
Gupta, Estelle Quain and Agnes LB Soucat. The list of contributors to the Handbook’s chap-
ters follows.
We would like to acknowledge a number of people and organizations whose contributions

were essential to the preparation of this publication. We wish to thank the data collection and
dissemination organizations and agencies who kindly lent permission to use their information
sources for the empirical analyses presented here, including: Federal Ministry of Health of
the Republic of the Sudan; International Health Policy Program, Thailand; Macro International;
Mexico National Institute of Statistics, Geography and Informatics; Minnesota Population
Center; Thailand National Statistical Office; Uganda Nurses and Midwives Council; and the
World Health Organization. Financial support for the production of this volume was provided
by the Bill & Melinda Gates Foundation, the Government of Norway, the United States Agency
for International Development, the World Bank and the World Health Organization. Elizabeth
Aguilar, Rosa Bejarano, Florencia Lopez Boo, Bénédicte Fonteneau and Dieter Gijsbrechts
provided research and analysis support. We also thank Yaw Ansu for his ongoing support.
Among the individuals who provided useful suggestions to earlier drafts of this final product
were: Adam Ahmat, Elsheikh Badr, James Buchan, Hartmut D Buchow, Arturo de la Fuente,
Delanyo Dovlo, Norbert Dreesch, Thushara Fernando, Paulo Ferrinho, Juliet Fleischl, Linda
Fogarty, Kara Hanson, Christopher H Herbst, Eivind Hoffmann, Beatriz de Faria Leao, Verona
Mathews, Robert McCaa, Edgar Necochea, Rodel Nodora, Jennifer Nyoni, Bjorg Palsdottir,
Jean Pierre Poullier, Magdalena Rathe, Martha Rogers, Kate Tulenko, Cornelis van Mosseveld
and Pablo Vinocur.
The named authors for each chapter alone are responsible for the views expressed in this
publication.
xi
List of contributors
Walid Abubaker, World Health Organization Hugo Mercer, World Health Organization
Ghanim Mustafa Alsheikh, World Health Organization Gustavo Nigenda,

Instituto Nacional de Salud Pública de México
Aarushi Bhatnagar, Public Health Foundation of India
Shanthi Noriega Minichiello, Health Metrics Network
Bates Buckner, MEASURE Evaluation/
University of North Carolina at Chapel Hill Ulysses Panisset, World Health Organization
Mario R Dal Poz, World Health Organization Priya Patil, MEASURE Evaluation/Constella Futures
Danny de Vries, Capacity Project/ Célia Regina Pierantoni,

IntraHealth International Universidade do Estado do Rio de Janeiro
Khassoum Diallo, United Nations High Commission Bob Pond, Health Metrics Network
for Refugees
Estelle Quain,
Gilles Dussault, Universidade Nova de Lisboa United States Agency for International Development
David B Evans, World Health Organization Shomikho Raha, World Bank
Bolaji Fapohunda, MEASURE Evaluation/JSI Inc. Krishna D Rao, Public Health Foundation of India
Nancy Fronczak, Felix Rigoli, Pan American Health Organization/

Social Sectors Development Strategies World Health Organization
Gülin Gedik, World Health Organization José Arturo Ruiz,

Instituto Nacional de Salud Pública de México
Neeru Gupta, World Health Organization
Indrani Saran, Public Health Foundation of India
Piya Hanvoravongchai, Asia-Pacific Action Alliance
on Human Resources for Health Catherine Schenck-Yglesias,
United States Agency for International Development
Christopher H Herbst, World Bank
Pieter Serneels, World Bank
Patricia Hernandez, World Health Organization
Dykki Settle, Capacity Project/IntraHealth International
David Hunter, International Labour Organization
Agnes LB Soucat, World Bank
Yohannes Kinfu, World Health Organization
Tessa Tan-Torres, World Health Organization
Teena Kunjumen, World Health Organization
Kate Tulenko, World Bank
René Lavallée, Consultant
Shannon Turlington, Capacity Project/
Tomas Lievens, Oxford Policy Management
IntraHealth International
Magnus Lindelow, World Bank
Pascal Zurn, World Health Organization
Pamela McQuide, Capacity Project/
IntraHealth International
xii
Acronyms and abbreviations
Acronyms and
abbreviations
COFOG Classification of the Functions of Government
COPP Classification of the Outlays of Producers According to Purpose
CPC Central Product Classification
EVIPNet Evidence-Informed Policy Network
HFA health facility assessment
HFC Health Facility Census
HRH human resources for health
HRIS human resources information system
ICSE International Classification of Status in Employment
IPUMS Integrated Public Use Microdata Series
ISCED International Standard Classification of Education
ISCO International Standard Classification of Occupations
ISIC International Standard Industrial Classification of All Economic Activities
JICA Japanese International Cooperation Agency
OECD Organisation for Economic Co-operation and Development
PAHO Pan American Health Organization
PALOP Países Africanos de Língua Oficial Portuguesa
PHRplus Partners for Health Reformplus
SAM Service Availability Mapping
SHA System of Health Accounts
SIGRAS sistema de informação sobre a graduação em saúde
SIGRHS sistema de informação e gestão de recursos humanos em saúde
SNA System of National Accounts
SPA Service Provision Assessment
USAID United States Agency for International Development
VCT voluntary counselling and testing
WHO World Health Organization
xiii
xiv
Part I:
OVERVIEW
1
Monitoring and evaluation of
1 human resources for health:

challenges and opportunities
MARIO R DAL POZ, NEERU GUPTA, ESTELLE QUAIN, AGNES LB SOUCAT
1.1 Introduction (HRH) development; continues with a discussion on

key challenges on the uses of health workforce infor-
Health systems and services depend critically on the mation for planning, policy and decision-making; and
size, skills and commitment of the health workforce. It proposes a comprehensive framework for HRH mon-
is now evident that in many low- and middle-income itoring. Presentation of this overview and framework
countries, meeting key Millennium Development Goal orients readers to the contents of the rest of the volume.
targets, specifically those relating to health, requires
a significant increase in the numbers of health work- This publication is the result of a collaborative effort
ers (1–3). The global shortage is estimated at around between the United States Agency for International
2.3 million physicians, nurses and midwives, and over Development (USAID), the World Bank and the World
4 million health workers overall. In some parts of the Health Organization (WHO) to document method-
world, notably in sub-Saharan Africa, the current work- ologies and share experiences in measuring and
force needs to be scaled up by almost 140% in order monitoring HRH, to encourage countries and part-
to overcome the crisis (4). And simply assessing num- ners to build upon these experiences and to compile
bers of health workers in relation to a given threshold recommendations for ministries of health and other
does not necessarily take into account all of a health stakeholders for health workforce monitoring and
system’s objectives, particularly with regard to accessi- evaluation. It builds upon and complements other inter-
bility, equity, quality and efficiency. nationally coordinated efforts for monitoring the building
blocks of health systems, including human resources
Countries with critical shortages and imbalances of (Box 1.2) (7). It is anticipated that this Handbook will
health workers also often lack the technical capacity to contribute to effective use of existing data and improve
identify and assess crucial policy issues related to the future data collection efforts for maximum utility of
health workforce. As a result, fundamental questions quantitative and qualitative HRH assessments within
regarding the status of the workforce, its level of per- countries, across countries and over time. The publi-
formance and the problems health workers face remain cation should be seen as a work in progress that will
largely unanswered. In addition, the lack of compre- result in an enhanced understanding of HRH and con-
hensive, reliable and up-to-date data, and the absence tribute significantly to the growing body of tools and
of commonly agreed definitions and analytical tools, applied research designed to address the challenge of
have made the task of monitoring the health workforce measuring and improving health workforce outcomes,
all the more difficult in all settings, from the global and strengthening health systems and, ultimately, improv-
regional to the national and subnational levels. Such ing population health.
challenges remain significant to many governments
striving to maintain a sufficient, sustainable and effec-
tive health workforce in their respective jurisdictions.
1.2 Global initiatives on HRH
This Handbook is a response to the need to have a and information systems
comprehensive, standardized and user-friendly refer-
The health workforce is increasingly recognized as
ence on health workforce monitoring and evaluation
key for scaling up health interventions for achieving
as a means to develop in-country capacity to build
the Millennium Development Goals (Box 1.3) (8). Even
the knowledge base needed to guide, accelerate and
with additional funds available in recent years from
improve country action (Box 1.1) (5, 6). This introduc-
international, multilateral, bilateral and private sources,
tory chapter begins with an overview of recent global
such as through poverty reduction strategies and debt
initiatives for supporting human resources for health
3
Box 1.1 Defining monitoring and evaluation
Monitoring. The ongoing process of collecting and using standardized information to assess progress
towards objectives, resource usage and achievement of outcomes and impacts. It usually involves
assessment against agreed performance indicators and targets. In conjunction with evaluation information,
effective monitoring and reporting should provide decision-makers and stakeholders with the knowledge
they need to identify whether the implementation and outcomes of a project, programme or policy initiative
are unfolding as expected and to manage the initiative on an ongoing basis.
Evaluation. The systematic and objective assessment of an ongoing or completed initiative, its design,
implementation and results. The aim is to determine the relevance and fulfilment of objectives, efficiency,
effectiveness, impact and sustainability. The development of an evaluation framework entails consideration
of a range of matters, including identification of the types of data that could inform an evaluation.
Indicator. A parameter that points to, provides information about or describes a given state. Usually
represented by a data element for a specified time, place and other characteristics, it gives value as an
instrument used in performance assessment.
Data. Characteristics or information, often numerical, that are collected through observation. Data can be
considered as the physical representation of information in a manner suitable for processing, analysis,
interpretation and communication.
Sources: Adapted from Deloitte Insight Economics (5) and Organisation for Economic Co-operation and Development (6).
Box 1.2 Toolkit for monitoring health systems strengthening
The ability to plan, monitor and evaluate health systems functioning is essential in order to correctly target
investments and assess whether they are having the intended impact. Health systems can be described
in many ways. The WHO framework delineates six core building blocks: service delivery, health workforce,
financing, information, leadership and governance, and medical products and technologies. Through
collaboration between WHO, the World Bank, country health information and systems experts and many
other organizations working in this field, a toolkit was developed proposing a limited set of indicators
and related measurement strategies covering each of the building blocks. As of mid-2008, the Toolkit
for monitoring health systems strengthening was being made available while still in draft form to invite
comments from a wide array of potential users.
Source: World Health Organization (7).
alleviation programmes, or through newer modalities Joint Learning Initiative (13), the WHO flagship publica-
such as the Global Fund to Fight AIDS, Tuberculosis tion The world health report 2006: working together for
and Malaria (9), the GAVI Alliance (10) and the United health (4), the resolutions of World Health Assemblies
States President’s Emergency Plan for AIDS Relief on health workforce development (14), and the launch
(11), country capacity to absorb funds and to put of the Global Health Workforce Alliance (15), as well
them to work can be severely crippled by the crisis as certain regional partnership mechanisms such as
in HRH. In many countries there is simply insufficient the Asia-Pacific Action Alliance on Human Resources
human capacity at all levels to absorb, deploy and use for Health (16), were among a cluster of international
efficiently the financing for scaling up health services activities that alerted national, regional and interna-
delivery offered by recent initiatives. tional policy-makers and stakeholders, including the
media, civil society and the general public, to the criti-
A series of high-level forums on the health Millennium cal importance of HRH worldwide, especially the HRH
Development Goals (12), the HRH strategy report of the crisis in sub-Saharan Africa.
4
Monitoring and evaluation of human resources for health: challenges and opportunities
Box 1.3 Health-related Millennium Development Goals
In September 2000, at the United Nations Millennium Summit, all countries committed to collectively
working towards a series of eight goals and 18 related targets for combating poverty and its determinants
and consequences, under a compact known as the Millennium Development Goals. Much investment
was subsequently allocated for measuring progress towards the achievement of these targets, including
a strong focus on monitoring coverage of prioritized health interventions and population health outcomes.
More recently, attention has been directed to addressing and monitoring the health systems inputs,
processes and outputs that impede or facilitate progress.
Goals Related targets

1. Eradicate extreme poverty and hunger Halve, between 1990 and 2015, the proportion of people
who suffer from hunger
4. Reduce child mortality Reduce by two thirds, between 1990 and 2015, the
under-five mortality rate
5. Improve maternal health Reduce by three quarters, between 1990 and 2015, the
maternal mortality ratio
6. Combat HIV/AIDS, malaria and other Have halted by 2015 and begun to reverse the spread of
diseases HIV/AIDS
Have halted by 2015 and begun to reverse the incidence

of malaria and other major diseases
7. Ensure environmental sustainability Halve by 2015 the proportion of people without
sustainable access to safe drinking-water and sanitation
8. Develop a global partnership for In cooperation with pharmaceutical companies, provide
development access to affordable, essential drugs in developing
countries
Raising awareness of the critical role of HRH places the Health Partnership (17) and the Global Campaign for
health workforce high on global public health agendas. the Health Millennium Development Goals (18), empha-
Countries, donors, international agencies and other size the principles of health systems support to achieve
stakeholders are increasingly willing not only to invest improved health outcomes. Increased opportunities for
in but also to contribute to HRH development overall. funding health systems strengthening through primary
The critical role of human resources in the achieve- health care means more opportunities for investing in
ments of health systems objectives is more and more improving the quality of human resources.
recognized and valued.
To this end, countries are called upon to provide clear
At the same time, there is an increased demand for trans- and consistent evidence in their requests for both
parency and performance measurement. For instance, new and ongoing resources for HRH development.
the Global Fund and GAVI Alliance, which spear- This is also true for decision-making and allocation of
head the principles of performance-based release of resources from national sources. Ministries of health
donor funding, have recognized the need to channel in many low- and middle-income countries face addi-
more of their disease-specific funds towards sustaina- tional challenges posed by the effects of decentralizing
ble, comprehensive health care that is accessible and responsibilities for both budget and information sys-
affordable to all. Other recent international initiatives tems to district authorities, with often incomplete or
seeking to accelerate progress towards achieving the inconsistent transfers of authority hampering proper
health-related Millennium Development Goals, together decision-making processes.
with all major stakeholders, including the International
5
Box 1.4 Assessment and strengthening of HRH information systems
Assessment and strengthening of national HRH information systems are integral components of efforts to
strengthen the evidence base on the health workforce. An assessment of the capacity of a country’s HRH
information system to support decision-making may include consideration of:
t timeliness of the system;
t validity of the information contained within the system;
t consistency across information sources (to allow for comparisons within and across countries and over time);
t level of disaggregation of the information within the system, to allow for in-depth analysis on issues of
relevance to HRH strategic planning.
Results of an assessment of the HRH information
systems in selected countries, 2006–2007
In 2006–2007, in collaboration with the Health
Metrics Network (19), assessments of the 1.6
Tracks stock 1.3
national health information system were and mix of HRH 2.0
conducted in selected low- and middle-income 1.6
1.9
countries, drawing on a standard assessment Tracks output of health 0.5
professions education
and monitoring tool. Using a scaled institutions
0.9
0.8
questionnaire in relation to four core dimensions 3.0
of national capacity and contents of the HRH Regularly updated 2.4
2.6
database, results of the self-evaluations 2.2
showed a score of 6.1 out of a possible 12 in Adequate human 1.0
1.0
Afghanistan, 6.9 in Eritrea, 5.2 in the Republic resources to maintain 1.4
databases
of Moldova and 7.5 in Sudan (see graph). Of 1.5
the four core dimensions, tracking the output 0 1 2 3

Average score
of health professions education institutions was
ranked lowest in most countries. Afghanistan Eritrea Moldova Sudan
The importance of sound empirical evidence for allocation of resources. In many countries, the prob-
informed policy development, decision-making and the lem partly emanates from the fragmentation of HRH
monitoring of progress towards achieving HRH devel- information and the shortages in human, financial and
opment and strengthening health systems is widely infrastructural resources available to collect, compile
recognized. Evidence is needed to support countries and analyse health workforce data (Boxes 1.5–1.7) (21–
to make the case for HRH both in national budget allo- 24). Moreover, the lack of standard tools, indicators,
cation and in their cooperation with donors. However, definitions and systems of classifying health workers
knowledge about what works and what does not is has placed further constraints on using HRH informa-
still very limited, signalling a need for more evidence tion for evidence-informed decision-making.
and further research. The launch of the Health Metrics
Network (19), a partnership aiming to increase the Strengthening HRH information and monitoring sys-
availability and use of timely and accurate health infor- tems requires a better foundation for policy-making,
mation by catalysing the joint funding and development planning, programming and accountability. A range of
of core country health information systems, was a big tools and resources exists to assist countries in devel-
step forward in improving the information and evidence oping a national HRH strategic plan (25–27); technical
base, including on HRH (Box 1.4). assistance for developing and costing these plans can
be sought, but having the necessary underlying data is
However, despite the global initiatives to assist in a prerequisite. The Kampala Declaration and Agenda
this core area of health systems strengthening, there for Global Action, adopted by the First Global Forum
is little consistency between countries in how HRH on Human Resources for Health (28), called upon
strategies are monitored and evaluated (20), thereby governments, in cooperation with international organ-
limiting the capacity of stakeholders to rationalize the izations, civil society, the private sector, professional
6
Box 1.5 Financial resource needs for strengthening HRH information systems
Little research has been undertaken into the financial investment levels needed to ensure a sound human
resources information and monitoring system, which can vary according to a country’s overall level of
development. Estimates of the cost of a comprehensive health information system, including a strong
HRH component, range from US$ 0.53 to US$ 2.99 annually per capita (21). In general, guidelines
suggest that health information, monitoring and evaluation costs comprise between 3% and 11% of total
project funds (22).
Box 1.6 Human resource needs for strengthening HRH information systems
Improvements to information systems require attention to be given to the training, deployment, remuneration
and career development of human resources at all levels. At the national level, skilled epidemiologists,
statisticians, demographers and computer programmers and technicians are needed to oversee data
quality and standards for collection, and to ensure the appropriate analysis and utilization of information. At
the district and facility levels, health information staff should be accountable for data collection, reporting
and analysis. Too often, such tasks are given to overburdened health service providers, who see this as
unwelcome additional work that detracts from their primary role. Appropriate remuneration and supervision
is essential to ensure the availability of high-quality staff and to limit attrition. This implies, for example, that
health information positions in ministries of health (and other bodies mandated with data analysis and use)
should be graded at a level equivalent to those of major disease programmes.
Source: Health Metrics Network (23).
Box 1.7 Technological resource needs for strengthening HRH information systems
Many countries lack access to the necessary information and communications technologies for strengthening
their HRH information systems. For instance, a 2004 study conducted by the WHO Regional Office for Africa
showed that 22% of health workforce departments of ministries of health in the region did not have computer
facilities, 45% had no electronic mail access, and fax machines were available in only 32% of the surveyed
departments. Under these circumstances, even a modest investment could yield significant returns.
associations and other partners, to “create health work- for collecting, processing and disseminating compre-
force information systems, to improve research, and hensive timely information on their health workforce,
to develop capacity for data management in order to including stock, distribution, expenditures and determi-
institutionalize evidence-based decision making and nants of change. Different pieces of information may
enhance shared learning”. be derived from health professional regulatory bod-
ies, district health information records on health facility
staffing, population or establishment censuses or sur-
1.3 Key issues and challenges veys, payroll records, work permits or other sources.
As a result, ministries of health and other stakeholders
It is an unfortunate truth that countries most in need of often depend on ad hoc reports compiled from differ-
strengthening their HRH tend to have the most frag- ent sources, for which the completeness, timeliness
mented and unreliable data and information. Most, if and comparability are widely variable.
not all, countries lack a harmonized dedicated system
7
An additional challenge is that most countries do not 1.4 Framework for health
have comprehensive data that capture the multitude
workforce monitoring: the
of health workforce engaged in the preservation, pro-
motion and restoration of health. For instance, few working lifespan approach
countries routinely collect and disseminate official sta- The present Handbook uses a “working lifespan”
tistics on their health system management and support approach to monitoring the dynamics of the health work-
workforce, which includes a large range of managerial, force. Introduced in The world health report 2006 (4),
administrative, professional and clerical occupations this approach focuses on the need for monitoring and
as well as many others working in the health sector evaluating each of the stages when people enter (or re-
who are not necessarily formally trained in health serv- enter) the workforce, the period of their lives when they
ices provision. In The world health report 2006, barely are part of the workforce, and the point at which they
a third of countries reported such data (4). Yet, these make their exit from it. The lifespan approach (Figure
workers – who are a critical component of the health 1.1) of producing, attracting, sustaining and retaining
workforce – are estimated to account for close to a the workforce offers a worker perspective as well as a
third of all HRH worldwide; excluding them from offi- systems approach to monitoring the dynamics of the
cial counts results in a substantial underreporting of the health labour market and the strategies of each stage.
health workforce stock and neglects a sizeable poten-
tial to strengthen health systems performance. Many From policy and management perspectives, the frame-
countries also lack timely and reliable information on work focuses on modulating the roles of both labour
the various dimensions of HRH imbalances, such as markets and state action at key decision-making
distribution by sector, geography, gender, labour force junctures:
activity, place of work and remuneration. t entry: preparing the workforce through strategic
investments in education and effective and ethical
Even in countries where data are relatively available, recruitment practices;
the translation of information into evidence that is use- t active workforce: enhancing workforce availabil-
ful for planners, decision-makers and stakeholders ity, accessibility and performance through better
has been greatly hindered by the lack of consistency human resources management in both the public
in occupational classification and the challenges of and private sectors;
combining information from multiple sources. Given t exit: managing migration and attrition to reduce
the differences in national developments and culture wasteful loss of human resources.
that result in variations in the roles and tasks of health
workers, any attempt to enhance cross-national and A central objective of policy and programmatic inter-
time-trend comparability needs to focus on ways to ventions at the entry stage is to produce and prepare
harmonize data collection, processing and dissemina- sufficient numbers of motivated workers with ade-
tion approaches. quate technical competencies, whose geographical
and sociocultural distribution makes them accessible,
A further challenge in the development and strength- acceptable and available to reach clients and popu-
ening of HRH information systems comes from the lack lations in an efficient and equitable manner. To do so
of consensus on standardized indicators and underin- requires active planning, management and budgeting
vestment in measurement strategies, which are core across the health workforce production pipeline, with a
for monitoring and evaluation of HRH interventions. focus on building strong health professions education
Important efforts are still needed to harmonize a mini- institutions, enhancing quality control mechanisms for
mum set of indicators that are broadly reflective of the skilled workers and strengthening labour recruitment
various dimensions and complexities of HRH dynam- capabilities.
ics, simple and ready to measure, but comprehensive
enough to be of use for public health decision-mak- Strategies to improve the performance of the active
ing in the area of HRH. There are as yet no commonly health workforce focus on the availability, competence,
agreed and systematically reported indicators inter- appropriateness, responsiveness and productivity of
nationally, apart from density of the most common those currently engaged in the health sector. This gen-
categories of health professionals (physicians, nurses erally involves assessment of HRH within the context
and midwives). Even this indicator has serious limita- of health services delivery among a wide variety of
tions for policy and planning, as its relevance in setting workplaces, and across the broader context of national
periodic targets may be influenced by changes in pop- labour markets.
ulation structure, burden of disease, health workforce
skill mix and other factors. Unplanned or excessive exits or losses of health work-
ers may compromise health systems performance
8
Figure 1.1 Working lifespan approach to the dynamics of the health workforce
ENTRY:
Preparing the
workforce
Planning
Education
Recruitment WORKFORCE: WORKFORCE PERFORMANCE
Enhancing worker
Availability
performance
Competence
Supervision Responsiveness
Compensation Productivity
Systems support
EXIT: Lifelong learning
Managing attrition
Migration
Career choice
Health and safety
Retirement
and exacerbate the fragility of some already weak sys- at various junctions to help guide readers towards a
tems. In some regions, worker illnesses, deaths and broader vision. Although many of the approaches pre-
out-migration (to other sectors or countries) together sented here have been selected with the expectation
constitute a haemorrhaging that threatens workforce of being applied in low- and middle-income countries,
stability. Strategies to counteract workforce attrition illustrative examples from countries with developed
include managing the market pressures that can lead to market economies are also included in order to opti-
migration, improving workplace conditions so that the mize the sharing of experiences and best practices.
health sector is viewed as a favourable career choice,
and reducing risks to health workers’ health and safety. In Chapter 2, Hunter, Dal Poz and Kunjumen dis-
cuss the definition of the health workforce and its
At each of these stages, and for each of the policies operationalization, reviewing the current uses of inter-
and interventions in place, there is a need to develop nationally standardized classifications relevant for
and measure appropriate indicators to inform strategy statistical delineation, description and analysis of the
development and monitor the impacts and cost-effec- health workforce. The timeliness of this chapter is man-
tiveness over time. ifest, given the recent 2008 revision to the International
Standard Classification of Occupations.
1.5 Road map In Chapter 3, Rigoli and colleagues detail the

approaches to and means of monitoring the active
The following chapters in this Handbook present in health workforce. They identify core indicators for char-
detail the opportunities, challenges and country expe- acterizing those currently participating in the health
riences in approaches to monitoring and evaluating labour market and review potential sources of data. The
these key aspects of health workforce dynamics and chapter continues with illustrative examples using case
the utility of different potential information sources and studies from various countries and sources, and con-
analytical techniques. The volume is structured in 12 cludes with recommendations for strengthening HRH
chapters, including this introductory chapter. While information and monitoring systems within countries.
each chapter may be read on its own, a number of
cross-references to other chapters are also included
9
Tulenko, Dussault and Mercer explore in Chapter 4 Qualitative measurement strategies for HRH analy-
ways in which entry into the health workforce can be sis are the focus of Chapter 10. Qualitative studies
monitored and assessed, and how the appropriate are used to understand health workers’ attitudes and
data can lead to formulation and evaluation of policies motivations, looking behind the numbers produced in
and programmes to address shortages and maldis- quantitative research. Lievens, Lindelow and Serneels
tribution of health workers. They identify a set of core provide some practical guidance on how to design and
indicators, and existing and new sources of data for implement qualitative health workforce studies, draw-
their measurement. ing on a rich methodological literature and a wealth of
applied research, including case studies from Ethiopia
In Chapter 5, Zurn, Diallo and Kinfu discuss the major and Rwanda.
factors influencing transition within and exit from the
health workforce, and propose comprehensive but Given the diversity of potential HRH information
readily measurable performance indicators. Using rele- sources, a strategy of triangulation – or cross-exami-
vant data, the chapter also provides illustrative analyses nation and synthesis of the available data from different
with discussion of the implications of observed pat- sources – can be effective in supporting decision-mak-
terns for policy and planning. ing, as it allows for a rapid understanding of the situation
and makes optimal use of pre-existing data. This is the
In Chapter 6, Hernandez, Tan-Torres and Evans aim rationale behind Chapter 11, in which Nigenda and
to encourage a greater number of countries to monitor colleagues present three country case studies (from
expenditure on HRH to inform decision-making. Several Mexico, India and Zambia) on the uses of triangulation
lines of action are presented, intended to be a how-to for HRH analysis, each focusing on a central theme.
guide for operational use by those actually monitoring
these expenditures. An introduction to the main proce- Lastly, Gedik and colleagues take stock of various glo-
dures for data collection and the associated results or bal, regional and national initiatives in place to build
indicators that would be generated is included, as well effective cooperative mechanisms for sharing of
as country cases to illustrate various procedures. The knowledge and best practices. Chapter 12 closes the
chapter focuses on issues of data collection and use at Handbook with a discussion on conceptual contribu-
country level, describing how to begin to construct and tions and frameworks intending to link health workforce
maintain a database on HRH expenditure. research, information and analysis to policy dialogue
and decisions. The authors highlight the potential
The next four chapters are each centred on a specific role and experiences of regional and national HRH
measurement tool that can potentially be a rich source observatories for getting information and evidence into
of policy-relevant information. In Chapter 7, Fapohunda policy-making and practice.
and colleagues highlight the usefulness of health facil-
ity assessments for HRH analysis, reviewing a broad
array of data collection techniques that focus on facil-
ity-based service delivery points.
In Chapter 8, Lavallée, Hanvoravongchai and Gupta

present approaches to using population census data
for exploring gender dimensions of the health work-
force, drawing on multicountry empirical applications.
The authors conclude with recommendations for pro-
moting the use of sex-disaggregated data, notably
from census sources, as a step towards monitoring and
evaluation of gender-sensitive human resources policy
planning and management.
In Chapter 9, McQuide and colleagues provide an

overview of the essential elements and lessons learnt
to date from various experiences in the implementation
of human resources information systems drawing on
administrative data sources. They present a series of
case studies in developing and strengthening routine
information systems from selected low- and middle-
income countries.
10
1.6 Further information

and comments
This Handbook is part of broader efforts to enhance
country capacities to generate, analyse and use data
to assess health workforce performance and track
progress towards their HRH-related goals. Requests
for further information on any of the tools, methods or
approaches described here are welcomed. In order
to ensure that future revisions of the Handbook are
improved and remain responsive to country needs
and situations, comments, feedback and sugges-
tions are solicited from readers and potential users.
Some specific issues on which feedback are welcome
include user-friendliness of the Handbook; feasibility
and sustainability of the recommended indicators and
related measurement and analysis strategies; and the
Handbook’s helpfulness in stimulating country owner-
ship and demand for strengthened HRH information,
monitoring and evaluation systems.
Please send your questions, comments and feedback to:

Coordinator, Health Workforce Information
and Governance
Department of Human Resources for Health
World Health Organization
Avenue Appia 20
Geneva 1211, Switzerland
Fax: +41–22–791–4747
Email: hrhstatistics@who.int
11
References 18. Global Campaign for the Health Millennium

Development Goals: launch of the first year report.
1. Anand S, Barnighausen T. Human resources and Norwegian Agency for Development Cooperation
health outcomes: cross-country econometric study. (http://www.norad.no/default.asp?V_ITEM_
Lancet, 2004, 364:1603–1609. ID=9263&V_LANG_ID=0, accessed 10 January
2. Anand S, Barnighausen T. Health workers and 2009).
vaccination coverage in developing countries: an 19. Health Metrics Network (http://www.who.int/
econometric analysis. Lancet, 207, 369:1277–1285. healthmetrics/en/, accessed 10 January 2009).
3. Speybroeck N et al. Reassessing the relationship 20. Diallo K et al. Monitoring and evaluation of human
between human resources for health, intervention resources for health: an international perspective.
coverage and health outcomes. Background Human Resources for Health, 2003, 1:3 (http://
paper prepared for The World Health Report www.human-resources-health.com/content/1/1/3,
2006. Geneva, World Health Organization, 2006 accessed 10 January 2009).
(http://www.who.int/hrh/documents/reassessing_
relationship.pdf, accessed 10 January 2009). 21. Stansfield SK et al. Information to improve decision-
making for health. In: Jamison DT et al., eds.
4. The world health report 2006: working together for Disease control priorities for the developing world.
health. Geneva, World Health Organization, 2006 Washington, DC, World Bank and Oxford University
(http://www.who.int/whr/2006, accessed 10 January Press, 2006.
2009).
22. Sullivan TM, Strachan S, Timmons BK. Guide
5. Deloitte Insight Economics. Impact monitoring and to monitoring and evaluating health information
evaluation framework: background and assessment products and services. Baltimore, MD, Johns
approaches. Barton, Cooperative Research Centres Hopkins Bloomberg School of Public Health,
Association of Australia, 2007. Constella Futures and Management Sciences for
6. Glossary of statistical terms. Paris, Organisation for Health, 2007.
Economic Co-operation and Development (http:// 23. Health Metrics Network. Framework and standards
stats.oecd.org/glossary/index.htm, accessed 10 for country health information systems, 2nd ed.
January 2009). Geneva, World Health Organization, 2008 (http://
7. Toolkit for monitoring health systems strengthening. www.who.int/healthmetrics/documents/framework/
Draft version. Geneva, World Health Organization, en/index.html, accessed 10 January 2009).
2008 (http://www.who.int/healthinfo/statistics/toolkit_ 24. Policy briefs. Briefs to complement The world health
hss/en/, accessed 10 January 2009). report 2006: working together for health. Geneva,
8. Health and the Millennium Development Goals. World Health Organization, 2006 (http://www.
Geneva, World Health Organization (http://www.who. who.int/hrh/documents/policy_brief, accessed 10
int/mdg, accessed 10 January 2009). January 2009).
9. Global Fund to Fight AIDS, Tuberculosis and Malaria 25. HRH Action Framework. Capacity Project (http://
(http://www.theglobalfund.org/en/, accessed 10 www.capacityproject.org/framework, accessed 10
January 2009). January 2009).
10. GAVI Alliance (http://www.gavialliance.org/, 26. Nyoni J, Gbary A, Awases M et al. Policies and
accessed 10 January 2009). plans for human resources for health: guidelines for
countries in the WHO African region. Brazzaville,
11. United States President’s Emergency Plan for AIDS WHO Regional Office for Africa, 2006 (http://www.
Relief (http://www.pepfar.gov, accessed 10 January afro.who.int/hrh-observatory/documentcentre/
2009). policies_plans_guidelines.pdf, accessed 10 January
12. High Level Forum on the Health MDGs (http://www. 2009).
hlfhealthmdgs.org/, accessed 10 January 2009). 27. Tools and guidelines for human resources for health
13. Joint Learning Initiative. Human resources for health: planning. Geneva, World Health Organization (http://
overcoming the crisis. Boston, Harvard University’s www.who.int/hrh/tools/planning/en/index.html,
Global Equity Initiative, 2004. accessed 10 January 2009).
14. WHO resolutions on health workforce development. 28. The Kampala Declaration and Agenda for Global
Geneva, World Health Organization (http://www.who. Action. Declaration adopted at the First Global
int/hrh/resolutions/en/, accessed 10 January 2009). Forum on Human Resources for Health, 2–7 March
2008, Kampala. Geneva, World Health Organization
15. Global Health Workforce Alliance (http://www.who.int/ and Global Health Workforce Alliance, 2008 (http://
workforcealliance/en/, accessed 10 January 2009). www.who.int/workforcealliance/Kampala%20
16. Asia-Pacific Action Alliance on Human Resources Declaration%20and%20Agenda%20web%20
for Health (http://aaahrh.org/, accessed 10 January file.%20FINAL.pdf, accessed 10 January 2009).
2009).
17. International Health Partnership (http://www.
internationalhealthpartnership.net/, accessed 10
January 2009).
12
Boundaries of the health workforce:
2 definition and classifications of

health workers
DAVID HUNTER, MARIO R DAL POZ, TEENA KUNJUMEN
2.1 Introduction informal caregivers and volunteers who contribute

to the improvement of health should also be counted
The health workforce represents one of the key build- as part of the health workforce. But in practical terms,
ing blocks of health systems and has been identified these are not often counted, due to lack of information
as a priority for action for strengthening those sys- on the unpaid workforce and the ensuing difficulty with
tems (1). However, international assessments of human regard to establishing the boundaries of what consti-
resources or other non-monetary inputs to health sys- tutes a health system.
tems tend to be less widespread than comparisons of
health-care expenditures (2). This is in part due to lack Even then, the definition of a health action for classi-
of a common framework and adequate data for compar- fying paid workers is not straightforward. Consider a
ative health workforce analysis. Imprecise professional nurse employed by a manufacturing company to pro-
boundaries and differences in defining and categoriz- vide on-site health-care services for its employees:
ing certain types of health workers across much of the the main goal of the actions of the nurse is to improve
world present further challenges in analysing health health, although the main goal of the actions of the
workforce data. For instance, a “nurse” in one country employer is not. Then take the case of a gardener
may be characterized by different educational require- employed by a hospital: the gardener’s own actions
ments, legislation and practice regulations, skills and do not directly improve health, although the actions of
scope of practice than a “nurse” elsewhere. In order to the employer – the hospital – do. There are many such
monitor trends in the health workforce situation across non-clinical workers in health industries, such as man-
countries or over time, or for countries to share experi- agers, computer operators, clerks and trades workers,
ences and best practices, it is necessary to know how who provide managerial and infrastructural support.
health workers are defined and classified in the original There are also many skilled health-care providers who
information source (3). work outside facility-based service delivery points,
including those in government ministries and depart-
This chapter discusses the definition of the health ments, public health offices, health and health systems
workforce and its operationalization. Current uses of research agencies, health professions education and
internationally standardized classifications for statisti- training institutions, company and school-based clin-
cal delineation, description and analysis of the health ics, residential care settings, rehabilitation centres,
workforce are reviewed, and some options for future correctional facilities, military service and others. A
consideration are outlined. classification system that considers the actions of the
individual alone, or those of the place of work alone,
may fail to capture them all in the health workforce.
2.2 Who are health workers?
In order to provide comparable and consistent data to
Any health workforce analysis requires precise defini-
inform decision-making, it is necessary to define the
tion of health workers. The World Health Organization
health workforce operationally. While there is no sin-
(WHO) defines the health workforce as “all people
gle measure of the health workforce, it is important to
engaged in actions whose primary intent is to enhance
specify which elements of the definition and classifi-
health” (4). This statement reinforces the WHO concept
cation structure are being considered. For example, if
of health systems as comprising “all organizations,
one study includes the above-mentioned nurse work-
people and actions whose primary intent is to promote,
ing for a private company while another does not,
restore or maintain health” (1). This infers, for example,
then the comparability of data from the two studies is
that family members looking after the sick and other
compromised.
13
Table 2.1 Framework for defining the health workforce
Individual’s training, occupation & Working in the health industry Working in a non-health industry or
place of work unemployed/inactive
Training in health and employed in a A. For example, physicians, nurses, C. For example, nurses working for
health occupation midwives working in health-care private companies, pharmacists
facilities working at retail outlets
Training in health but not employed in A. For example, medically trained C. For example, medically trained
a health occupation managers of health-care facilities university lecturers, unemployed nurses
Training in a non-health field or no B. For example, economists, clerks, D. For example, primary school teachers,
formal training gardeners working in health-care garage mechanics, bank accountants
facilities
Table 2.1 provides a useful framework to capture health public expenditure, staffing and payroll, professional
workers employed (or not employed) in the health and training, registration and licensure). Health facility
non-health industries. Three categories of workers rele- assessments or payroll records will only provide data
vant for health workforce analysis can be distinguished: for categories A and B, while data from professional
regulatory associations tend to be limited to A and C. In
A. those with health vocational education and training contrast, nationally representative population censuses
working in the health services industry; and labour force surveys with properly designed ques-
B. those with training in a non-health field (or with tions on occupation, place of work and field of training
no formal training) working in the health services can provide information on all three components.
industry;
C. those with health training who are either working in
a non-health-care–related industry, or who are cur- 2.3 Health workforce classification
rently unemployed or not active in the labour market.
Health workers play different roles and often have dif-
Categories A and C together form the trained (skilled) ferent national history, culture and codes of practice.
health workforce (active or inactive) available in a given Any attempt to compare the size and characteris-
country or region, while A and B represent the work- tics of the health workforce across countries or over
force employed in the health industry. The sum of the time requires some level of harmonization of the
three elements A, B and C provides the total potential available information. In order to compare and inte-
health workforce available. A fourth category, D, encap- grate data from different sources and countries, it is
sulates all non-health workers, that is, those workers necessary to use internationally consistent or harmo-
without training for a health occupation and not work- nized classification systems. Although some countries
ing in the health industry. disseminate data using national educational, occupa-
tional or industrial classifications that are not always
The advantage of this framework lies in the fact that it comparable, most use classification systems that are
integrates the elements of training, current occupation either based on or linked to internationally standard-
and industry. In this context, “training” refers to the (for- ized classifications, such as the International Standard
mal and informal) education undertaken by individuals to Classification of Education (ISCED), the International
equip them with the skills necessary to perform the tasks Standard Classification of Occupations (ISCO) and
required for competent performance in a job, “occupa- the International Standard Industrial Classification of
tion” refers to the tasks and duties performed in a job by All Economic Activities (ISIC). These classifications
individuals, and “industry” refers to the activities of the provide a coherent framework for categorizing fields
establishments or enterprises in which individuals are and levels of training, occupations and industries of
employed. Considering all three elements is essential to employment, respectively, according to shared char-
gain an understanding of the dynamics of the workforce. acteristics (5–7).
The framework can be a useful tool for identifying poten-

2.3.1 Classification of
tial data sources and gaps for health workforce analysis.
education and training
A number of sources can be used to provide informa-
tion and evidence to inform policy – notably, population There are important challenges in clearly identifying the
censuses and surveys, health facility assessments and different types of education and training programmes for
routine administrative records (including records on health workers from different institutions, having different
14
Boundaries of the health workforce: definition and classifications of health workers
Table 2.2 Relevant levels of education and training for health occupations according to the
International Standard Classification of Education (ISCED-1997)
Level Name Description Typical duration Complementary dimensions

3 Upper secondary Typically begins at the end of Typically requires the Considerations for classifying
education full-time compulsory education completion of some this level include type of
for those countries that have a 9 years of full-time subsequent education or
system of compulsory education. education since the destination; programme
beginning of basic orientation; and cumulative
(primary) education. theoretical duration.
Level 3 corresponds to the

typical minimum entrance
requirement for education and
training for a health occupation.
4 Post-secondary Programmes that straddle Typical full-time May include three programme
non-tertiary the boundary between upper equivalent duration of orientations: (i) general
education secondary and post-secondary between 6 months and education; (ii) prevocational
education from an international 2 years. or pretechnical education; and
point of view, even though they (iii) vocational or technical
might clearly be considered education.
as upper secondary or post-
secondary programmes in a Includes adult education (for
national context. example, technical courses
given during an individual’s
professional life on specific
subjects).
5 First stage of Tertiary programmes having an Typical full-time Programmes usually giving
tertiary education advanced educational content equivalent duration access to occupations with high
(but not leading directly to the of at least 2 years, skills requirements.
award of an advanced research although some
qualification). programmes are of 4 Includes programmes leading to
or more years. a master’s degree.
6 Second stage of Tertiary programmes leading Very restricted scope at this

tertiary education to the award of an advanced level.
research qualification (i.e. entails
advanced study and original
research, not only coursework).
Source: United Nations Educational, Scientific and Cultural Organization (8).
entrance criteria, curricula and durations of training, and The latest version of ISCED (referred to as ISCED-97)
then grouping them into categories that are nationally classifies seven educational levels, and nine broad
and internationally comparable. Comparability can be fields (in other words, at the one-digit coding level)
enhanced through the collection, processing and dis- and 25 subfields (two-digit level). Table 2.2 shows the
semination of data following the ISCED standard, which ISCED educational levels relevant for education and
provides a framework for the compilation and presen- training leading to a health occupation. Most relevant
tation of national and international education statistics specializations fall by level under subfield 72, “health”,
and indicators for policy analysis and decision-making, including education in medicine, medical and health
whatever the structure of the national education systems services, nursing and dental services.
and whatever the stage of economic development of a
country (8). ISCED covers all organized and sustained Certain tools also exist that aim at providing guidelines
learning activities for children, young people and adults. on how to apply the ISCED classification. In one such
It allows a variety of types of education programmes to manual, the two-digit fields of education from ISCED-
be classified by level and field of education, such as ini- 97 are expanded to the three-digit level, capturing more
tial formal education, continuing education, non-formal details for vocational education and training while still
education, distance education, apprenticeships, techni- ensuring cross-national comparability (9). The man-
cal-vocational education and special needs education. ual is intended to serve as a guide in countries where
15
comprehensive national classifications are not devel- Most health occupations can be identified at the two-
oped, based on an analysis of the descriptions of the digit or three-digit levels of ISCO-08. However, a
content of training programmes. Table 2.3 presents the four-digit code is needed to distinguish practitioner
three-digit details for health-related specializations. specializations (such as dentists from pharmacists),
and also to separately identify some other allied health
workers, such as psychologists and social work pro-
2.3.2 Classification of occupations
fessionals, classified in the same minor group as other
Another useful classification system for health workforce social scientists, including economists and sociologists.
analysis is ISCO, developed by the International Labour
Organization (10). This system of classification enables A significant improvement in ISCO-08 for the purposes
jobs to be arranged into a hierarchical system specified of health workforce analysis is the creation of addi-
according to the precision needed, in major (one-digit tional unit groups, notably for distinguishing generalist
level), sub-major (two-digit level), minor (three-digit from specialist medical doctors, and for identifying a
level) and unit (four-digit level) groups. The basic cri- number of types of allied health workers. For example,
teria used to define the grouping system are the “skill a growing number of countries, especially low-income
level” and “skill specialization” required to carry out the countries with critical shortages of highly skilled
tasks and duties of the occupations (6). Skill level refers medical and nursing professionals, are turning to
to the complexity and range of tasks required for the “community health workers” – community health aides
job. Skill specialization is related to the field of knowl- selected, trained and working in the communities from
edge required, tools and machinery used, the materials which they come – to render certain basic health serv-
worked on or with and the goods or services produced. ices (12). This category is specified in ISCO-08 (unit
group code 3253).
In the most recent version of ISCO, revised in 2008
(known as ISCO-08), the main occupations of inter- Another improvement is in the treatment of veterinary
est with health care-related specialization fall within occupations. In ISCO-88, veterinary occupations were
two sub-major groups: sub-major group 22, “health found in the same minor groups as human health occu-
professionals” (generally well-trained workers in jobs pations; therefore coding was needed at the four-digit
that normally require a university degree for competent level to distinguish veterinarians and veterinary assist-
performance); and sub-major group 32, “health associ- ants. However, in ISCO-08, these two occupational
ate professionals” (generally requiring knowledge and groups are now classified in separate minor groups.
skills acquired through advanced formal education As a result, data disseminated at even the three-digit
and training but not equivalent to a university degree). level will allow human health occupations to be dis-
Health professionals include medical doctors (an occu- tinguished from veterinary occupations at both the
pational title used interchangeably with “physicians” in professional and associate professional levels.
this Handbook), nursing and midwifery professionals,
and others such as dentists and pharmacists. Health Other occupations of interest include managers and
associate professionals include medical and pharma- personal care workers in health services. In ISCO-
ceutical technicians, nursing and midwifery associate 08, managers of health and aged care services are
professionals and others such as dental assistants, separately identified at the four-digit level (unit group
physiotherapy technicians and dispensing opticians codes 1342 and 1343, respectively). “Personal care
(Table 2.4, page 18). workers in health services” are identified in a separate
minor group (code 532), which includes unit groups
For the earlier version of ISCO (adopted in 1988, or for health-care assistants and home-based personal
ISCO-88) – against which currently available data care workers. Information classified according to the
were being classified at the time of publication of previous ISCO-88 was required at the most detailed
this Handbook – the relevant information needed to four-digit level in order to differentiate childcare workers
be coded to a degree of detail that minimally corre- from personal care workers in health services.
sponded to the three-digit level in order for health
occupations to be properly identified. Information at the Certain documents produced by the International
two-digit level did not allow distinction of health occu- Labour Organization guide countries on how to develop,
pations from other life sciences occupations. However, maintain and revise a national occupation classification
drawing on consultations between the International and its mapping to the international standard (13, 14).
Labour Organization, WHO and other stakeholders, the The most up-to-date information and latest advice can
newly adopted 2008 version stemmed from the recog- be found on the ISCO web site (10).
nition that the previous version was outdated in some
areas (11).
16
Table 2.3 Fields of vocational training related to health according to Fields of training manual
Code Field Description

Fields of training directly related to health
721 Medicine The study of the principles and procedures used in preventing, diagnosing, caring for
and treating illness, disease and injury in humans and the maintenance of general health.
Principally, this field consists of training of physicians.
Programmes with the following main content are classified here: medicine, medical
science, medical training. Medical specializations are included here, such as
anaesthesiology, anatomy, cardiology, dermatology, epidemiology, forensic medicine,
gerontology, haematology, internal medicine, neurology, obstetrics and gynaecology,
oncology, ophthalmology, paediatrics, preventive and social medicine, psychiatry, surgery.
722 Medical services The study of physical disorders, treating diseases and maintaining the physical well-being
of humans, using non-surgical procedures.
Training programmes classified here comprise a wide range of services such as ambulance
service, chiropractic, hearing aid technology, medical laboratory technology, medical
X-ray techniques, nutrition and dietetics, occupational therapy, optometry, orthopaedic
prosthetics, emergency paramedical technologies, pharmacy, physiotherapy, radiotherapy,
speech pathology and therapy, vocational rehabilitation.
723 Nursing The study of providing health care for the sick, disabled or infirm and assisting physicians
and other medical and health professionals diagnose and treat patients.
Training programmes with the following main content are classified here: assistant
nursing, basic nursing, care of old people, care of the disabled, infant hygiene (nursing),
midwifery, nursing aide/orderly, psychiatric nursing.
724 Dental studies The study of diagnosing, treating and preventing diseases and abnormalities of the teeth
and gums. It includes the study of designing, making and repairing dental prostheses and
orthodontic appliances. It also includes the study of providing assistance to dentists.
Training programmes with the following main content are classified here: clinical dentistry,
dental assisting, dental hygiene, dental laboratory technology, dental nursing, dental
science, dental surgery, odontology, orthodontics.
Fields of training associated with health
762 Social work and The study of the welfare needs of communities, specific groups and individuals and the
counselling appropriate ways of meeting these needs. The focus is on social welfare with emphasis on
social policy and practice.
Among the training programmes classified here: alcohol and drug abuse counselling, crisis
support, social practice, social work (welfare).
850 Environmental The study of the relationships between living organisms and the environment in order
protection to protect a wide range of natural resources. Programmes in services to the community
dealing with items that affect public health, such as hygiene standards in food and water
supply, are included here.
Among the training programmes classified here: air pollution control, community
sanitation, environmental toxicology, garbage disposal, water pollution control.
862 Occupational health The study of recognizing, evaluating and controlling environmental factors associated with
and safety the workplace.
Among the training classified here: ergonomics (occupational health and safety), health
and safety in the workplace, labour welfare (safety), occupational health and industrial
hygiene, stress at work.
Source: Adapted from European Centre for the Development of Vocational Training and Eurostat (9).
17
Table 2.4 Occupational titles related to health according to the International Standard Classification of Occupations (ISCO), 1988 and 2008 revisions
18
ISCO-2008 ISCO-1988
Group code Occupational title Group code Occupational title

Sub- Minor Unit Sub- Minor Unit
major major
22 Health professionals 22 Life science and health professionals
221 Medical doctors 222 Health professionals (except nursing)
2211 Generalist medical practitioners 2221 Medical doctors
2212 Specialist medical practitioners
222 Nursing and midwifery professionals 223 Nursing and midwifery professionals
2221 Nursing professionals 2230 Nursing and midwifery professionals
2222 Midwifery professionals
223 Traditional and complementary medicine professionals
2230 Traditional and complementary medicine professionals
224 Paramedical practitioners
2240 Paramedical practitioners
226 Other health professionals 222 Health professionals (except nursing)
2261 Dentists 2222 Dentists

2262 Pharmacists 2224 Pharmacists
2263 Environmental and occupational health and hygiene professionals
2264 Physiotherapists
2265 Dieticians and nutritionists
2266 Audiologists and speech therapists
2267 Optometrists and ophthalmic opticians
2269 Health professionals n.e.c. 2229 Health professionals (except nursing) n.e.c.
32 Health associate professionals 31 Physical and engineering science associate professionals
321 Medical and pharmaceutical technicians 313 Optical and electronic equipment operators
3211 Medical imaging and therapeutic equipment technicians 3133 Medical equipment operators
3212 Medical and pathology laboratory technicians
32 Life science and health associate professionals
322 Modern health associate professionals (except nursing)
3213 Pharmaceutical technicians and assistants 3228 Pharmaceutical assistants
3214 Medical and dental prosthetic and related technicians
Continues…
ISCO-2008 ISCO-1988
Group code Occupational title Group code Occupational title

Sub- Minor Unit Sub- Minor Unit
major major
322 Nursing and midwifery associate professionals 323 Nursing and midwifery associate professionals
3221 Nursing associate professionals 3231 Nursing associate professionals
3222 Midwifery associate professionals 3232 Midwifery associate professionals
323 Traditional and complementary medicine associate 324 Traditional medicine practitioners and faith healers
professionals
3230 Traditional and complementary medicine associate professionals 3241 Traditional medicine practitioners
325 Other health associate professionals 322 Modern health associate professionals (except nursing)
3251 Dental assistants and therapists 3225 Dental assistants
3252 Medical records and health information technicians
3253 Community health workers
3254 Dispensing opticians 3224 Optometrists and opticians
3255 Physiotherapy technicians and assistants 3226 Physiotherapists and related associate professionals
3256 Medical assistants 3221 Medical assistants
3257 Environmental and occupational health inspectors and associates 3222 Sanitarians
3258 Ambulance workers
3259 Health associate professionals n.e.c. 3229 Modern health associate professionals (except nursing) n.e.c.
53 Personal care workers 51 Personal and protective workers
532 Personal care workers in health services 513 Personal care and related workers
5321 Health care assistants 5132 Institution-based personal care workers
5322 Home-based personal care workers 5133 Home-based personal care workers
5329 Personal care workers in health services n.e.c.
Additional health-related unit groups Additional health-related unit groups
1342 Health service managers
1343 Aged care service managers
2634 Psychologists 2445 Psychologists
2635 Social work and counselling professionals 2446 Social work professionals
3344 Medical secretaries
Notes: This table presents an overview of the treatment of health occupations in the 1988 and 2008 versions of ISCO and should not be used for correspondence. Occupations related to the veterinary
field are excluded from the thematic view.
n.e.c. = not elsewhere classified.
19
Source: International Labour Organization (10).

Table 2.5 Economic sectors related to health activities on the basis of the International Standard
Industrial Classification of All Economic Activities (ISIC) Revision 4
Code Economic activity
Section Division Group Class

Core health industry groups and classes
Q Human health and social work activities
86 Human health activities
861 8610 Hospital activities
862 8620 Medical and dental practice activities
869 8690 Other human health activities
Selected associated classes
C 21 210 2100 Manufacture of pharmaceuticals, medicinal chemical and botanical products
32 325 3250 Manufacture of medical and dental instruments and supplies
E 36 360 3600 Water collection, treatment and supply
37 370 3700 Sewerage
G 47 477 4772 Retail sale of pharmaceutical and medical goods, cosmetic and toilet articles
in specialized stores
K 65 651 6512 Non-life insurance (including provision of health insurance)
M 71 712 7120 Technical testing and analysis (include testing in the field of food hygiene;
testing and measuring air and water pollution)
O 84 841 8412 Regulation of the activities of providing health care, education, cultural
services and other social services
8430 Compulsory social security activities (including funding and administration
of government-provided social security programmes for sickness, work-
accident, temporary disablement, etc.)
Q 87 871 8710 Residential nursing care facilities
872 8720 Residential care activities for mental retardation, mental health and
substance abuse
88 881 8810 Social work activities for the elderly and disabled (without accommodation)
Source: United Nations Statistics Division (15).
2.3.3 Classification of branches levels of classification in a four-level hierarchy: sections

of economic activity (one-letter code), divisions (two-digit code), groups
(three-digit code) and classes (four-digit code). This
As mentioned earlier, in addition to health-care service
grouping is done according to similarities in the char-
providers, there are many non-health-trained work-
acter of the goods and services produced, the uses to
ers acting to keep health institutions functioning. It is
which the goods and services are put, and the inputs,
estimated that about one third of the global health work-
process and technology of production.
force is composed of health management and support
workers (4). Capturing them requires consideration of
Relevant information for health workforce analysis
occupations across almost all ISCO groups, so further
essentially falls under ISIC division 86, “human health
information on place of work may be required. The ISIC
activities”. Data available at the group (three-digit) or
classification can form a basis for such analysis, as it
class (four-digit) level will allow disaggregating the
allows pooling of information on workers in health serv-
different types of health systems activities, including
ices across different types of economic systems within
service provision, supplies procurement and financing.
a comparative framework (15). In ISIC, economic pro-
Table 2.5 presents selected health-related categories
ducing units are grouped into successively broader
from the latest ISIC revision.
20
2.3.4 Other classifications ask each active person for both the occupational title
It is also of significance to countries and stakehold- and a brief description of the main tasks and duties
ers to be able to distinguish the different categories of performed on the job. It is expected that possibilities for
human resources within health systems, such as those health workforce analyses will be strengthened in the
who are regular employees of the systems and those current global series of censuses, known as the 2010
who are not, or those whose basic salaries are drawn round (covering the period 2005 to 2014), which will
from the government budget in comparison with health largely be able to exploit the new ISCO-08 revision.
workers who are funded by other sources. A full list of
For some countries, human resources for health anal-
international classifications for the collection and dis-
yses based on population census and survey data
semination of economic and social statistics is available
can be facilitated through collaborative research
at the United Nations Statistics Division web site (16).
projects aiming to harmonize microdata variables
and structures for public use. Key microdata provid-
ers include the Integrated Public Use Microdata Series
2.4 Summary and conclusions (21), the African Census Analysis Project (22) and the
Comparative health workforce analysis is meaningful Luxembourg Income Study (23). Such projects pro-
only when the available information is based on com- cess census and survey microdata series for multiple
mon definition and classification of health workers. countries – with education, occupation and indus-
There is no single operational boundary of what con- try variables mapped where possible to ISCED, ISCO
stitutes the health workforce. Many assessments use and ISIC, respectively – and help disseminate the rele-
country-specific or even tool-specific definitions and vant documentation for scholarly and policy research.
titles that are not always comparable across countries Chapter 8 of this Handbook presents a multicountry
or over time. However, a growing number of coun- analysis of health workforce statistics making use of the
tries are disseminating health labour data that can Integrated Public Use Microdata Series (21). The analy-
be mapped to international standard classifications sis draws on available occupational data from the 2000
– such as the International Standard Classification of round of censuses mapped to ISCO-88.
Education, the International Standard Classification of
Occupations, and the International Standard Industrial Even with ongoing improvement and revisions, given
Classification of All Economic Activities. These classifi- their nature, standardized classifications are inherently
cations provide a coherent framework for categorizing generalized and attempt to simplify a very complex
key workforce variables (vocational training, occupation system for statistical purposes. They may not always
and industry of employment, respectively) according to capture the full complexity and dynamics of the
shared characteristics. Using this trichotomy allows the health labour market. The World Health Organization,
identification of people with training in health, of people International Labour Organization and other partners
employed in health-related occupations, and of people are continually engaged in initiatives to improve inter-
employed in health services industries. national classifications relevant for health workforce
analysis and promote their use. This includes ongoing
Health workforce analyses can draw on data from enumeration of the various sources of data and types
a number of sources, including standard statistical of classifications used for monitoring health workers
sources outside the (traditional) health sector. Selected (7). This may facilitate definitional harmonization of the
tools for guiding the collection and coding of statisti- health workforce within and across countries, and be
cal information on economic activity from population used to develop a road map on how to improve health
censuses and surveys can be found online at the hand- workforce classifications at the national and interna-
books, guidelines and training manuals section of the tional levels. Such exercises continue to benefit from
United Nations Statistics Division web site (17) (see exchanges and interactions among those that produce
also 18, 19). The United Nations Statistics Division (20) and use this information from diverse perspectives,
recommends the collection and processing of census including national governments (ministries of health,
data on education, occupation and industry catego- labour and education, and central statistical offices),
rized in accordance with, or in a manner convertible to, health professional associations, WHO regional and
the latest revision available of the relevant international country offices, other international bodies with health
classification (i.e. ISCED, ISCO and ISIC, respec- and statistical interests, nongovernmental and private
tively). It is further recommended that countries code organizations working in health and statistics, and aca-
the collected responses at the lowest possible level of demic and research institutions.
classification detail supported by the information given.
In particular, in order to facilitate detailed and accurate
coding for occupation data, the questionnaire should
21
References 12. Lehmann U, Sanders D. Community health workers:

what do we know about them? Follow-up paper
1. Everybody’s business: strengthening health systems to The world health report 2006. Geneva, World
to improve health outcomes – WHO’s framework Health Organization, 2007 (http://www.who.int/
for action. Geneva, World Health Organization, hrh/documents/community_health_workers.pdf,
2007 (http://www.who.int/healthsystems/strategy/ accessed 11 January 2009).
everybodys_business.pdf, accessed 10 January
2009). 13. Methodological issues concerning the development,
use, maintenance and revision of statistical classifi-
2. Anell A, Willis M. International comparison of health cations. Geneva, International Labour Organization,
care systems using resource profiles. Bulletin of the 2004 (http://www.ilo.org/public/english/bureau/stat/
World Health Organization, 2000, 78(6):770–778 isco/docs/intro5.htm, accessed 11 January 2009).
(http://www.who.int/bulletin/archives/78(6)770.pdf,
accessed 10 January 2009). 14. Embury B. Constructing a map of the world of
work: how to develop the structure and contents of
3. Dal Poz MR et al. Relaciones laborales en el sector a national standard classification of occupations.
salud: fuentes de informacion y metodos de analisis, STAT Working Paper No. 95–2. Geneva, International
v. 1. Quito, Organizacion Panamericana de la Salud, Labour Office, 1997 (http://www.ilo.org/public/
2000. english/bureau/stat/download/papers/map.pdf,
4. The world health report 2006: working together for accessed 11 January 2009).
health. Geneva, World Health Organization, 2006 15. International Standard Industrial Classification of All
(http://www.who.int/whr/2006, accessed 10 January Economic Activities, fourth revision. Statistical Papers
2009). Series M, No. 4/Rev.4. New York, United Nations
5. Diallo K et al. Monitoring and evaluation of human Statistics Division, 2008 (http://unstats.un.org/unsd/
resources for health: an international perspective. demographic/sources/census/2010_PHC/docs/
Human resources for health, 2003, 1:3 (http:// ISIC_rev4.pdf, accessed 11 January 2009).
www.human-resources-health.com/content/1/1/3, 16. List of international family of economic and social
accessed 10 January 2009). classifications. United Nations Statistics Division
6. Hoffmann E. International statistical comparisons (http://unstats.un.org/unsd/class/family/family1.asp,
of occupational and social structures: problems, accessed 11 January 2009).
possibilities and the role of ISCO-88. In: Hoffmeyer- 17. Handbooks, guidelines and training manuals.
Zlotnik JHP, Wolf C, eds. Advances in cross-national United Nations Statistics Division (http://unstats.
comparison. New York, Kluwer Plenum Publishers, un.org/unsd/demographic/standmeth/handbooks/,
2003. accessed 11 January 2009).
7. Dal Poz MR et al. Counting health workers: 18. Hussmanns R, Mehran F, Verma V. Surveys of
definitions, data, methods and global results. economically active population, employment,
Background paper prepared for The world health unemployment and underemployment: an ILO
report 2006. Geneva, World Health Organization, manual on concepts and methods. Geneva,
2006 (http://www.who.int/hrh/documents/counting_ International Labour Office, 1990.
health_workers.pdf, accessed 10 January 2009).
19. Handbook on measuring the economically active
8. International Standard Classification of Education: population and related characteristics in population
ISCED 1997. Paris, United Nations Educational, censuses. Studies in Methods Series F, No. 102.
Scientific and Cultural Organization, 1997 (http:// New York, United Nations and International Labour
www.uis.unesco.org/TEMPLATE/pdf/isced/ Organization, 2009 (http://unstats.un.org/unsd/
ISCED_A.pdf, accessed 10 January 2009). demographic/sources/census/Entire%20Handbook.
9. Fields of training: manual. Thessaloniki, European pdf, accessed 19 May 2009).
Centre for the Development of Vocational Training 20. United Nations Statistics Division. Principles and
and Eurostat, 1999 (http://www.trainingvillage.gr/etv/ recommendations for population and housing
Upload/Information_resources/Bookshop/31/5092_ censuses, revision 2. Statistical Papers Series M, No.
en.pdf, accessed 10 January 2009). 67/Rev. 2. New York, United Nations, 2008 (http://
10. International Standard Classification of Occupations. unstats.un.org/unsd/demographic/sources/census/
International Labour Organization (http://www.ilo.org/ docs/P&R_Rev2.pdf, accessed 11 January 2009).
public/english/bureau/stat/isco/index.htm, accessed 21. Integrated Public Use Microdata Series. Minnesota
11 January 2009). Population Center (http://www.ipums.umn.edu/,
11. Options for the classification of health occupations accessed 11 January 2009).
in the updated International Standard Classification 22. African Census Analysis Project. University of
of Occupations (ISCO-08). Background paper for Pennsylvania (http://www.acap.upenn.edu,
the work to update ISCO-88. Geneva, International accessed 20 January 2009).
Labour Organization, 2006.
23. Luxembourg Income Study (http://www.lisproject.
org/, accessed 11 January 2009).
22
Part II:
MONITORING THE
STAGES OF THE
WORKING LIFESPAN
23
Monitoring the active health
3 workforce: indicators, data sources

and illustrative analysis
FELIX RIGOLI, BOB POND, NEERU GUPTA, CHRISTOPHER H HERBST
3.1 Introduction Various permutations and combinations of what consti-

tutes the health workforce potentially exist, depending
Human resources for health (HRH) have long been on each country’s situation and the means of monitor-
recognized as “the cornerstone of the [health] sec- ing. A framework for harmonizing the boundaries and
tor to produce, deliver, and manage services” (1). constituency of the health workforce across contexts is
Assessments of HRH are required for various pur- presented in Chapter 2 of this Handbook. To facilitate
poses, notably for planning, implementing, monitoring data collection and analysis processes, it is important
and evaluating health sector strategies, programmes to focus on a limited and essential number of indica-
and interventions. The importance of sound empir- tors that are comparable and measurable regularly
ical evidence for informed policy decision-making using standard data sources (2). Such data sources
and monitoring of progress in strengthening health include population-based sources (censuses and sur-
workforce development and management is widely veys), health facility assessments and administrative
recognized. Precisely describing HRH can help to records. For specialized or in-depth HRH assessments,
identify opportunities and constraints for scaling up information can further be drawn from, for example, pro-
health interventions. fessional registries, national health accounts, records
of health education and training institutes, and qualita-
The size and distribution of the health workforce is
tive studies.
the result of the inflow into, outflow from and circu-
lation of workers between, for example, different The development of a comprehensive evidence base
sectors (public or private), industries (health services generally requires combining different types of infor-
or other), regions (rural or urban), countries and sta- mation that may exist, frequently scattered across
tuses (employed, unemployed or inactive) (Figure 3.1). different sources. This chapter focuses on describing
Figure 3.1 Stocks and flows of the health workforce
Labour force activity

Pre-entry to Exit from
labour force labour force
Health sector
Training in health- Rural Urban
related field Migration
Public Private
Salaried Self-employed Retirement
Training in Full time Part time
non-health field
Patient care Administrative
or other tasks Work-limiting
disability of death
Non-health sector
No formal Other reasons
training Health-related Non-health- (e.g. family care)
tasks related tasks
Unemployed
25
the tools and means to monitor the active health work- the assessment), persons having been educated in a
force, that is, all people currently participating in the health-related field regardless of place of employment,
health labour market. Core indicators for characterizing or persons having been educated in a health-related
HRH are first identified, with an emphasis on optimiz- field regardless of current labour force status.
ing comparability across countries and over time. Key
potential sources of data are then reviewed; both pri- Measuring the skills mix of the health workforce offers
mary sources and standard statistical sources are a means to assess the combination of categories
examined, and the opportunities and challenges they of personnel at a specific time and identify possible
offer for health workforce analysis are considered. imbalances related to a disparity in the numbers of vari-
Illustrative examples are presented, using case stud- ous health occupations. Statistics on skills mix can help
ies from various countries and sources. Lastly, some inform strategies to ensure the most appropriate and
lessons learnt and recommendations for strengthening cost-effective combination of roles and staff. Because
HRH information and monitoring systems within coun- counts of workers in the private sector are likely to be
tries are discussed. The present chapter is primarily less accurate when drawing on administrative data
devoted to monitoring current health workforce activity; sources than counts of those in the public sector,
measurements of entry (notably, pre-service education and because private for-profit providers are often less
and training) and exit (attrition due to various factors, accessible to low-income populations, it is also rec-
including migration, retirement and death) are the focus ommended that indicators be used to monitor workers’
of the next two chapters in this Handbook, respectively. employment sector (public, private for-profit or private
not-for-profit).
As detailed in the previous chapter, comprehensive

3.2 Core indicators for
assessments require accurate information on occupa-
HRH analysis: what needs tion, industry and training. Drawing upon a combination
to be monitored? of these types of information will enable the identi-
fication of, for example, employment in non-health
Effective monitoring and evaluation of HRH in coun-
activities among those with a health-related educa-
tries requires agreement upon a core set of indicators
tion, and employment in health activities with jobs that
at the subnational, national and international levels to
do not require clinical skills (see Chapter 2). Additional
inform decision-making among national authorities
indicators on labour productivity, unemployment and
and other stakeholders. Ideally, the indicators retained
underemployment, and emigration, for instance, will
should be characterized by “SMART” properties: spe-
allow monitoring of workforce wastage, or excess loss
cific (measures exactly the result); measurable (so
in utility due to attrition or poor productivity that could
that the result can be tracked); attainable (so that the
have been prevented or managed (4). Health workforce
result can be compared against a realistic target); rele-
metrics, or measurements of particular characteris-
vant (to the intended result); and timebound (indicates
tics of performance or efficiency of HRH development
a specific time period). The ongoing and consistent
strategies, can further be assessed by means of indi-
measurement of these indicators allows monitoring of
cators on HRH renewal and migration (2).
how HRH-related programmes and policies are being
implemented. Once the baseline data have been gen-
Comparability of HRH statistics across countries and
erated, an evaluation framework can be established
over time can be enhanced through the setting and
with periodic targets for analysis in terms of change
use of common definitions and classifications for mon-
and progress over time, that is, whether activities have
itoring the labour market. This includes the collection,
been implemented in the right direction in accordance
processing and dissemination of data following inter-
with the original plans and strategic objectives.
nationally standardized classifications, including the
International Standard Classification of Occupations
Table 3.1 presents a series of indicators that, when sys-
(ISCO), the International Standard Industrial
tematically measured, can be used to track the active
Classification of All Economic Activities (ISIC), the
health workforce (2, 3). At the most basic, there is a
International Standard Classification of Education
need to know how many people are working in the field
(ISCED) and the International Classification of Status in
of health, their characteristics and distribution. In con-
Employment (ICSE).
sidering the size of a country’s health workforce at a
given moment, or measuring the stock of health work-
Depending on the data source used, indicators on
ers, it is crucial to distinguish whether the snapshot
HRH may be disaggregated by selected character-
includes workers employed at health-care facilities
istics for further analysis. Disaggregation of relevant
(differentiating between those on facility duty rosters
indicators allows for monitoring progress in health
versus those physically headcounted on the day of
26
Monitoring the active health workforce: indicators, data sources and illustrative analysis
worker training, recruitment and management policies for example, surveys with questions on care-seeking
among underserved communities or other nationally behaviour have been used to help understand how fac-
prioritized population groups. Disaggregating infor- tors such as demographics, health insurance coverage
mation on earned income among health workers by and distance to a health facility influence not only cli-
sex, for example, can be useful for monitoring gen- ents’ choice of whether or not to seek the services of
der gap in occupational earnings. Stratification of a health-care provider, but also from whom services
workforce statistics by district, province or region is have been obtained (for example public or private sec-
particularly important for monitoring equity of geo- tor, formal or informal provider).
graphical access to health services. HRH renewal
can be indirectly assessed through the age distribu- Health facility assessments can be conducted using
tion of the active health workforce, notably in terms of different sampling approaches (establishment census
the ratio of younger workers (under 30 years) to those or sample survey) and methodologies (self-admin-
close to retirement age.Depending on the nature of the istered postal, fax or Internet-based questionnaire;
indicator and the data source, an evaluation of HRH telephone or face-to-face interview). Depending on the
programmes and policies can be carried out in the nature of the data collection procedures and instru-
short, medium or longer terms. For example, certain ments, in-depth information can be obtained on health
aspects of HRH dynamics are only likely to change to workforce metrics, for instance, in-service training and
a significant degree over the long term, such as the provider productivity. In addition, the nature of facility-
production of physicians over at least a decade or so, based assessments facilitates the collection of data for
given the lengthy pre-service training requirements for numerous other indicators pertinent to health system
this category of health workers. performance assessment, such as infrastructure, avail-
ability of supplies and costs (7).
In many countries, the computerization of administra-

3.3 Overview of potential
tive records – including public expenditure, staffing
data sources and payroll, work permits, trade union memberships
Policies and programmes for the health sector should and social security records – is greatly facilitating the
be informed by timely, reliable and valid data. Despite possibilities for HRH analysis. Many skilled health-
a prevailing view that statistics on the health workforce care providers require formal training, registration and
are scarce, diverse sources that can potentially pro- licensure to practise their occupation; as such, the
duce relevant information exist even in low-income administrative records of health training institutions
countries, including population censuses and sur- and professional licensing bodies are potentially valu-
veys, facility assessments and routine administrative able sources for tracking the health workforce. These
records. There are strengths and limitations to each sources offer the advantage of producing continuously
source that need to be evaluated (Table 3.2) (2, 5). updated statistics. In addition, depending on the char-
Drawing upon a combination of these complementary acteristics of the registries, notably whether individuals
tools can result in useful and rich information for mon- are assigned a unique identifier, it may be possible to
itoring and evaluation of the health workforce and its track workers’ labour force entry, career progression
impact in health systems. and exit.
All countries collect at least some data on their popu- The fundamental challenges for data compilation,
lation, mainly through periodic demographic censuses analysis and use include identification of appropri-
and household sample surveys that produce statistical ate sources and gaining timely access to the data (2).
information about people, their homes, their socio- Other issues include the decision whether to com-
economic conditions and other characteristics. Most plement existing sources with new data collection
censuses and labour force surveys ask for the occu- activities, such as a specialized (quantitative or qual-
pation of the respondent (and other adult household itative) study of health human resources. In particular,
members) along with other demographic characteris- periodic health sector-specific labour surveys can pro-
tics, including age, sex and educational attainment. vide more detailed information, which can typically be
Labour force surveys generally delve into greater details disaggregated to a finer detail for distinct categories
on, for example, place of work, industrial sector, remu- of health workers compared to general labour force
neration, time worked and secondary employment (6). surveys. Mobilizing the required resources (human,
Many meaningful results pertinent to HRH analysis can technical and financial) for specialized HRH data col-
be produced through tabulation of population-based lection activities can often be difficult, especially in
data on labour activity. Other kinds of national house- low-income countries, but the instruments can be cus-
hold surveys can also provide relevant information; tomized to gather in-depth information on almost any
27
Table 3.1 Selected key indicators for monitoring and evaluation of human resources for health
28
Indicator Description Numerator Denominator Measurement/comparability issues
Basic indicators of HRH stock and distribution
Stock (and density) Total number of health human Total number of health workers in a (Total population of the same Definition and boundaries of HRH, such as
of HRH resources (relative to the population) given country country) by occupation (e.g. physicians, nurses, etc.),
industry or training – with distinction between
headcounts versus job positions
Skills mix Distribution of HRH by occupation, Number of physicians, nurses and Total number of health workers Occupational classification – with distinction
specialization or other skill-related midwives (or other categories of health between headcounts versus job positions (with
characteristic service providers) positions weighted for full-time equivalency on
the basis of working hours)
Geographical Distribution of HRH by geographical Number of health workers in rural areas Total number of health workers Definition of rural (or other geographical
distribution location (or other epidemiological, administrative delimitation)
or economic region)
Age distribution Distribution of HRH by age group Number of health workers of a given Total number of health workers
age group
Gender distribution Distribution of HRH by sex Number of female (or male) health Total number of health workers
workers
Indicators of HRH labour activity
Labour force activity Proportion of HRH currently active in Number of persons with health-related Total number of persons of Occupational/educational classifications as well
rate the labour force skills active in the labour force working age with health-related as age range for labour force eligibility
skills
Employment/ Proportion of HRH currently Number of persons with health- Total number of persons with Definitions of labour force participation and
unemployment rate employed (or unemployed) related skills currently employed (or health-related skills active in employment status
unemployed) the labour force
Industrial sector Distribution of workers by industry Number of persons employed in health Total number of persons Industrial classification
of activity services industry currently employed
Institutional sector Distribution of health workers by Number of health workers employed Total number of health workers Definition of operating authority of the place of
sector of activity in the public (versus private or work
nongovernmental) sector
Dual employment Proportion of HRH currently Number of health workers currently Total number of health workers
employed at more than one location employed at more than one location
Continues…
Indicator Description Numerator Denominator Measurement/comparability issues
Occupational Average occupational earnings and Total income from labour over a Total number of health workers Distinction between net/gross income, sources
earnings and income among health workers given period (from wages, practice or of income, non-monetary benefits, as well
income business) among health workers as definition of reference period for income
reporting
Indicators of HRH productivity
Absenteeism Days of absenteeism among health Number of days of employee absences Total number of scheduled Delimitation and reporting of causes of
workers over a given period in the health working days among absenteeism (e.g. duty absence, sickness or
workplace employees over the same other emergency leave, maternity or parental
period in the same place leave, unauthorized absence)
Provider Relative number of specific tasks Specific tasks performed over a Total number of specific tasks Delimitation of tasks in terms of quantity/quality
productivity performed among health workers given period (e.g. ambulatory visits, performed over the same
immunizations, surgeries) by a given period among all health service
health service provider providers
Indicators of HRH renewal and loss
Workforce Ratio of entry to the health workforce Number of graduates of health Total number of health workers Educational classification as well as processes of
generation ratio professions education institutions in the professional credentialing/ deployment for new
last year graduates
National HRH Proportion of nationally trained Number of health workers who received Total number of trained health Occupational/educational classifications across
self-sufficiency health workers their professional training in the workers in the same country the country of origin and receiving country for
reference country foreign-educated workers
Workforce loss ratio Ratio of exits from the health Number of health workers who left the Total number of health workers Delimitation and reporting of reasons for exit
workforce active labour force in the last year (e.g. retirement, mortality, out-migration, career
break or change)
Sources: Adapted from Diallo et al. (2) and WHO and University of Technology Sydney (3).
29
Table 3.2 Potential data sources for monitoring the health workforce
Source Strengths Limitations

Population tProvides nationally representative data on stock tPeriodicity: usually only once every 10 years
census of HRH: headcount of all occupations (including tDatabase management can be
private sector, management and support staff, health computationally cumbersome
occupations in non-health sectors) tDissemination of findings often insufficiently
tData can be disaggregated for specific subgroups (e.g. precise for HRH analysis, but microdata that
by age, sex) and at lowest geographical level would allow for in-depth analysis often not
tRigorous collection and processing procedures help released
ensure data quality tCross-sectional: does not allow tracking of
workforce entry and exit
tUsually no information on labour productivity
or earnings
Labour force tProvides nationally representative data on all tVariable periodicity across countries: from
survey occupations monthly to once every 5 years or more
tProvides detailed information on labour force tDissemination of findings often insufficiently
activity (including place of work, unemployment and precise for HRH analysis
underemployment, earnings) tSample size usually too small to permit
tRigorous collection and processing procedures help disaggregation
ensure data quality tCross-sectional: does not allow tracking of
tRequires fewer resources than census workforce entry and exit
Health facility tProvides information on health facility staff, including tUsually conducted infrequently and ad hoc
assessment management and support staff (headcounts and full- tPrivate facilities and practices often omitted
time equivalents) from sampling
tData can be disaggregated by type of facility, staff tCommunity-based workers may be omitted
demographics (age, sex) and geographical area tMay double-count staff working at more than
tCan be used to track wages and compensation, one facility
in-service training, provider productivity, absenteeism, tCross-sectional: does not allow tracking of
supervision, available skills for specific interventions workforce entry and exit
tUsually requires fewer resources than household- tNo information on unemployment or health
based assessments occupations in non-health services sector
tCan be complemented with routine reporting (e.g. (e.g. teaching, research)
monthly) of staff returns from each facility (such tVariable quality of data across countries and
statistics are frequently cited in official publications) over time
Civil service tProvides information on public sector employees tExcludes those who work exclusively in
payroll (headcounts and full-time equivalents) the private sector (unless they receive
registries tData are usually accurately and routinely updated government compensation)
(given strong government financial incentive for quality tDepending on the nature of the registry, may
information, which can also be validated through double-count staff with dual employment or
periodic personnel audits) exclude locally hired staff not on the central
tData can sometimes be disaggregated by age, sex, payroll
place of work, job title and pay grade tMany countries have persistent problems
eliminating ghost workersa and payments to
staff who are no longer active
Registries of tProvides headcounts of all registered health tVariable coverage and quality of data across
professional professionals countries and over time, depending on
regulatory tData are routinely updated for entries to the national the characteristics and capacities of the
bodies health labour market regulatory authorities
tData can typically be disaggregated by age, sex and tUsually limited to highly skilled health
sometimes place of work professionals
tDepending on the characteristics of the registry, it may
be possible to track career progression and exit of
health workers
a. Ghost workers are personnel formally on payroll but providing no service (notably as a strategy among health personnel to over-
come unsatisfactory remuneration or working conditions).
Sources: Adapted from Diallo et al. (2) and Pond and Kinfu (5).
30
variable of interest. An overall strategy of cross-exam- will vary depending on the users’ needs in terms of the
ination or triangulation across different data sources level of detail and technical specificity required. It has
and using different methodologies can be used to been suggested that graphs and maps can display
monitor consistency and validity of results, optimize information in a form more easily understood among
the information and evidence retrieved, and ultimately non-specialist audiences (9, 10).
provide better-quality measurements of health work-
force characteristics (see also Chapter 11). Given the For instance, if imbalance in the geographical distri-
diversity of information sources, it is especially impor- bution of health personnel is considered a constraint
tant that data dissemination includes the metadata – or to health systems strengthening – as it is in almost
details on the definition, construction and coverage of all countries (11) – the information should show the
each data point (literally: data about data) – in order to nuances of that distribution, for example by highlighting
help understand the background of the information and regions of the country with higher versus lower densi-
judge its appropriateness for the decision at hand (8). ties of health workers. In addition to being a nationally
representative data source, population censuses offer
an advantage in terms of sample sizes that allow for
3.4 Illustrative analysis disaggregation at the subnational level. Census-based
data on occupation can be used to map maldistribution
How a health or health system challenge is perceived in the health workforce, as exemplified in Figure 3.2 for
affects how related data are organized and presented two countries. Such data can also be used to calculate
to support decision-making. Even the most robust a range of summary measures of workforce imbalance,
research findings may fail to reach policy- and deci- such as the Gini coefficient or other indices of relative
sion-makers unless they are presented in a way that inequality (12, 13).
they can understand and use. A critical aspect of HRH
assessment is the identification and critical review of Given the relatively long periodicity of censuses (usu-
data from the most appropriate source(s), and synthe- ally once every 10 years), an examination of labour
sis and presentation based on the construction of a force dynamics for shorter term HRH planning and pol-
few relevant indicators. The final presentation of data icy monitoring might be better addressed by drawing
Figure 3.2 Geographical distribution of the stock of health workers (per 100 000 inhabitants), Viet Nam
and Mexico censuses
Viet Nam,1999 Mexico, 2000
<350
350–549
550–749
750+
<175
175–249
250–324
325+
Source: Gupta et al. (12).
31
Figure 3.3 Trends in the distribution of workers Figure 3.4 Distribution of health workers’
in health services by main occupational group, salaries by source, according to facility owner-
Namibia labour force surveys, 1997–2004 ship, Rwanda health facility assessment, 2006
N=14 010 100%
Source of health workers' salaries

14 N=13 136 11% 14% 12%
Workers in health services industry
12
N=10 650 21% 40%
80% 19% 22%
10
30%
(thousands)
60%
38%
8
52%
40%
6 36% 70% 66%
42% 56%
4
20%
2
27% 24%
20% 0%
0
Government / Faith-based Total
1997 2000 2004 Ministry of Health organizations
Facility operating authority

Professionals
Nongovernmental organization
Technicians & associate professionals Faith-based organization
Managers, clerks & others Government / Ministry of Health
Source: Institute for Public Policy Research (14). Source: Herbst and Gijsbrechts (16).
on survey data. Figure 3.3 presents time trends in of the proportion of staff working in facilities classified
the mix of workers in health services by broad occu- in a given sector but on the payroll of another. Such a
pational group in Namibia, drawing on results from distinction is often not made but is crucial for accurate
repeated labour force surveys (14). Labour force and statements on health labour metrics, for example, work-
employment surveys offer the advantage of being able ers in public sector facilities are often (imprecisely) all
to grid health management and support workers, a counted as public sector employees (16).
group often overlooked in HRH analyses but critical to
the functioning of health systems. On the other hand, Lastly, Figure 3.5 presents a synthesis of findings on
disaggregation for specific subgroups might be con- nurse demographics and workforce capacity cap-
strained due to sample size limitations, especially in italizing on the development of an electronic nursing
many low-income countries facing the most serious workforce database in Kenya. The database con-
health worker shortages. Based on statistics from the tains some 40 000 administrative records covering a
Global atlas of the health workforce (15), even a larger- 45-year span from the national nursing regulatory body
scale survey (for example sample size in the vicinity (17). The analyses gleaned from the data indicate the
of 0.6% of the total population) would identify only a dominance of enrolled nurses compared to their reg-
small number of health professionals in most of the 50 istered counterparts with more specialized training, as
countries with the lowest density of medical person- well as the feminization of the profession – except at
nel: fewer than 20 physicians in three quarters of these the most highly skilled level, that is, among those with
countries, and fewer than 20 nurses in half of them. a Bachelor of Science degree in nursing. At the same
time, the quality of information derived from this source
Where the existing data remain inadequate (notably must be called into question; while nursing licensure
from sources outside the health sector), a well-designed and renewal are mandatory in Kenya for both the pub-
and carried-out health facility assessment may obtain lic and private sectors, the database does not track
more detailed data for an HRH situation analysis. exits by death, out-migration or retirement. The need
Assessments that collect information at both the level to improve completeness and timeliness of available
of the facility and of the worker offer a unique opportu- data, notably for monitoring stock and flows of health
nity for measuring certain labour market indicators. In workers, is present in virtually all HRH information sys-
the example given (Figure 3.4), a health facility assess- tems, and may be even more urgent in most low- and
ment in Rwanda collected data on facility ownership middle-income countries, given the status of their
and employee salary source, allowing cross-reference workforce situation.
32
Figure 3.5 Gender and skills mix of the nursing needed at the national and subnational levels (and
workforce in Kenya, Nursing Council of Kenya even at larger facilities) for data collection, process-
database, 1960–2005 ing, reporting and analysis. Training may be required
among those providing information to decision-makers
30
Number of qualified nurses (thousands)
N=27 245 to strengthen analytical and presentation skills. Training

among policy-makers and their aides could also help
25 Male
Female them to better identify and use high-quality research.
20
Support for a centralized national HRH database
(for example at the ministry of health) is among the
15
N=11 770 potentially effective means to enhance national
77%
10 monitoring and evaluation performance (18). The coor-
dinating mechanism for this unit will ideally function in
5 79% the following capacities: ongoing assessment of data
N=265
availability and quality; data management, analysis and
female=56%
0 synthesis; generation of national indicators estimates;
Registered with Registered Enrolled information and research dissemination; advocacy
B.Sc. in nursing
and communication of HRH monitoring and evalua-
tion efforts; coordination, supportive supervision and
Source: Riley et al. (17). capacity building among state or provincial and district
monitoring and evaluation offices; and coordination with
international partners and other stakeholders. The unit
may be involved in developing and costing the national
HRH monitoring and evaluation strategy, and drafting
terms of reference for technical consultancy input. At
the decentralized level, activities of regional and dis-
trict monitoring and evaluation officers may include
3.5 Putting it all together: maintaining an updated registry of all health facilities
governance and use of HRH and health service providers; collecting routine data
on health workers from various health and non-health
information sources
sources; conducting basic data analysis; and reporting
Increasing attention is being paid at the national and data and facilitating its use for decision-making.
international levels to the need for new and improved
information for planning and monitoring HRH devel- While there is no single best model of what a health
opment as a core parameter of health systems workforce information and monitoring system should
strengthening. However, in many countries, informa- look like, much depends on the main reasons for
tion on the health workforce is fragmented, inadequate building such a system. An important consideration is
and not timely. Statistics generated by various sources planning and monitoring of the monitoring system itself
receive limited public dissemination and are generally (19). This includes delineating, as part of the process
underused. Moreover, even in countries where good of setting up and using the system, not only its contents
and reliable data are available, the information is not but also its ownership and accessibility, as well as the
always used for decision-making. mechanisms for maintaining data security and regular
updating. A key challenge is ensuring the information
Often, limited human, technical and financial resources needs of all stakeholders are being met, from the inter-
contribute to the current poor status of information and national to district level, while also being realistic about
evidence on the health workforce situation. In most ability to operationalize the system. The set of indicators
low- and middle-income countries, information and proposed earlier in Table 3.1 is neither exhaustive nor
communications technologies (for example computer absolute. Rather, it is an attempt to build a framework
hardware and software applications for the manipu- for HRH monitoring and evaluation focusing on globally
lation and communication of information) needed for identified areas of concern for workforce development.
implementing a comprehensive HRH information sys- Collection and sharing of data on such standardized
tem are sorely lacking. At the same time, strengthening indicators can be helpful for cross-national analyses,
HRH information systems means attention to each of allowing ministries of health to benchmark against
its components – not just the infrastructure and tech- regional and global performance and compare, for
nology, but also the persons needed to capture and example, what service levels and health outcomes
use the data. Dedicated health information staff are other countries are able to achieve with similar human
33
resources. It will be imperative to review this selection starting point for any investigative exercise of the HRH
at the national and subnational levels, particularly with situation should be a rigorous review of existing stand-
regard to feasibility and cost of measurement, and in ard statistical sources, including those from outside
the process of establishing appropriate country-spe- the health sector: population statistics generated by
cific baselines and targets. census bureaus and central statistical offices; work
permits from labour departments; income files from
A number of critical requirements are identified for tax departments; and others seldom used by health
developing and sustaining a comprehensive health system planners and managers. Decision-making
workforce information system, including: should draw on a meta-analysis, or investigation of
the results across several information sources. Ideally,
t the political decision to place monitoring of the HRH all HRH data sources should be integrated into one
situation high on the national agenda; comprehensive information system, whereby routine
t the establishment of a set of explicit benchmarks administrative records are complemented with regu-
and targets within the national HRH strategic plan, larly conducted population-based and facility-based
each linked with appropriate indicators and an iden- surveys and censuses.
tified minimum dataset for their measurement;
t preparing an enabling work environment for health The optimization of use of such sources, however, can
workforce information system strengthening; be hindered by the dichotomy that often exists between
t recruitment and training of a sufficient number of the providers of the data and potential users. In partic-
staff for developing, implementing and managing ular, while variables on occupation and place of work
the information system; are typically integral to population census and labour
t involving all key stakeholders in the process from the force survey questionnaires, often the final results are
initial planning stages. not disseminated using a categorization permitting the
identification of those with a health-related occupation
Some possible indicators for monitoring implementation or working in the health services industry. Even when
of the HRH information and monitoring system, along they are, the results are often not comparable across
with potential means of verification, are presented in countries or over time, due to differences in the occu-
Table 3.3 (10, 20–22). These indicators may not require pation, education and industrial classifications used.
a specific numeric answer, but they are at the centre of
monitoring usefulness of the information system for sup- As such, monitoring and evaluation of HRH requires
porting evidence-based HRH policy-making, budget good collaboration between the ministry of health and
decisions, management and accountability, and doing other sectors that can be reliable sources of informa-
so in ways that are open and transparent and optimize tion, notably the central statistical office, ministry of
stakeholder participation. Systematic use of the infor- education, ministry of labour, professional licensing or
mation system for decision-making, with most national certification bodies, and individual health-care facilities
and international stakeholders accepting its contents and health training institutions. Ideally a commitment
as reliable and valid, would be the ultimate measure to should be established in advance to investigate pur-
evaluate its performance. poseful ways to put the data to use. Discussions
between representatives of the ministry of health, cen-
tral statistical office and other stakeholders, such as
3.6 Summary and conclusions professional associations and development partners,
are recommended from the beginning to set an agenda
There is growing concern around the world about the for data harmonization, publication and use, taking into
current and future availability of health workers for account the timeline for data collection and processing
maintaining effective health systems (23). The lack of and the information needs for HRH policy and planning.
reliable, up-to-date information on numerous aspects
of the HRH situation – including skills mix, sources and
levels of remuneration, workforce feminization, and
even basic stock – greatly restricts the ability to develop
evidence-based strategies at the national and interna-
tional levels to address the health workforce crisis.
Data and evidence are necessary to inform discussion,

prioritization and decision-making among countries
and other stakeholders. Even in many low-income
countries, a variety of potential information sources
exist but remain underutilized in health research. The
34
Table 3.3 Selected indicators and means of verification for monitoring implementation and use of the
national HRH information and monitoring system to support decision-making
Indicator Description Timeline Potential means of

verification
Indicators of sound governance of the HRH information system
Existence of an operational National HRH strategy developed, From the initial National HRH strategy, key
national strategy with including a set of SMART a indicators planning stages of informants (e.g. ministry
explicit objectives, and targets, and with costed (budgeted) the national HRH of health, ministry of
indicators and targets to prioritized workplan for implementation strategy planning)
address HRH planning and and monitoring at the national and
management subnational levels
Existence of an advisory Regular meetings and consultations From the initial National HRH strategy
body to monitor among national and international planning stages of progress reports, reports/
implementation of the HRH stakeholders in health, development the national HRH minutes of advisory body
information and monitoring and information management to steer strategy meetings, key informants
system in accordance with and monitor implementation of the HRH (e.g. ministries, agencies,
the national strategy information and monitoring system institutions, associations,
NGOs, private initiatives)
Existence of a functional Existence of a national coordinating From the initial National HRH strategy
national coordinating mechanism with a dedicated unit with planning stages of progress reports, minutes
mechanism for the HRH sufficient resources (human, financial the HRH information of the coordinating
information and monitoring and technical) to develop, implement and monitoring mechanism meetings, key
system and monitor the information system system informants
HRH information system Contents of the HRH information Throughout National HRH strategy
in place and used for HRH system used to inform decision- implementation of progress reports, key
decision-making at all levels making among health authorities at the the HRH information informants
national and subnational levels on a and monitoring
regular basis (e.g. annual planning and system
management review)
Indicators of a strengthened HRH information system
Timeliness of the HRH National HRH information and Throughout Dissemination reports, key
information and monitoring monitoring system populated with data implementation of informants (e.g. ministry
system at the subnational and national levels on the HRH information of health, district health
a regular basis (e.g. quarterly/annually) and monitoring managers, professional
system bodies/associations,
private providers)
Validation of the HRH Comprehensive review of all available Throughout Dissemination reports, key
information and monitoring HRH data sources conducted and used implementation informants (e.g. ministry of
system to update and calibrate the national HRH health, ministry of labour,
information and monitoring system ministry of education,
on a regular basis (e.g. biennially/ central statistical office)
quinquennially)
Consistency of the HRH All indicators and data within the HRH Throughout Dissemination reports,
information and monitoring information and monitoring system implementation international standard
system use a common set of definitions and classifications (ISCO,
classifications allowing for consistent ISCED, etc.)
comparisons over time, across sources
and at the international level
Disaggregation of data in All relevant indicators and data within Throughout Dissemination reports
the HRH information and the HRH information and monitoring implementation
monitoring system system can be disaggregated by cadre,
gender, geographical area, sector or
other characteristics
a. SMART = specific, measurable, attainable, relevant, timebound.
Sources: Adapted from Health Metrics Network (10), World Health Organization (20), Capacity Project (21) and Islam (22).
35
References 11. Dussault G, Franceschini MC. Not enough there, too

many here: understanding geographical imbalances
1. Chankova S, Kombe G, Muchiri S et al. Rising to the in the distribution of the health workforce. Human
challenges of human resources for health in Kenya: Resources for Health, 2006, 4:12 (http://www.human-
developing empirical evidence for policy making. resources-health.com/content/4/1/12, accessed 11
Bethesda, MD, Partners for Health Reformplus January 2009).
Project, Abt Associates Inc., 2006 (http://www.
healthsystems2020.org/content/resource/ 12. Gupta N et al. Uses of population census data for
detail/1654, accessed 11 January 2009). monitoring geographical imbalance in the health
workforce: snapshots from three developing
2. Diallo K et al. Monitoring and evaluation of human countries. International Journal for Equity in
resources for health: an international perspective. Health, 2003, 2:11 (http://www.equityhealthj.com/
Human Resources for Health, 2003, 1:3 (http:// content/2/1/11, accessed 11 January 2009).
www.human-resources-health.com/content/1/1/3,
accessed 10 January 2009). 13. Anand S et al. China’s human resources for health:
quantity, quality, and distribution. Lancet, 2008,
3. World Health Organization and University of 372:1774–1781.
Technology Sydney. WHO human resources for
health minimum data set. Manila, WHO Western 14. Economic database. Windhoek, Institute for Public
Pacific and South-East Asia Regional Offices, 2008 Policy Research, 2007 (http://www.ippr.org.na/
(http://www.who.int/hrh/documents/hrh_minimum_ database.php, accessed 11 January 2009).
data_set.pdf, accessed 11 January 2009). 15. Global atlas of the health workforce: August 2007
4. Dovlo D. Wastage in the health workforce: some revision. Geneva, World Health Organization, 2007
perspectives from African countries. Human (http://www.who.int/globalatlas/autologin/hrh_login.
Resources for Health, 2005, 3:6 (http://www.human- asp, accessed 11 January 2009).
resources-health.com/content/3/1/6, accessed 11 16. Herbst CH, Gijsbrechts D. Information on stock,
January 2009). profiles, and distribution of health workers in Rwanda:
5. Pond B, Kinfu Y. Health workforce statistics: summary analysis of the 2005/06 health worker census data.
on data sources. Presented at the World Health World Bank Human Resources for Health Program
Organization/Health Metrics Network Technical Paper. Washington, DC, World Bank, 2007.
Working Group Meeting on Health Workforce 17. Riley PL et al. Developing a nursing database
Statistics, Geneva, 11–13 July 2006. system in Kenya. Health Services Research, 2007,
6. Hussmanns R, Mehran F, Verma V. Surveys of 42(3):1389–1405.
economically active population, employment, 18. Centers for Disease Control and Prevention.
unemployment and underemployment: an ILO Monitoring and evaluation capacity building for
manual on concepts and methods. Geneva, program improvement: field guide. Atlanta, GA,
International Labour Office, 1990. United States Centers for Disease Control and
7. Lindelow M, Wagstaff A. Assessment of health Prevention, Global AIDS Program, 2003.
facility performance: an introduction to data and 19. Mackay K. How to build M&E systems to support
measurement issues. In: Amin S, Das J, Goldstein M, better government. Washington, DC, World Bank,
eds. Are you being served? New tools for measuring 2007 (http://www.worldbank.org/oed/ecd/better_
service delivery. Washington, DC, World Bank, 2008 government.html, accessed 11 January 2009).
(http://go.worldbank.org/F6KIIC0700, accessed 11
January 2009). 20. Establishing and monitoring benchmarks for scaling
up health workforce education and training. Report
8. Pencheon D. The good indicators guide: prepared for the Global Health Workforce Alliance/
understanding how to use and choose indicators. Task Force for Scaling Up Education and Training for
Coventry, NHS Institute for Innovation and Health Workers. Geneva, World Health Organization,
Improvement, 2008 (http://www.apho.org.uk/ 2007.
resource/item.aspx?RID=44584, accessed 11
January 2009). 21. HRH Action Framework. Capacity Project (http://
www.capacityproject.org/framework, accessed 10
9. MEASURE Program Dissemination Working Group. January 2009).
Connecting people to useful information: guidelines
for effective data presentations. Calverton, MD, ORC 22. Islam M, ed. Health systems assessment approach:
Macro, 2004 (http://www.measuredhs.com/pubs/ a how-to manual. Arlington, VA, Management
pdf/OD41/OD41LG.pdf, accessed 11 January 2009). Sciences for Health, 2007 (http://healthsystems2020.
org/content/resource/detail/528/, accessed 11
10. Strengthening country health information systems: January 2009).
assessment and monitoring tool (version 2.00).
Geneva, Health Metrics Network, 2007 (http://www. 23. The world health report 2006: working together for
who.int/healthmetrics/support/tools, accessed 11 health. Geneva, World Health Organization, 2006
January 2009). (http://www.who.int/whr/2006, accessed 10 January
2009).
36
Framework and measurement
4 issues for monitoring entry

into the health workforce
KATE TULENKO, GILLES DUSSAULT, HUGO MERCER
4.1 Introduction of core indicators are identified, and existing and new
sources of data for their measurement are examined.
One of the main causes of shortages and maldistribu- The objective is to assist researchers, managers, pol-
tion of health workers in a country is the lack of entry into icy-makers and others to identify, obtain and use data
the labour market of adequate numbers of persons with critical to sound health workforce analysis and plan-
appropriate education and training. Yet in most coun- ning, through a consideration of the questions: What
tries few data on health worker entry are available for needs to be measured, and how?
use by the education system or the ministry of health,
impeding the ability of the health system to respond For the purpose of the chapter, “entry” means inclu-
to labour market forces or develop effective health sion into the pool of workers available for employment
workforce strategies. Measuring entry and tracking in the health sector in a country. Entry does not nec-
subsequent service are essential to the planning, man- essarily imply employment, only eligibility and desire
agement and quality control of the health workforce in a for employment. Consideration is given to all health
country (1). Gathering timely, accurate and comprehen- service providers who require vocational education
sive data on health worker entry is crucial to planning and training in a health-related field for recruitment in
the delivery of health services and effecting health pol- a job, including clinical cadres, public health work-
icy reform. Beyond the simple but commonly reported ers, community health workers and laboratory health
indicator of numbers of health education graduates workers. Those not considered include self-declared
(see also Chapter 3 in this Handbook), it is important traditional healers, drug sellers and others without for-
that the factors surrounding entry in a country are well mal training, working in the informal health sector or
understood. The availability of baseline and ongoing working illegally; those working in a non-health sector;
data enables decision-makers to monitor the progress and qualified health service providers who had previ-
of interventions and to make periodical corrections. This ously not been working but who are now re-entering
is critical as adjustments can often take years to imple- the health labour market. Some of these issues will be
ment, with effects potentially lasting even decades, and covered in Chapter 5.
can involve the reallocation or addition of significant
amounts of funding. Also, since human resources for It is important to keep in mind that decisions made
health (HRH) planning involves sectors beyond the min- regarding health worker entry affect more than just the
istry of health – including the ministries of education, aggregate number of workers. The geographical mald-
labour and finance, civil service and professional reg- istribution of health workers (usually favouring urban
ulatory bodies, and the private sector – strengthened areas over rural areas and wealthy communities over
information and monitoring systems are necessary for poor communities) can be attributed in part to pre-
following trends in the education, employment and service education and training factors. These include
regulation of health labour; assisting the decisions of insufficient recruitment of students with attributes that
managers and policy-makers; and creating a sound would lend them to serve underserved communities;
evidence base for the policy dialogue with professional the training of cadres whose skills are not matched to
associations and development partners. the needs of underserved communities; lack of student
exposure, during training, to underserved clinical set-
This chapter explores ways in which entry into the tings or populations; and failure to instil students with
health workforce can be monitored and assessed, a sense of professional obligation to underserved pop-
and how the appropriate data can lead to formulation ulations and to counterbalance students’ perceptions
and evaluation of policies and programmes to correct that there is less prestige and lower income prospects
shortages and maldistribution of health workers. A set associated with working with such populations (2–5).
37
The location of training institutions also appears to be Chapter 1 of this Handbook (see section 1.4). The pro-
related to the choice of a location of practice by grad- duction and availability of health workers can be viewed
uates (6–8). In addition, entry decisions can affect as a pipeline tracking the processes related to health
the overall health wage bill: for example, training more worker training and the development of the institutions
primary and community-based health workers than that train them, with the outputs of each component
specialist practitioners can result in a lower wage cost feeding into the next (Figure 4.1) (9–11). How each of
per unit of health service provided. the components in the pipeline can be measured, mon-
itored and evaluated will be examined.
Following this introduction, the chapter is organized
into three parts: framework for monitoring entry into the The seven components are: (i) pool of eligible can-
workforce; measurement issues; and summary and didates for health education; (ii) recruitment and
conclusions. selection of students to health education programmes;
(iii) accreditation of health education institutions; (iv)
capacity and output of health education institutions; (v)
4.2 Framework for in-migration of trained health workers from other coun-
tries; (vi) certification and licensing of health service
monitoring entry providers (nationally or internationally trained); and (vii)
This section proposes a conceptual framework that recruitment into the health labour market. In the context
divides the entry process into seven distinct but inter- of this chapter, health worker education refers to pre-
related components. Each component is explored and service vocational education and training in the field of
the policy implications discussed. The framework is health, as opposed to in-service training for upgrading
an expansion of the HRH education and in-migration skills among workers already employed in the health-
paths within the working lifespan framework developed care industry. An illustrative example of monitoring
by the World Health Organization (9) and introduced in in-service training will be presented in Chapter 7.
Figure 4.1 Framework for

Pool of eligible
monitoring entry into the health candidates for health
workforce education
Recruitment and
selection of students
to health education
programmes
Health education and training
Capacity and Attrition from

Accreditation of
output of health health education
health education
education programmes (failure,
institutions
institutions transfer, etc.)
In-migration of Completion of health

trained health education programmes
workers from other
countries
Occupational entry Professional

Out-migration
requirements met certification or licensure
Health labour market
Provision of health services
Health Health
research administration
Health
education and Health policy
Sources: Adapted from WHO (9), training
Allen et al. (10) and Dussault et al. (11).
38
Framework and measurement issues for monitoring entry into the health workforce
4.2.1 Pool of eligible candidates Educational institutions may post the opening of the
for health education acceptance of applications on their public notice board,
but there is generally little outreach to the pool of eligi-
Within each country, the size of the pool of eligible peo-
bles. From a policy point of view, the need for active
ple from which health training institutions recruit their
recruitment to correct gender, economic, ethnic, urban/
students depends primarily on the admission criteria
rural and regional imbalances should be examined.
for each training programme and the strength of the pri-
Also to be considered is the provision of assistance to
mary and secondary school system. For tertiary-level
potential students in their choice of institution and in fill-
programmes producing the most highly skilled health
ing out the application forms, which may be especially
service providers, this pool traditionally consisted of
daunting to those from disadvantaged socioeconomic
students having graduated at the upper secondary
groups or underserved communities who may be the
level – equivalent to level 3 of the International Standard
first in their family to apply for higher education.
Classification of Education (12) – and with strong sci-
ence backgrounds.
Recruiting students based on their motivations for
pursuing a health career can help improve worker
With the increasing emphasis on the need for locally
retention. Evidence from an observational study of
recruited cadres for providing basic preventive and
Ethiopian nurses and physicians revealed that students
curative care and referral services, especially among
with higher reported rates of altruism (measured as will-
underserved communities, some countries have
ingness to help the poor) were willing to work in rural
recognized that their requirements for eligibility for
areas for a lower rural bonus, and were more likely to
admittance into certain types of health training pro-
still be practising in a rural area when followed up two
grammes were unnecessarily stringent, and have
years later (4). A study on the migration of health work-
made the criteria more suitable to the responsibilities
ers from Ghana to the United Kingdom and the United
of the cadres. The pool of eligibles has been widened
States of America found that many nurses and physi-
to include those without upper secondary diplomas or
cians had entered the health field with the intention to
strong science backgrounds. For example, to increase
migrate, and that a health career was often seen as a
retention of nurses in rural areas of Pakistan, the Aga
“ticket” out of Ghana (14).
Khan School of Nursing developed a programme that
recruited young women who had graduated from rural
secondary schools but essentially, due to the weakness 4.2.3 Accreditation of health
of their underresourced schools, at only the lower sec- education institutions
ondary level. The remedial programme quickly brought
All health professions education institutions, public or
their knowledge level up to meet the qualifications to
private, should be accredited to assure the match of
enter nursing school, and they then joined the regular
health workers and their skills with the country’s health-
nursing programme along with their counterparts who
care needs, and to ensure the quality of education
graduated from urban secondary schools (13).
provided. The accreditation process should be driven
by the national health policy and be conducted in a
The pool of eligibles is one of the most underappre-
manner that makes it socially responsible, while main-
ciated policy issues within HRH and, as a result, few
taining the independence of the accreditation agency
ministries of health have accurate data or effective pol-
(15, 16). The mechanism consists of an initial formal rec-
icy on eligibles. With appropriate data and analysis,
ognition of training institutions by a representative body
policies on the entry requirements for students can be
(usually at the national or sometimes subregional level)
made or changed so that they are more aligned to the
that certain predetermined educational requirements
country context; outreach programmes can be started
have been met – covering such aspects as instruc-
to interest high-school students or others to become
tors’ qualifications, curriculum and clinical rotations
health workers; and training programmes can be set
– followed by periodic assessments to ensure mainte-
up within high schools. Or the analysis may reveal that
nance of standards. Elements of proper accreditation
more sweeping changes are needed in the primary
and quality assurance processes of health education
and secondary curriculum to properly prepare stu-
institutions include authoritative mandate and deci-
dents for health careers.
sion of the accreditation agency; social accountability;
independence from government and providers; trans-
4.2.2 Recruitment and selection of parency; predefined general and specific criteria for
students to health education programmes education standards; procedures using a combination
In most low- and middle-income countries, especially of institutional self-evaluation and site visits by exter-
those that rely exclusively on public training institutions, nal reviewers; and publication of reports and decisions
recruitment of students remains a passive process. (17).
39
Potential uses of accreditation data to support deci- 4.2.5 In-migration of trained health
sion-making include identifying practices from workers from other countries
high-performing training institutions that can be rep-
Countries with better wages, working conditions and
licated at other institutions, and identifying poorly
quality of life tend to attract health workers from other
performing institutions in need of increased attention. A
countries. In order to legally exercise their occupation
large number of institutions failing to meet accreditation
in the destination country, in-migrants must receive
or reaccreditation standards could indicate that insti-
working visas and, for certain skilled health service
tutional management may need to be improved, that
providers, be licensed or certified by the appropriate
the standards are unrealistically high or that education
regulatory body.
institutions are underresourced.
Policy options on migration for destination or receiving
Several barriers exist to the effective use of accredi-
countries include adjusting the number of visas ear-
tation data. In Ghana it has been found that health
marked for health workers, the degree to which their
science training institutions did not receive copies of
visa applications are facilitated and expedited, and how
their accreditation report, nor were the reports pub-
actively the government allows the public or private sec-
licly available (18). Another barrier to data use is that
tor to recruit internationally (19). Policies and practices
many countries do not have an accreditation body, or
can be active, for example when the government posts
the one that exists is underresourced. There is a need
advertisements in other countries, sends recruiters or
to promote national and regional policies to enhance
negotiates bilateral arrangements with other countries;
the accreditation of health professions education insti-
or passive, that is, simply considering health workers
tutions as a way to ensure the quality of health services
like all others who apply for visas on their own accord.
delivery.
For sending countries, options can include increased
4.2.4 Capacity and output of funding for the production of health workers to meet
health education institutions demand abroad, and policies of return (20). Of note
is the Philippines’ policy on assistance to international
Capacity in pre-service training includes physical
migration for its nurses to many receiving countries. As
infrastructure (for example classrooms, laboratories,
part of its managed migration strategy, the Philippines
libraries, clinics for internships, campus residencies),
negotiates the number of workers the destination coun-
human resources (quantity and quality of instructors
tries will receive and the terms of their service (the
and auxiliary staff), financial resources, organizational
agreements signed with the United Kingdom in 2003
and operational capacity (managerial structure and
and with Japan in 2006 are examples). This may be
processes) and other non-infrastructure physical inputs
considered a mutually advantageous process for both
(pedagogical tools, reference books and journals,
countries, as it allows both the Philippines and the
computer equipment) (11). Health worker training insti-
receiving country to conduct long-range nurse work-
tutions, accredited or not, can vary greatly by capacity,
force planning and minimize sudden shocks (21). On
and in many cases the training institution may not be
the other hand, unexpected (and undesirable) side-
aware of its own capacity or potential capacity.
effects of this strategy may include encouraging nursing
teachers and trainers to leave, encouraging physicians
The various components of capacity determine the
and other health professionals to retrain as nurses to
overall output capacity: the number, type and quality
improve their chances to emigrate (22) and weakening
of cadres that graduate from the institution. Combining
of the health system, particularly in rural areas.
data on output with that on financing can be used to
calculate how much it costs to train each type of health
Increasing attention is being focused on the incorpo-
worker, and to estimate how much it would cost to train
ration of ethical codes of practice into national practice
additional workers based on current capacity.
(20). Policy options being explored by a country should
recognize the right of individual workers to migrate, and
For policy purposes, it is critical to monitor each of
denounce unethical recruitment practices that exploit
the components of capacity and output. By draw-
health workers or mislead them into accepting job
ing on assessments of oversupply or undersupply of
responsibilities and working conditions that are incom-
various cadres of the active health workforce, institu-
patible with their qualifications, skills and experience
tional capacity for producing new health workers can
(23, 24). Among receiving countries, they should also
be decreased or increased, or training programmes for
acknowledge that the flow of international migration of
new cadres can be developed. This information can
skilled health professionals is generally from poorer to
be used to identify the specific bottlenecks in capac-
wealthier countries, who gain a valuable resource with-
ity so that if rapid increases in production are required,
out paying the education and training expenses.
capacity can be increased as rapidly as possible.
40
4.2.6 Certification and licensing communities. Since these workers will often be the first
of health service providers point of contact with the formal health-care system, and
therefore will represent the system at the community
Certification and licensing are used by countries to con-
level, guaranteeing the quality of the workers through
trol the quality of health-care workers practising in their
proper certification is of extreme importance.
country and to control the size of the health labour mar-
ket. Certification and licensing purposefully weed out
For in-migrants, typical requirements for professional
unqualified workers because those whose knowledge
certification or licensing in the receiving country vary
and skills do not match the minimal requirements for
greatly. Most countries require graduation from a train-
their cadre can do more harm than good to the health
ing institution recognized by the receiving country (for
of their patients and can erode the confidence that the
example, based on recognition of meeting the qual-
public has in the health system, especially of govern-
ity assurance standards of the World Federation of
ment-provided services. The location of certification
Medical Education) and professional certification or
and licensing can vary for different cadres and coun-
licensing in the country of education. Most countries
tries and this affects how the data can be gathered.
have credentialing staff in medical boards or the minis-
For certain cadres, quality at the level of the individual
try of health who document the certification or licensing
worker is controlled by graduating from an accredited
of the workers and their work history. Some countries
training institution, while for other cadres, the require-
have streamlined this process and designed reciprocal
ment is passing a national professional qualification
recognition procedures with other countries.
exam. Also affecting potential monitoring and evalu-
ation efforts is the fact that certification to practise a
A special issue with regard to immigrant health work-
profession usually does not need to be renewed, while
ers is that of language and cultural competency. Some
a licence usually needs to be periodically renewed
countries may require passing a proficiency exam for
based on certain criteria such as passing a renewal
the language most used in professional communica-
exam, demonstrating continuing learning, being
tion at the national level. However, there tend to be few
employed in the field or simply paying a fee (25).
or no requirements for knowledge of local languages
and culture, which may not favour retention of immi-
Monitoring trends in professional certification and
grant health workers in rural areas. A provider’s lack
licensing numbers and success rates can help identify
of common language with patients may also affect cul-
a variety of problems in the entry process. For exam-
tural acceptability of the health system, and negatively
ple, an increase in the licensing exam failure rate may
impact care-seeking behaviours and treatment compli-
indicate insufficiencies in training curricula or exams
ance (26).
that are outdated in relation to changes and innovations
in clinical practice. Another policy issue that may need
In any case, professional certification and licensure
to be addressed is whether conflicting quality con-
can document the quality of health workers at entry, but
trol criteria exist between the government, individual
does not necessarily reflect the quality of care they pro-
training institutions and professional regulatory bod-
vide. Quality of service provision is affected by many
ies. There are cases where graduates succeed in the
factors, such as workload, motivation, supervision,
institutional proficiency tests, but fail the professional
available resources (for example equipment, supplies,
association certification exam, indicating a mismatch in
support staff) and lifelong learning.
the level of proficiency expected at the institutional ver-
sus the association level. This points to the need to set
and use common standards within a country, and align 4.2.7 Recruitment into the
training curricula with professional knowledge and skill health labour market
requirements. Monitoring the recruitment of newly trained health work-
ers into the national health labour market is critical in
With the expected global increase in the production of order to reduce inefficiencies in the hiring system,
front-line cadres that are usually certified rather than identify potential gaps between supply and demand
licensed, such as community health workers and aux- for health workers, and monitor achievements in
iliary nurses, it is crucial that the certification process health workforce planning. Policies and strategies for
is better monitored. This increase will occur because deployment of health workers vary according to the
many countries with extreme shortages or maldistribu- context and dynamics of countries’ health, education
tion of highly skilled health service providers, especially and labour markets. Countries with only government-
medical and nursing professionals, are consider- operated education institutions and few health-related
ing or opting for rapid production of large cadres of private sector jobs have simple health labour markets;
lower-skilled workers to meet the immediate needs for active recruitment of workers to health-care service is
basic health services among underserved, mainly rural not needed since all graduates are directly employed
41
by the government, or do not work if posts are not avail- 4.3 Measurement issues
able. But countries with private training institutions or a
significant formal private health sector have more com- Each section of the health workforce entry framework
plex health labour markets, requiring active recruitment can be monitoring by means of its own set of indicators
to fill job vacancies. for measurement. Table 4.1 (page 44) presents a series of
indicators on entry and their potential means of verifica-
Policy issues include making sure the application and tion. This list is by no means exhaustive, and additional
posting process is as transparent and timely as possi- indicators may be required. The issue of a minimal set
ble. Health workers should be able to apply to specific of essential indicators is a complex one, and there may
posts and the criteria for selection should be clear. be conflicting priorities between what data are needed
Governments can improve the efficiency of the health at the national level and the international level. Every
labour market by establishing free, easily accessed country and region has a unique HRH situation, and will
job boards on which all job seekers and employers need to collect and analyse the necessary information
can post. In some contexts, offering incentives (mone- most suited to its health system needs, objectives and
tary and non-monetary) may be needed to encourage targets (Box 4.1) (28). Throughout this discussion, it is
workers to apply for posts in underserved areas. important to keep in mind the need, where possible, to
Establishing early links between potential employ- routinely compile, analyse and act on data collected
ers and educational institutions is also an option that through existing national administrative processes (Box
enlarges the students’ knowledge on their future labour 4.2) (12, 29). This routine data collection can then be
trajectory. supplemented and validated through periodic or ad
hoc surveys and other standard statistical sources (for
If a country imports or exports health workers, rele- example population census or labour force survey).
vant questions include: What are the recruitment rules
that need to be followed? How many health workers in- Within the framework, capacity and output of train-
migrate or out-migrate each year? Is reimbursement ing institutions is the field with the largest number of
needed from the receiving to the sending country to proposed indicators. This is also probably where lies
compensate the latter for health workers trained with the greatest potential for changes in the shorter term
public funds? How should this recruitment be con- in response to policy or programmatic interventions.
ducted? For example, in some African countries, One entry indicator that each country should routinely
recruiters for foreign health systems have been allowed measure, analyse and disseminate is the annual output
to set up tables at nursing graduation ceremonies and (or number of graduates) of health vocational training
directly recruit the new nurses (27). institutions (see Box 4.3 for an illustrative example) (11).
This is the aggregate of multiple pieces of information
Box 4.1 Illustrative example of the establishment of health worker education

and training goals and targets: region of the Americas, 2007–2015
An HRH strategic plan for the Americas proposed a set of goals, targets and indicators for the countries
in the region in several areas of education and training. Under the goal “adapt the education of the health
workers to a universal and equitable model of providing quality care to meet the health needs of the entire
population”, specific benchmarks for 2015 included:
t 80% of schools of clinical health sciences will have reoriented their education towards primary health
care and community health needs and adopted interprofessional training strategies;
t 80% of schools of clinical health sciences will have adopted specific programmes to recruit and train
students from underserved populations with, when appropriate, a special emphasis on indigenous, or
First Nations, communities;
t attrition rates in schools of nursing and medicine will not exceed 20%;
t 70% of schools of clinical health sciences and public health will be accredited by a recognized
accreditation body.
Source: Pan American Health Organization (28).
42
Box 4.2 Some notes on data collection, processing and use
t Among the recommended requirements for strengthening the collection, processing, analysis and use
of data on HRH production is the replacement of paper-based administrative records and registers
with electronic data processing systems. This includes securing the necessary human, financial and
technical resources for developing and strengthening the information system, which would capture
data at different levels (including basic education, institution-based training and community-based
training) for ongoing monitoring of progress. However, in certain local conditions, such as unreliable
electricity supply, dust problems or poor access to computer repair services, a robust paper system
may be preferable to an unreliable computerized system.
t Comparability of HRH education statistics within and across countries and over time can be enhanced
through the setting and use of common definitions and classifications. This includes the collection,
processing and dissemination of data following or mapped to the International Standard Classification
of Education (12) (or national equivalent).
t Special permission may be needed to access certain types of data for research and policy purposes,
such as student professional qualification exam or institution accreditation scores. Individual-level
records should be accessible only to those who need to work directly with them, and all identifiers
(such as name and unique identification number) must be removed from the dataset prior to distribution
and use for analysis. In some cases, the level of precision of certain variables that may not be common
to several individuals (for example age, district of residence or clinical specialization) may need to
be changed to reduce the risk of indirect personal identification. Various techniques can be used
for anonymizing microdata from administrative and survey sources, such as those developed by the
International Household Survey Network (29).
Box 4.3 Illustrative example of data on trends in medical education

institutional capacity, recruitment and output: Rwanda, 2002–2006
Annual number of admissions and graduates,

Faculty of Medicine, National University of Rwanda,
120 2002–2006
90
2002
2003
Number
60 2004
Maximum
capacity 2005
2006
30
0
Enrolments Graduates
Source: Dussault et al. (11).
43
44
Table 4.1 Key indicators and means of verification for measuring entry into the health workforce
Indicators Potential data sources Complementary dimensions

Pool of eligible candidates for health education
tNumber of students graduating from primary school, e.g. expressed Ideally assessed through routine administrative Data on eligible students ideally disaggregated by age, sex,
as % of all children of primary schooling age records submitted by individual primary and urban/rural or other characteristic that would lend them to serve
tNumber of students graduating from secondary school, e.g. secondary schools (ministry of education). underserved communities.
expressed as % of all children of secondary schooling age Can also be assessed by interviews with key Additional qualitative information may be required on the quality/
informants (e.g. district school managers). relevance of the secondary science curriculum.
tNumber and % of students graduating from secondary schools
with science concentrations (or other entry requirements for health Information on the total number of children Further information may also be needed on requirements to enter
vocational training) belonging to the age group that officially cor- training for lower-skilled occupations (such as community health
responds to primary and secondary schooling workers).
should be periodically validated against data
from a population census or other nationally
representative source (central statistical office).

Recruitment and selection of students to health education programmes
tNumber of applicants per training place, per cadre Ideally assessed through routine administrative Data on applicants ideally disaggregated by age, sex, urban/rural
tNumber and % of applicants meeting entry requirements per place, records submitted by individual health training or other characteristic that would lend them to serve underserved
per cadre institutions (ministry of health, ministry of communities.
education). Can also be assessed through a Additional qualitative information may be useful on recruitment
tNumber and % of applicants accepted for training programmes, per quantitative survey of training institutions or
cadre strategies (especially targeting certain population groups),
interviews with key informants (e.g. managers reasons applicants did not qualify for training and reasons
tNumber and % of accepted applicants who register for training, per of training programmes). accepted applicants did not eventually register for the programme.
cadre
Accreditation of health education institutions
tExistence of an accreditation agency of health education and training Can be assessed through document reviews Data on accreditation results ideally disaggregated by type of
institutions (e.g. evaluation reports) or interviews with institution (public/private) and region.
tNumber and % of health training institutions meeting accreditation key informants (ministry of health, ministry Additional qualitative information may be required on the authority
and reaccreditation standards of education, national or subregional experts of and resources available to the accreditation agency, and on the
of accreditation processes and education main barriers to institutional accreditation (e.g. reasons for failure
standards). to obtain accreditation, most commonly missed criteria).
Continues…
Indicators Potential data sources Complementary dimensions
Capacity and output of health education institutions
tNumber of education and training places per cadre Indicators on training capacity and output Data on training capacity, attrition rates, output, expenditures and
tNumber of places in laboratories or clinical internships, per cadre ideally assessed through routine administrative costs disaggregated by type of institution (public/private) and
records submitted by individual health training region. Data on graduates should be disaggregated by age, sex,
tNumber of students per qualified instructor, per cadre
institutions (ministry of health, ministry of urban/rural or other sociodemographic characteristics.
tNumber of students per personal computer, per cadre education). Can also be assessed through a Additional qualitative information may be required on main
tNumber of library books and journals per student, per cadre quantitative survey of training institutions. bottlenecks in training capacity (e.g. recruitment, qualifications
tAttrition (drop-out) rate per student cohort, per cadre Data on government expenditure ideally and retention of instructors), opinions on accessibility to clinical
tAttrition (turnover) rate among instructors, per cadre available from ministry of finance. Additional environments and other resources, career expectations (for
tNumber of students graduating each year, per cadre data on training costs required to take account both instructors and students), career counselling/mentoring
of private expenditure (e.g. tuition fees, budget programmes for students, and reasons for student attrition (e.g.
tGovernment expenditure on health vocational training, per cadre of private institutions, household expenditure failure, transfer to a non-health programme, migration).
tPrivate expenditure on health vocational training, per cadre survey).
tTotal cost per graduate for health vocational training, per cadre
In-migration of trained health workers from other countries
tNumber of non-national health workers applying for entry visas, Ideally assessed through routine administrative Data on in-migrants ideally disaggregated by age, sex and country of
per cadre records (ministry of foreign affairs). origin. Additional follow-up data could be useful, including eventual
tNumber of entry visas issued to non-national health workers, posting (urban/rural) and length of stay in the destination country.
per cadre Qualitative information on special visa programmes for trained
health workers and bilateral agreements for managed migration
may also be required.
Certification and licensing of health service providers
tNumber and % of new nationally trained health workers granted Ideally assessed through routine administrative Additional qualitative information may be required on main
professional certification/licensure, per cadre records (professional regulatory bodies). reasons for unsuccessful certification/licensing.
tNumber and % of new internationally trained health workers granted
professional certification/licensure, per cadre
Recruitment into the health labour market
tExistence of job boards to facilitate recruitment of newly trained Ideally assessed through routine administrative Data on new entrants to the health labour market ideally
health workers records (ministry of health, ministry of labour, disaggregated by age, sex, urban/rural, and place of work (public/
tNumber of newly graduated health workers who are employed in ministry of foreign affairs, professional private).
the health labour market within 3 months of graduation (or other regulatory bodies, associations of private Additional qualitative information may be required on regulations
nationally defined time period), per cadre providers). and practices for internal and external recruitment, such as
tNumber of newly graduated or licensed health workers who are Information on labour market participation transparency of government practice, offering incentives to serve
should be periodically validated against data in rural areas, and ethical recruitment of foreign workers.
45
diverted from the national health labour market (e.g. unemployed,
migrate, choose not to work, or work in a non-health job), per cadre from a national labour force survey.
depending on the number of cadres in the health sys- example associations of private providers and non-
tem. The number and type of newly trained health governmental or faith-based organizations that provide
workers is relevant everywhere: in countries that need health services). Ideally, the information system would
increased production among all cadres, countries that be characterized with the use of unique personal
need increased training for workers tailored to rural identification numbers, which would allow tracking of
and underserved areas, and receiving countries that individual workers from the time they enter pre-service
are aiming towards national self-sufficiency of health training and throughout their career; this will be dis-
worker entry. cussed further in Chapter 9.
At the international and regional levels, it is important Entry can be politically sensitive: through admission
to have standardized definitions, indicators and meas- criteria to health education programmes and proc-
ures that can be compared and aggregated across esses for credentialing and regulation, political actors
countries for global health workforce monitoring. The negotiate their respective interests. In many low- and
most commonly reported measures related to entry middle-income countries, health-care careers, espe-
are production of physicians, nurses and midwives (9). cially among physicians and nurses, are among the
However, additional information on other categories most highly respected, remunerated and sought-
of health workers should be considered to reflect the after. Many stakeholders watch closely entry into the
vital importance of all human resources in health sys- health professions. Failure to understand and address
tems, such as pharmacists, public health workers and these stakeholders’ interests can result in opposition to
community health workers. As discussed in Chapter health workforce reform. In addition, in many countries
2, given the differences across countries in occupa- the pool of eligible candidates for advanced training in
tional titles, training requirements and responsibilities, health restricts recruitment, and the search for alterna-
data should be processed and disseminated such as tive solutions has led to populist (but not necessarily
to enhance comparability across countries and over sustainable) pathways for expanding the health work-
time, notably by means of mapping to the International force, such as engaging volunteers or importing health
Standard Classification of Education. workers from other countries. Another sensitive issue
is the division of tasks among various cadres in the
Also of particular importance is information on govern- health-care team – for example, those who can pre-
ment expenditure and financing for health vocational scribe medicines or perform a Caesarean section – or
education and training. These data can be used to the creation of new cadres, such as paramedical prac-
estimate the current costs of producing health workers, titioners (sometimes called physician assistants or
and are needed for planning purposes to project future clinical officers) or community health workers, which
costs. Combining information on expenditure with data typically provokes resistance from established cadres.
on student attrition or professional licensing exam pass In some contexts, entry into the health workforce can
rates will enable more efficient training systems to be be also a lever for cultural and social changes, such
identified, and wasteful programmes to be improved or as an increase in the proportion of women or students
even eliminated. from rural or other underserved communities trained for
a professional career in this sector.
Many data already exist within countries on heath
worker production but they are often difficult to access
and analyse. They tend to be fragmented, not shared
and stored in a form that make them difficult to com-
pare with data from other sources. Administrative
records may not be computerized and archived, and
statistics on the different components of HRH pro-
duction and entry are not routinely compiled, updated
or analysed. Further challenges include the fact that
many information systems only contain data for the
public sector. As such, the usefulness of the data for
policy-makers and researchers will depend on the
completeness, reliability and timeliness of the informa-
tion and monitoring system, which itself is dependent
on the level of collaboration between key stakeholders,
including the ministry of health, ministry of education,
ministry of finance, individual training institutions, pro-
fessional regulatory bodies and the private sector (for
46
4.4 Summary and conclusions References

As has been discussed in this chapter, timely and 1. Pan American Health Organization. Public health
accurate data on entry into the health labour market in the Americas: conceptual renewal, performance
assessment and bases for action. Washington, DC,
are essential for evidence-informed planning and man- Pan American Health Organization/World Health
agement of health systems. During implementation of Organization, 2002.
national HRH plans and strategies, policies and inter-
2. Dunbabin J, Levitt L. Rural origin and rural medical
ventions must be monitored, managed and adjusted exposure: their impact on the rural and remote
if necessary. The labour or educational market may medical workforce in Australia. Rural and Remote
change quickly as the economy changes, positively Health, 2003, 3:212 (http://www.rrh.org.au/articles/
or negatively affecting the quality of candidates for subviewnew.asp?ArticleID=212, accessed 12
January 2009).
health occupations. The population size and structure,
immigration patterns and burden of disease may also 3. Dussault G, Franceschini MC. Not enough there, too
many here: understanding geographical imbalances
change, necessitating a change in training curricula or
in the distribution of the health workforce. Human
workforce size or skill mix. All these changes have to Resources for Health, 2006, 4:12 (http://www.human-
be taken into account in the policy-making process. resources-health.com/content/4/1/12, accessed 11
Health sector actors must actively monitor the whole January 2009).
situation to be able to respond appropriately, and to be 4. Serneels P et al. For public service or money:
able to gauge the success of interventions. understanding geographical imbalances in the
health workforce. Health Policy and Planning, 2007,
Seven components of the health worker production 22(3):128–138.
pipeline were identified for which data are needed. 5. Wibulpolprasert S. Inequitable distribution of
They were: (i) the pool of eligible candidates for health doctors: can it be solved? Human Resources
education; (ii) recruitment and selection of students Development Journal, 1999, 3:1 (http://www.who.int/
hrh/hrdj/en/index4.html, accessed 12 January 2009).
to health education programmes; (iii) accreditation
of health education institutions; (iv) capacity and out- 6. Wang L. A comparison of metropolitan and rural
medical schools in China: which schools provide
put of health education institutions; (v) in-migration of
rural physicians? Australian Journal of Rural Health,
trained health workers from other countries; (vi) cer- 2002, 10:94–98.
tification and licensing of health service providers
7. Salafsky B, Glasser M, Ha J. Addressing issues
(nationally or internationally trained); and (vii) recruit- of maldistribution of health care workers. Annals
ment into the health labour market. As such, measuring of the Academy of Medicine Singapore, 2005,
and monitoring the entry function requires comprehen- 34(8):520–525 (http://www.annals.edu.sg/
sive information on education and training at different pdf/34VolNo8200509/V34N8p520.pdf, accessed 12
levels, including basic education. January 2009).
8. Smucny J et al. An evaluation of the rural medical
Policy-makers have a responsibility to analyse the rele- education program of the State University of New
York Upstate Medical University, 1990–2003.
vant data and formulate policy on all of the seven entry
Academic Medicine, 2005, 80(8):733–738.
issues. The process of collecting, analysing and acting
on these entry data is not just a one-time exercise, but 9. The world health report 2006: working together for
health. Geneva, World Health Organization, 2006
would be an ongoing activity of the ministry of health
(http://www.who.int/whr/2006, accessed 10 January
(or other agency mandated by the government for that 2009).
purpose). At certain critical points of HRH planning and
10. Allen MK et al. Educating health workers: a statistical
monitoring, special surveys or studies may be needed portrait, 2000 to 2004. Ottawa, Statistics Canada,
to validate or gather additional entry data that are not 2007 (http://www.statcan.gc.ca/pub/81–595-m/81–
feasible to gather on a routine basis. It is important for 595-m2007049-eng.pdf, accessed 13 January
the ministry of health to partner with other responsi- 2009).
ble ministries, education and training institutions, and 11. Dussault G et al. Assessing the capacity to produce
stakeholders to work together to provide the coun- health personnel in Rwanda. Leadership in Health
try with the proper numbers of appropriately trained Services, 2008, 21(4):290–306 (http://www.
emeraldinsight.com/10.1108/17511870810910092,
health-care workers. accessed 13 January 2009).
12. International Standard Classification of Education:
ISCED 1997. Paris, United Nations Educational,
Scientific and Cultural Organization, 1997 (http://
www.uis.unesco.org/TEMPLATE/pdf/isced/
ISCED_A.pdf, accessed 10 January 2009).
13. Bryant NH, ed. Women in nursing in Islamic
countries. Oxford, Oxford University Press, 2003.
47
14. Ozden C. Migration of nurses and physicians from

Ghana. World Bank Working Paper. Washington, DC,
World Bank, 2008.
15. Hennen, B. Demonstrating social accountability in
medical education. Canadian Medical Association
Journal, 1997, 156(3):365–367 (http://www.
pubmedcentral.nih.gov/picrender.fcgi?artid=122695
8&blobtype=pdf, accessed 13 January 2009).
16. Boelen C. Building a socially accountable health
professions school: towards unity for health.
Education for Health, 2004, 17(2):223–231.
17. Karle H and World Federation for Medical Education
Executive Council. International recognition of basic
medical education programmes. Medical Education,
2008, 42(1):12–17.
18. Beciu H, personal communication. Washington, DC,
World Bank, 2008.
19. Aluwihare APR. Physician migration: donor country
impact. Journal of Continuing Education in the
Health Professions, 2005, 25(1):15–21.
20. Bach S. International migration of health workers:
labour and social issues. Sectoral Activities
Programme Working Paper No. 29. Geneva,
International Labour Office, 2001 (http://www.ilo.ch/
public/english/dialogue/sector/papers/health/wp209.
pdf, accessed 13 January 2009).
21. Lorenzo FME et al. Nurse migration from a source
country perspective: Philippine country case study.
Health Services Research, 2007, 42(3p2):1406–1418.
22. Brush B, Sochalski J. International nurse migration:
lessons from the Philippines. Policy, Politics, and
Nursing Practice, 2007, 8(1):37–46.
23. Ethical nurse recruitment. Geneva, International
Council of Nurses, 2001.
24. The World Medical Association statement on
ethical guidelines for the international recruitment
of physicians. Ferney-Voltaire, World Medical
Association, 2003.
25. Schuwirth L. The need for national licensing
examinations. Medical Education, 2007,
41(11):1022–1023.
26. Shengelia B et al. Access, utilization, quality, and
effective coverage: an integrated conceptual
framework and measurement strategy. Social
Science and Medicine, 2005, 61(1):97–109.
27. Xu Y, Zhang J. One size doesn’t fit all: ethics of
international nurse recruitment from the conceptual
framework of stakeholder interests. Nursing Ethics,
2005, 12(6):571–581.
28. Pan American Health Organization. Regional
goals for human resources for health 2007–2015.
Washington, DC, Pan American Health Organization
and World Health Organization, 2007 (http://www.
paho.org/english/gov/csp/csp27–10-e.pdf, accessed
12 January 2009).
29. International Household Survey Network (http://www.
internationalsurveynetwork.org/home/, accessed 13
January 2009).
48
Monitoring health workforce
5 transitions and exits

PASCAL ZURN, KHASSOUM DIALLO, YOHANNES KINFU
5.1 Introduction factors associated with exit from the active workforce;
a comprehensive set of indicators that could poten-
The world health report 2006, published by the World tially be measurable from standard statistical sources
Health Organization (WHO), drew global attention to is proposed; and the underlying data requirements are
the human resources for health (HRH) crisis, man- specified. Illustrative analyses from several contexts
ifested by shortages and imbalances in the health are also provided, and the implications of observed
workforce undermining the performance of health sys- patterns for policy and planning are discussed.
tems and exercising adverse impacts on the ability of
many countries to promote and enhance the health of
their population (1). The HRH crisis has several causes
and consequences. In many low- and middle-income
5.2 Transitions within and exits
countries, in addition to past and current investment from the health workforce:
shortfalls in pre-service education and training, interna- a framework for analysis
tional migration of skilled health workers and premature
Building an adequate workforce supply, capable of
exit from the health labour pool due to career change,
addressing the health needs of the population, begins
early retirement, work-limiting morbidity and prema-
with the education and deployment of skilled health
ture mortality are among the main responsible factors.
workers. Once having entered the health system, they
Yet, in many countries, the dynamics of movements
then experience different forms of transition along the
within and exits from the health workforce are poorly
working lifespan. There is a growing body of litera-
understood, limiting the capacity of governments and
ture showing that health workforce turnover is directly
stakeholders to design and implement effective, equi-
influenced by inadequate compensation, poor work-
table and cost-efficient intervention programmes for
ing conditions (such as lack of medical equipment and
enhancing workforce retention. This challenge ema-
poor workplace safety) and job dissatisfaction (due
nates in part from the absence of timely and relevant
to various reasons, including low work autonomy, lim-
data and information on inflows and outflows of health
ited opportunities for professional development and
workers, but also from the lack of widely accepted
inflexible working hours), combined with better career
standardized indicators associated with the measure-
options elsewhere and other factors exogenous to the
ment of workforce flows.
health system (for example living conditions and edu-
The most commonly monitored dimensions of health cational opportunities for children) (2–4). This includes
workforce metrics have traditionally been limited to two geographical movements of health workers within and
indicators: current health workforce stock (the number across countries, professional movements within and
of workers participating in the health labour market at outside the health sector and other types of move-
a given time) and the number of newly trained health ments. For the purpose of the present chapter, all
service providers (number of graduates of health labour movements that occur within the national health
professions education programmes over a given refer- sector are considered as “transitions” (or intrasectoral
ence period) (as detailed in Chapters 3 and 4 of this mobility); and those that involve a movement away from
Handbook, respectively). However, adequate under- the health sector (intersectoral mobility) or to another
standing of health workforce dynamics requires an country (international out-migration) are considered as
analysis not only of new entries and current stock, but “exits” from the health workforce. These dynamics are
also of the flows within and out of the health workforce. summarized in Figure 5.1.
In this chapter, a framework is provided for identifying
and analysing major health workforce transitions and
49
Figure 5.1 Transitions within and exits from the health workforce: a framework for analysis
Permanent or temporary outflow due to death,

retirement, morbidity, unemployment, other
Other Health Other Health

sectors sector sectors sector
International migrants
Country I Country II
Transition within the health sector Exit from the health sector
Transitions within the national health labour market may pattern transitions – such as from full-time to part-time
involve changes in work patterns, places and positions. positions – may be prompted by changes in individ-
The flow of health workers from rural and remote areas ual circumstances or overall economic conditions in
to urban and more affluent regions, known as geo- the country. Monitoring such trends is critical to under-
graphical transition, is probably the most common type standing HRH dynamics within and across occupations
of transition. Also of importance is movement of work- and the possible implications for overall workforce skills
ers across sectors (for example from the public to the mix and capacity for provision of quality services.
private sector) or from one type of service delivery point
to another (for example from a primary-level health- At the same time, it is important to keep in mind that
care centre to a tertiary-level hospital). Given the social such transitions are not necessarily mutually exclusive;
and political importance of ensuring universal access a single move could actually combine different types
to essential and affordable health services, notably of transitions. For example, the move of a nurse prac-
through primary health-care strengthening, monitoring titioner working in a public sector health centre in a
the movement of workers away from public health-care rural area to a job as manager of nursing care services
facilities is a concern for decision-makers and stake- in a private hospital in the capital city simultaneously
holders in many low- and middle-income countries. involves occupational, sectoral, institutional and geo-
graphical transition.
Another type of transition that needs to be considered
is occupational transition: health workers may move With regard to exits from the national health system, these
from one health occupation to another, or to a non- include movements that are either voluntary or invol-
health occupation within the health sector. Often this is untary, permanent or temporary in nature. Temporary
a reflection of career progression, for instance a nurs- exits are categorized as those where the health worker
ing care provider being promoted to nursing services is assumed to have a chance to re-enter the system at
manager. An example that has attracted considera- some future point, usually in the shorter term. Maternity
ble interest in recent years, as mentioned in Chapter or family care leave, sickness or other emergency leave,
4, is the phenomenon observed in the Philippines of unemployment and return to studies are among the most
doctors retraining as nurses in order to facilitate their frequent causes of temporary exit from the economically
chances of international migration. In some contexts, productive workforce. Retirement, death, work-limiting
there is concern over the exodus of health profession- chronic disability and international out-migration are
als previously providing direct patient care (usually in examples of “permanent” exits (although the possibility
lower-paying jobs in public sector facilities) to project remains that a worker who has retired early or migrated
management jobs in the expanding donor-supported abroad may still opt to eventually return to the national
nongovernmental sector (5, 6). Other types of work health labour market).
50
Monitoring health workforce transitions and exits
5.3 Indicators and stayed in their public sector job over the same period);
or by taking snapshots of the public–private distribution
measurement strategies
of clinical staff at two points five years apart. In other
Measuring and monitoring health workforce transitions words, when measuring transitions and exits, either
and exits remains challenging for a number of reasons, flow-based or stock-based indicators can be used. The
including scarcity of the required data (almost no coun- former are usually expressed as rates, while the latter
try has reliable data on international out-migration, for are expressed as proportions or ratios.1
one); underuse of available data; lack of disaggrega-
tion (information sources often combine transition and The choice of whether to examine flow-based or
exit factors, limiting the ability to estimate separate indi- stock-based estimates generally depends on the
ces for each indicator); and lack of standardization of nature of the underlying information source. As such,
measurement techniques (different methods can be understanding the full complexity of exit and transi-
used to produce different indicator estimates, render- tion patterns requires analysis of data across multiple
ing comparisons across information sources difficult or sources. Table 5.1 presents an overview of potential
impossible). data sources that can be used for measuring various
indicators of workforce transitions and exits. As can
One inherent complexity in analysing workforce transi- be seen, no single source will be able to provide all
tions and exits is that most relevant indicators can be the information needed; if used in a complementary
measured in two different ways: following all moves that manner, a wide range of data collection techniques –
have taken place among a given group of individuals including population censuses, labour force and other
over the course of a specified time period (longitudinal
analysis); or observation of the current state of all indi- 1 The general rate of workforce transition and exit can
viduals in the target population at a single point in time be represented by the following algebraic formula:
Tij (t n )
(cross-sectional analysis). For instance, with regard to M ij (t n ) = , where M ij (t n ) represents the observed rate
sectoral movements, this can be assessed either by Wi
of transfer from origin state i to destination state j between
counting all moves of clinical personnel from the public time period t and t+n; Tij (t n ) ,is the observed number of moves
to private sector within the last five years (or, from a per- (exits or transitions) from state i to state j between period t
spective of monitoring retention, the number who have and t+n; and Wi is the stock of health workers in state i at
midpoint. n time
Table 5.1 Potential sources of data on health workforce transitions and exits
Indicators Potential data sources

Flow-based measures Stock-based measures
Transition indicators
tBetween sectors (public/private) tRoutine administrative records, including tPopulation census or labour
payroll records and health professional force survey (with questions
tBetween occupations registries on occupation, place of work,
tBetween areas/regions working hours)
tSpecial HRH assessments (longitudinal
tBetween institutions design or retrospective questions) tHealth facility assessment
(module on staffing)
tFull time to part time (or vice versa)
Exit indicators
tUnemployment/loss of job tRoutine administrative records, including tPopulation census or labour
payroll records, health professional force survey (with questions
tLeave for further education and training registries, social security records on education, labour force
tMaternity or family care leave activity, occupation, reasons for
tSpecial HRH assessments (longitudinal inactivity)
tSickness or other emergency leave design or retrospective questions)
tHealth facility assessment
tInternational out-migration tVital registration (module on staffing)
tRetirement
tWork-limiting chronic disability
tDeath
51
household surveys, health facility assessments, routine health workers moving from a given region to another
administrative records and specialized quantitative or over the specific time period. Flow estimates can
qualitative HRH studies – can provide a more complete also be obtained indirectly from a census or survey if
picture of the dynamics. A general review of standard respondents are asked about their place of residence
statistical sources and their strengths and limitations five or ten years earlier (assuming they held the same
for HRH analysis is presented in Chapter 3. occupation at that previous time).
To give an example, one way of measuring occupational Ideally, international migration should be measured
transition using a stock-based approach is through adopting the same approach. In reality, few coun-
using labour force survey data to calculate the propor- tries have accurate and timely data on the numbers of
tion of individuals in the country with education and nationals living abroad or leaving the country. Concerns
training in a given health field who are currently work- about the adverse impact of the migration of health pro-
ing in a different occupation (irrespective of when they fessionals, especially from poorer to richer countries,
moved from qualification for one occupation to practice have thrust the issue to the forefront of the global health
in another). Measuring the same dynamic using a flow- and development policy agenda in recent years (1).
based approach – for example to calculate an annual However, the evidence needed to monitor and eval-
rate of movement from one health occupation to another uate the phenomenon remains weak or non-existent
– can be accomplished using health professional reg- (7, 8). Most available analyses are excessively reliant
istries, provided they are continuously updated to on indirect quotations, largely based on extrapolation
reflect current work activities. Specially designed HRH of measures compiled and disseminated in destina-
assessments for collecting job histories and trajecto- tion countries: (i) census-based estimates of lifetime
ries among health workers, either by reinterviewing the migration (by country of birth of individuals practising
same cohort at periodic intervals over an extended a health occupation in the destination country at the
time or by singular in-depth interviews with questions time of enumeration, regardless of place of education);
on labour activities at some point in the past, also allow (ii) registry-based estimates of foreign-trained workers
for estimates of occupational flows. newly obtaining professional licensure (which do not
count skilled workers who fail to satisfy national practice
Similarly, unemployment among skilled health work- regulations); and (iii) numbers of residency or work per-
ers can be expressed in different ways depending on mits issued to foreigners according to the self-reported
the data source. Population survey or census data on occupation held in their country of origin (regardless
education, labour force participation and reasons for of eventual work activity in the destination country). In
inactivity can be used for stock estimates (for example this context, ensuring comparability of such measures
the number of currently unemployed individuals with – including occupational definitions, education equiv-
education and training leading to a health occupation, alencies and professional practice regulations across
relative to the total number of employed health work- the source and destination countries – is particularly
ers of the same occupation). Routine data from health imperative.
professional registries can potentially offer information
on numbers of qualified health workers who were ever Another challenging area is measuring mortality among
unemployed over the past year. Health facility assess- health workers: even in countries where vital registration
ments with modules on staffing levels and patterns can coverage is very high or complete, occupation-specific
help shed light on the magnitude of the problem by tal- mortality data are rarely tabulated and disseminated.
lying the numbers of health service providers who lost Again, most available measures use indirect estimation
their jobs in the previous year. Such results can then be techniques. This includes examination of facility staff-
analysed within the context a monitoring and evalua- ing records for deaths while in employment, or the use
tion framework, for example to gauge trends over time, of model life tables applying age- and sex-specific sur-
across regions or by occupational group. vival ratios against the demographic distribution of the
active health workforce to estimate numbers of prema-
Geographical transitions within a country can also be ture deaths.
assessed with a number of measures. A stock-based
assessment drawing on periodic census data could By contrast, measuring workforce exits due to retire-
compare, for instance, the density of health workers in ment is relatively straightforward. Estimates can often
a given region at the time of enumeration against the be obtained from payroll or social security records.
same density enumerated in the previous census. A In the absence of reliable data from administrative
related flow-based measure could be obtained through sources, population-based census and survey tools
administrative records designed to allow tracking of are another option if they include data on educational
individuals, enabling measurement of the number of attainment by field (or on previous occupation) and
52
reasons for labour force inactivity. In many cases it is surveys in six African countries revealed substantially
possible to use the legal age of retirement as a proxy high proportions of health professionals – from 26% in
measure, and assume that all health workers remaining Uganda to 68% in Zimbabwe – declaring an intention
alive and in the country will retire at that age. to emigrate, mostly to high-income countries of Europe
and North America but also to some other African coun-
tries (12). Another indirect measure is through numbers
5.4 Illustrative analyses of verifications of professional licensure with national
regulatory bodies. For instance, prospective foreign
In this section, illustrative analyses of different types of employers of nurses seeking employment abroad may
workforce transitions and exits are presented from var- request verification of licensure with the nursing coun-
ious contexts, and using different methodologies and cil in the country of origin. In Kenya, such data from the
data sources as described in the previous section. national Nursing Council pointed to the United States
of America as one of the main intended host countries
5.4.1 Transitions within the for Kenyan nurses looking to practise abroad (13). In
national health labour market both of these cases, while useful for weighing general
trends, the data refer only to migration intentions, and
As mentioned earlier, different approaches can be
do not confirm that a health worker has actually (or will
used to collate information on transitions within the
ever have) out-migrated.
health workforce, largely depending on data availa-
bility, relevance and quality. An example of sectoral
The main means of assessing levels of international
transitions can be gleaned from a special HRH survey
migration is through examining data in the destination
in Sri Lanka, which included interviews among a sam-
country according to the migrant worker’s country of
ple of facility-based health workers across the country
origin (in terms of birth, citizenship or professional edu-
(9). According to retrospective questions on place of
cation). As illustration, data from the United Kingdom
work prior to the current location, the large majority
Nursing and Midwifery Council (14) allow estimation
(96%) of hospital-based health workers in the public
of trends in nursing exits from source countries (Table
sector also listed a government hospital as their previ-
5.2). The data reveal major changes between 1998
ous workplace. Among staff in private hospitals, 62% of
and 2007 in the distribution of foreign-trained nurses
those who had previously worked elsewhere reported
in the United Kingdom: from mostly high-income
having moved from a government hospital. Altogether,
countries at the beginning of the period of observa-
fewer than 0.5% of facility-based workers reported their
tion (Australia and New Zealand), to more and more
previous place of work as outside the health sector.
nurses from low- and middle-income countries (nota-
bly India and the Philippines). Some countries, such as
In another example based on a special HRH survey,
Nepal and Pakistan, which previously had no or very lit-
higher levels of occupational transition – that is, move-
tle out-migration towards the United Kingdom, are now
ment from one occupation to another while remaining
among the main source countries. Migration levels from
employed in the health sector – were observed among
selected African countries appear to have peaked in
medical practitioners compared to nursing personnel
around 2002/03, especially for South Africa. It may fur-
in Lesotho. The turnover rate was highest among med-
ther be noted that South Africa itself is home to many
ical specialists (16.7%) and considerably lower among
foreign-qualified health professionals: data from the
nursing officers (4.4%) (10).
Health Professions Council of South Africa indicate
that one quarter (24%) of registered doctors had been
Another potential means to measure and monitor
trained in a different country (15).
occupational transitions is by using professional and
academic registries. In the Philippines (probably the
Combining migration data across multiple destina-
largest exporter of nurses worldwide), it was estimated
tion countries can help present a broader picture. In
that of students who took the national exams for nurs-
this context, population census data are often valu-
ing licensure, more than 4000 were previously doctors,
able, as census measurement tools tend to be more
a figure that represents about 10% of the total number
standardized, allowing for international comparisons.
of doctors in the country (11).
Figure 5.2 presents selected findings using merged
census data on foreign-born health professionals liv-
5.4.2 International migration ing in 24 high-income countries of the Organisation for
International outflows of health workers are rarely meas- Economic Co-operation and Development (OECD),
ured directly at the country level. Some inferences including Australia, the United Kingdom and the United
can be made, for example through qualitative stud- States of America. While migration patterns may vary
ies on migration intentions. Findings from special HRH significantly from one country to another, one key
53
Table 5.2 Annual numbers of overseas-trained nurses obtaining national licensure to practise in the
United Kingdom, 1998–2007 (main countries of origin outside the European Economic Area)
Country Year professional licensure obtained in the United Kingdom

of nursing
education 1998/99 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07
India 30 96 289 994 1830 3073 3690 3551 2436

Philippines 52 1052 3396 7235 5593 4338 2521 1541 673
Australia 1335 1209 1046 1342 920 1326 981 751 299
Nigeria 179 208 347 432 509 511 466 381 258
Pakistan 3 13 44 207 172 140 205 200 154
Nepal 0 0 0 0 71 43 73 75 148
Zimbabwe 52 221 382 473 485 391 311 161 90
China 0 0 0 0 0 0 60 66 80
New Zealand 527 461 393 443 282 348 289 215 74
Ghana 40 74 140 195 254 354 272 154 66
Zambia 15 40 88 183 133 169 162 110 53
South Africa 599 1460 1086 2114 1368 1639 933 378 39
Kenya 19 29 50 155 152 146 99 41 37
Canada 196 130 89 79 52 89 88 75 31
Source: United Kingdom Nursing and Midwifery Council (14).
Figure 5.2 Foreign-born nurses and doctors enumerated in 24 OECD countries by main countries of
origin (population census data, around 2000)
Nurses Doctors
Poland Cuba
Trinidad and Tobago Morocco
Former Soviet Union Egypt
China South Africa
Mexico Viet Nam
Former Yugoslavia Iran
Haiti Canada
Pakistan
Nigeria
Algeria
Ireland
Former Soviet Union
Canada
China
Jamaica Philippines
Germany United Kingdom
United Kingdom Germany
Philippines India
0 20,000 40,000 60,000 80,000 100,000 120,000 0 10,000 20,000 30,000 40,000 50,000 60,000
Source: Dumont and Zurn (8).
finding stands out: nurses born in the Philippines (some help gain some further sense of the magnitude of inter-
110 000 in number) and doctors born in India (56 000) national emigration. Such an analysis is depicted in
represent a major part of the immigrant health work- Figure 5.3, where census data on foreign-born doctors
force in OECD countries (about 15% of the total stock residing in OECD countries at the time of enumera-
each) (8). tion are compared to official statistics on the domestic
workforce for those same origin countries. Some coun-
Extending such analysis by combining data on in- tries – in particular, small island developing nations
migrants in destination countries with information on of the Caribbean and South Pacific and certain sub-
the active health workforce in countries of origin can Saharan African countries with severe shortages of
54
Figure 5.3 Estimated lifetime emigration rate of Table 5.3 Estimates of annual losses due to
physicians born in selected non-OECD countries mortality under age 60 among health workers
and working in OECD countries at the time of in selected countries of the WHO Africa Region,
the census based on life table analysis
Country Premature death rate

Liberia
per 1000 workers
Trinidad and Tobago
United Republic of Tanzania Physicians Nursing and
Sierra Leone midwifery
Fiji personnel
Angola Central African 25 21
Mozambique Republic
Guyana
Grenada Congo, Democratic 23 19
Antigua and Barbuda Republic
0% 25% 50% 75% 100% Côte d’Ivoire 25 22
Source: Dumont and Zurn (8). Ethiopia 23 20

Kenya 23 23
Liberia 24 20
Madagascar 21 20
medical personnel – appear to be disproportionably Rwanda 25 19
affected by out-migration. These are countries with
estimated lifetime emigration rates above 50% (which Sierra Leone 26 22
means that there are as many doctors born in these Uganda 26 22
countries working in the OECD region as there are doc-
United Republic of 24 22
tors working in the home country) (8).
Tanzania
Zambia 28 22
5.4.3 Workforce exits due to mortality
Total 24 21
In several southern African countries, especially where
HIV prevalence is high, death is emerging as one of the Source: From the authors.
most important causes of exits from the health work-
force (16). Effects include both the permanent loss of
individual workers and temporary increases in staff
absenteeism to attend funerals. However, systematic
data collection on this issue is lacking: few countries example of the application of such techniques for 12
routinely compile and disseminate occupation-specific African countries is presented in Table 5.3. The results
mortality data via their vital registration system (usually were obtained by dividing the projected annual number
the main source of information on deaths and births in of premature deaths among active health workers
a country). (based on age- and sex-specific mortality quotients
extracted from national life tables) by the baseline total
Measuring and monitoring the extent and effects of number of health workers (according to official work-
premature health worker mortality often calls for special force statistics compiled by WHO). Premature deaths
tools and approaches. In one pilot study conducted in were defined as those occurring under age 60. Overall,
Zambia, time trends were estimated at two hospitals each year these countries are expected to lose about
based on the numbers of archived death certificates 2% of their medical, nursing and midwifery workforce to
for female nurses compared to the numbers of person- premature mortality. As could be expected (given the
years of service (17). Results suggested that mortality underlying assumption that mortality patterns among
among the female nursing workforce increased more health workers follow those of the total population),
than tenfold between 1980 and 1991, from 2.0 to 26.7 somewhat lower estimated death rates among nursing
per 1000. The observed increase was largely attributed and midwifery personnel compared to doctors would
to HIV. be the reflection of higher proportions among the
former of women, for whom age-specific death rates
Often, workforce mortality must be estimated indirectly tend to be lower compared to their male counterparts.
using model life tables and other demographic and
epidemiological projection techniques. An illustrative
55
5.4.4 Workforce exits due to workforce for other reasons, either in the short or the
retirement and other reasons long term. With regard to temporary exits, special
absenteeism surveys including unannounced visits
Generally, three types of data are used to shed light
may capture data on proportions of facility-based prac-
on health worker retirement: (i) routinely compiled num-
titioners who were not at their assigned post. In one
bers of individuals retiring from the health workforce
such application in Bangladesh, over one third (35%)
as recorded in the payroll, social security records or
of health workers were found to be temporarily absent
professional registries; (ii) counts of retirees among
on the day of visit (for explained or unexplained rea-
those with the educational background to qualify for a
sons). Proximity of the worker’s residence to the health
health occupation as reported in a population census
facility, opportunity cost of the worker’s time and other
or survey; and (iii) in the absence of direct measures,
indicators of general socioeconomic conditions (road
qualitative survey data on retirement intentions among
access, rural electrification) were identified as the main
health workers. Retirement data plotted against the
correlates of absenteeism patterns (19).
age distribution of the active workforce are commonly
used in workforce projection models to estimate future
It is possible that some recorded long-term workforce
supply.
exits are due to change of occupation, notably to one
outside the health sector. Labour force surveys – which
Regardless of the information source, one complex-
include data across all economic branches in the
ity to monitoring trends in workforce retirement is that
national economy – can help shed some light on this
there is no universal definition of what “retirement” actu-
point, although the ability to draw comparisons across
ally means. Some analyses may consider retirement as
countries or over time will depend on the question-
the period immediately following gainful employment,
naire and sampling design of the original source. For
or the period of life above a certain age. In a case from
instance, an analysis of labour force survey data from
Canada, where many parts of the country do not have
four countries showed wide cross-national differences
mandatory retirement, the national medical association
in the proportion of respondents reporting an occupa-
records any physician who exits the medical workforce
tion in medicine or nursing but not working at a health
as “retired”, regardless of age and cause (18). In this
service delivery point (and thus who might not be cap-
context, as seen in Figure 5.4, even workers as young
tured in facility- or payroll-based data sources) (Figure
as 30 are considered to have retired.
5.5). The highest proportion was found in the Denmark
Most HRH databases do not allow differentiation sample, which included those who were unemployed
between retirement, death and departure from the or had returned to school at the time of the survey (20).
Figure 5.4 Retirement rate among physicians Figure 5.5 Proportion of survey respondents
by age group according to the National Medical reporting a health occupation but not working
Association registry, Canada, 2005 in the health services industry at the time of
interview, selected countries
4
40%
Physicians
3
Nursing and midwifery professionals
30%
Retirement rate
2
20%
1 10%
0 0%
30 or less 31–40 41–50 51–60 61–65 66–70 71 or Denmark Netherlands United Kingdom United States
older
Age group
Source: Pong, Lemire and Tepper (18). Source: Gupta et al. (20).
56
5.5 Concluding remarks: Even data on health workforce retirement, increasingly

becoming a major issue given global patterns of work-
implications for policy
force ageing, are often deficient. Relying on a given
and planning age (for example statutory retirement) to project num-
This chapter has reviewed the current state of knowl- bers of retirees is likely to be inadequate, as differences
edge on measuring and monitoring health workforce in actual age at retirement are often observed across
transitions and exits for policy and planning purposes. occupations or by sex. Having a good sense of retire-
Understanding these workforce dynamics could help ment rates is valuable to plan future health workforce
in identifying imbalances in health worker distribution supply, and can also be used to prepare flexible work
within and across countries, and in implementing reten- policies encouraging delayed retirement. For instance,
tion policies that encourage workers to stay (or return a flexible-retirement initiative in the United Kingdom
to) where their skills and services are needed most. It enabled physicians nearing retirement to move into
is also being increasingly recognized that the key to part-time work while preserving pension entitlements
maintaining a sufficient workforce for achieving health (22).
systems goals is to educate, recruit and retain young
health practitioners, while also reinvesting in the mature Recruiting back individuals trained in health services
workforce. In this context, having and making use of delivery but either working outside the health sector
appropriate data from different sources to monitor and or economically inactive can represent an attractive
address HRH challenges across the working lifespan option to increase health system capacity. For that pur-
is critical. pose, it is crucial to determine the size of the potential
pool of individuals concerned in order to evaluate the
Information on past and projected movements of the opportunity and potential impact of such an initiative.
workforce is needed to make future workforce sup- In the United States of America, as of 2004, almost
ply projections, necessary as a basis for formulation 17% (or some 488 000) of surveyed nurses included
of evidence-based HRH planning and rationalized in the national professional registry were not employed
decision-making. Given their nature, few types of tran- in nursing (23). Although many were older, and thus
sitions and exits (aside perhaps from retirement) can unlikely to return to active nursing service, the number
be accurately predicted; however, all types need to be below the age of 50 totalled approximately 160 000
accounted and planned for in national HRH develop- potentially employable nurses. Considering that hospi-
ment strategies. In most countries, transitions within tals of the United States have reported some 116 000
the health workforce – such as change of employ- vacant nursing positions (24), policies and strategies to
ment sector or rural–urban migration while remaining in attract back qualified nurses could have high returns.
health services – are likely to be an important workforce
dynamic; however, timely and reliable data tend to be International migration of health professionals from
scarce. Often, policy- and decision-makers must rely low- and middle-income countries to wealthier coun-
on information from special HRH assessments (ad hoc tries is another issue of increasing global attention.
in nature) or periodic assessments based on stock- Doctors and nurses represent a small proportion of all
based estimates of the current situation observed at highly skilled workers who migrate, but the HRH loss
two different points in time (and then making infer- for developing countries can mean that the capacity of
ences on the flows that actually occurred between the the health system to deliver health services equitably
two points). When it comes to monitoring exits from the is compromised (25). Many developed countries that
national health labour market, data must often be col- previously actively recruited health workers abroad as
lated from sources outside the health sector or even a solution to (real or perceived) workforce shortages at
outside the country altogether. home have now recognized the need to address the
adverse impacts in some of the main sending coun-
Notably lacking in many countries is systematic data tries, notably in sub-Saharan Africa (26, 27). One policy
collection on morbidity and mortality among health option for receiving countries is the formulation of ethi-
workers, crucial for monitoring workplace health and cal practices for the international recruitment of health
safety – often a major factor influencing attrition from workers. For example, in 2001 the United Kingdom’s
the health sector. Understanding the causes of work- Department of Health adopted a code of practice
force losses can help inform, for example, strategies for for employers in the national health system seeking,
preventing HIV and other diseases among health work- among other things, to prevent targeted recruitment
ers to reduce premature mortality in the longer-term, from developing countries experiencing severe short-
while providing appropriate treatment for health work- ages of health-care staff. While assessing the impact
ers who need it to enable them to work longer (21). of such a code is challenging, due to the numerous
factors driving labour migration, it can be noted that
57
nursing migration from Africa to the United Kingdom References

would seem to have declined substantially in the period
1. The world health report 2006: working together for
following its adoption (as previously illustrated in Table health. Geneva, World Health Organization, 2006
5.2). (http://www.who.int/whr/2006, accessed 10 January
2009).
At the same time, it should be recognized that health 2. Coomber B, Bariball L. Impact of job satisfaction
personnel movement and migration is bi-directional. components on intent to leave and turnover for
Health workers move from rich to poorer countries and hospital based nurses: a review of the literature.
from urban to rural areas for a variety of reasons and International Journal of Nursing Studies, 2003,
44(2):297–314.
through a number of mechanisms, although in much
smaller numbers – and even less well documented. In 3. Padarath A et al. Health personnel in southern
many instances, health personnel migrate abroad for Africa: confronting maldistribution and brain drain.
EQUINET Discussion Paper No. 3. Harare, Regional
a shorter period of time and return to their country of Network for Equity in Health in Southern Africa,
origin, which can be beneficial to source countries as Health Systems Trust and MEDACT, 2003 (http://
these workers return with more experience, skills and www.equinetafrica.org/bibl/docs/DIS3hres.pdf,
personal resources than when they left (3). Sending accessed 15 January 2009).
countries therefore need to consider policy options 4. Zurn P, Dolea C, Stilwell B. Nurse retention and
regarding whether outflow of health workers should recruitment: developing a motivated workforce.
be supported or encouraged (for example to stimulate ICN Issue Paper No. 4. Geneva, International
Council of Nurses, 2005 (http://www.icn.ch/global/
remittance income or to address oversupply relative to
Issue4Retention.pdf, accessed 15 January 2009).
national health labour market absorption capacity) –
or constrained or reduced (to counteract “brain drain”) 5. Davey G, Fekade D, Parry E. Must aid hinder
attempts to reach the Millennium Development
(26). In all cases, monitoring international flows is pre- Goals? Lancet, 2006, 367(9511):629–631.
requisite for evaluating policy effectiveness.
6. McCoy D et al. Salaries and incomes of health
workers in sub-Saharan Africa. Lancet, 2008,
The contribution of existing data sources and analyti- 371(9613):675–681.
cal approaches is growing significantly for monitoring
7. Diallo K. Data on the migration of health-care
health workforce transitions and exits, and for sup-
workers: sources, uses, and challenges. Bulletin of
porting HRH policy development at the national and the World Health Organization, 2004, 82(8):601–607
international levels; however, some areas need further (http://www.who.int/bulletin/volumes/82/8/601.pdf,
consideration. One of the main factors constraining in- accessed 15 January 2009).
depth analysis is the general lack of disaggregated data 8. Dumont JC, Zurn P. Immigrant health workers in
on the different types of workforce exits, which makes it OECD countries in the broader context of highly
difficult for decision-makers to address specific reten- skilled migration. In: International migration outlook:
SOPEMI 2007 edition. Paris, Organisation for
tion issues in the most pertinent manner. Strengthening
Economic Co-operation and Development, 2007
efforts for systematic data gathering and improved (http://www.oecd.org/dataoecd/22/32/41515701.pdf,
coordination among stakeholders in data collection accessed 15 January 2009).
and use across different sectors and countries should 9. Assessment of human resources for health: country
act as a catalyst for improving the availability, quality profile. Geneva, Ministry of Health of Sri Lanka and
and comparability of HRH data, thereby strengthening World Health Organization, 2002.
the evidence base needed to advocate policy options 10. Schwabe C, McGrath E, Lerotholi K. Health sector
and guide decision-making. human resources needs assessment. Silver Spring,
MD, Medical Care Development International, 2004.
11. Lorenzo FME, Galvez-Tan J, Icamina K, Javier L.
Nurse migration from a source country perspective:
Philippine country case study. Health Services
Research, 2007, 42(3p2):1406–1418.
12. Awases M et al. Migration of health professionals in
six countries: a synthesis report. Brazzaville, World
Health Organization Regional Office for Africa,
2004 (http://www.afro.who.int/hrh-observatory/
researchpolicies/migration_en.pdf, accessed 15
January 2009).
13. Riley PL et al. Developing a nursing database
system in Kenya. Health Services Research, 2007,
42(3):1389–1405.
58
14. Statistical analysis of the register: 1 April 2006 to 26. Buchan J. How can the migration of health service
31 March 2007. London, Nursing and Midwifery professionals be managed so as to reduce
Council, 2008 (http://www.nmc-uk.org/aArticle. any negative effects on supply? Copenhagen,
aspx?ArticleID=36, accessed 15 January 2009). WHO Regional Office for Europe and European
Observatory on Health Systems and Policies, 2008
15. South Africa yearbook 2002/03. Pretoria,
(http://www.euro.who.int/document/hsm/7_hsc08_
Government of South Africa Communication and
epb_10.pdf, accessed 19 January 2009).
Information System, 2003.
27. Robinson M, Clark P. Forging solutions to health
16. Cohen D. Human capital and the HIV epidemic
worker migration. Lancet, 2008, 371(9613):691–693.
in sub-Saharan Africa. ILO Programme on HIV/
AIDS and the World of Work, Working Paper No. 2.
Geneva, International Labour Organization, 2002
(http://www.ilo.org/public/english/protection/trav/
aids/publ/wp2_humancapital.pdf, accessed 15
January 2009).
17. Buve A et al. Mortality among female nurses in the
face of the AIDS epidemic: a pilot study in Zambia.
AIDS, 1994, 8(3):396.
18. Pong RW, Lemire F, Tepper J. Physician retirement
in Canada: what is known and what needs to be
done. Presented at the 10th International Medical
Workforce Conference, Vancouver, Canada, 20–24
March 2007 (http://www.cranhr.ca/pdf/10_retCAN.
19. Chaudhury N, Hammer JS. Ghost doctors:
absenteeism in rural Bangladeshi health facilities.
World Bank Economic Review, 2004, 18(3):423–441.
20. Gupta N et al. Assessing human resources for
health: what can be learned from labour force
surveys? Human Resources for Health, 2003,
1:5 (http://www.human-resources-health.com/
content/1/1/5, accessed 15 January 2009).
21. Simoens S, Hurst J. The supply of physician services
in OECD countries. Health Working Papers, No.
21. Paris, Organisation for Economic Co-operation
and Development, Directorate for Employment,
Labour and Social Affairs, 2006 (http://www.oecd.
org/dataoecd/27/22/35987490.pdf, accessed 15
January 2009).
22. Kinfu Y et al. The health worker shortage in
Africa: are enough physicians and nurses being
trained? Bulletin of the World Health Organization,
2009, 87(3):225–230 (http://www.who.int/bulletin/
volumes/87/3/08–051599.pdf, accessed 20
February 2009).
23. The registered nurse population: national sample
survey of registered nurses, March 2004 –
preliminary findings. Washington, DC, United States
Department of Health and Human Services, 2005.
24. Research and trends: health and hospital trends
2007 – AHA survey of hospital leaders. Washington,
DC, American Hospital Association, 2007 (http://
www.aha.org/aha/research-and-trends/health-and-
hospital-trends/2007.html, accessed 15 January
2009).
25. Stilwell B et al. Developing evidence-based
ethical policies on the migration of health workers:
conceptual and practical challenges. Human
Resources for Health, 2003, 1:8 (http://www.human-
resources-health.com/content/1/1/8), accessed 15
January 2009).
59
60
Part III:
MEASUREMENT
STRATEGIES AND
CASE STUDIES
61
Measuring expenditure on the
6 health workforce: concepts,

data sources and methods
PATRICIA HERNANDEZ, TESSA TAN-TORRES, DAVID B EVANS
6.1 Introduction decision-making. Tracking the financial resources con-

tributed to HRH can facilitate monitoring of resources
Payments to labour have been assessed as consti- to achieve the Millennium Development Goals, the
tuting the largest single expenditure item for national national poverty reduction strategy plan and other ini-
accounts and, where they have been measured, in tiatives. Disentangling HRH expenditure within existing
health services provision. However, in many countries, frameworks focusing on resource use helps provide
data on the extent and nature of expenditure on health that information. If implemented on a regular basis,
workers are not routinely available and information is accounting systems can track health labour expend-
often scattered across multiple sources. Where data iture trends, an essential element in HRH monitoring
collation occurs, it may only have partial coverage and and evaluation.
is rarely used for policy and planning. Timely acces-
sibility among decision-makers and stakeholders to Several lines of action are presented here, intended
information about expenditures on the health workforce to be a how-to guide for operational use by people
would require a systematic consolidation and harmo- actually monitoring these expenditures. Following this
nization effort, with modifications in the way data are introduction, the chapter offers a brief presentation on
recorded. The nature and intensity of the effort implied the purpose of the exercise and the core indicators pro-
would, of course, vary by country. posed. It moves to a description on how to begin to
construct and maintain a minimum database on HRH
Notwithstanding the weight of human resources for expenditure. Issues of data collection and use at coun-
health (HRH) in overall health expenditure, there is try level are discussed, enriched with case studies to
no full documentation of health accounting meth- illustrate various procedures and recommendations
ods specifically related to the health workforce. Other for enhancing comparability across countries and over
measurement approaches describe how to measure time.
expenditure on the labour force in general, notably
the provision of health-care services in the system of Recognizing that many countries are unlikely to
national accounts – that is, a conceptual framework develop sophisticated, integrated systems of data
that sets the international statistical standard for the collection and collation in the short run, this chapter
measurement of the market economy – and in meth- suggests how the different sources of data typically
ods used to guide the measurement of expenditures on found in countries can be used to establish the order
government-funded health services. of magnitude of expenditure on HRH. The approach
proposed here does not posses the attributes of a fully
The construction of a comprehensive, reliable and inte- fledged HRH account linked to official health accounts,
grated system of HRH expenditure measures typically strictly covering the same boundaries. It is offered as a
requires a compilation of data from routine administra- short-term solution while waiting for some of the main
tive records and periodic surveys. Some countries are internationally agreed-upon estimating methods to be
beginning to store all labour-related surveys that have updated and refined – notably the updates, expected
been undertaken in a single repository. In many cases, to be completed by 2012, of the current versions of the
additional surveys are required to complement existing System of National Accounts 1993 (henceforth referred
sources in order to cover the field adequately. to as SNA93) (1), the System of Health Accounts (or
SHA1.0) (2) and the Guide to producing national health
The main aim of this chapter is to encourage a
accounts (3).
greater number of countries to monitor expendi-
ture on human resources in health systems to inform
63
6.2 What should be measured Countries lacking reliable private sector data can use
the general government or public sector data as an
Before addressing measurement issues and data entry point. General government expenditure refers to
sources, it is important to be clear about what is to expenditures incurred by central, state or regional, and
be measured and why. Choices of indicators on HRH local government authorities, as well as social security
expenditure should be driven by policy needs, although schemes and non-profit institutions that are controlled
it is also important to take into account feasibility and and mainly financed by government units. Monitoring
costs of data collection and processing. public expenditure is related to the question of how
much funding is raised for HRH development and can
6.2.1 Defining a core set of indicators also be considered to reflect government commitment.
A first step is to define a desirable minimum set of
The sixth proposed indicator is meant to provide more
indicators. There are two goals: to offer a means for
detail by breaking down expenditure data into various
countries to develop a practical reporting system for
components that would be useful for policy, such as
their own policy purposes, and to facilitate compari-
place and sector of work, employment status or occu-
sons over time and across settings. Standardization
pational function.
and harmonization of information enables countries to
track the impact of changes, and allows opportunities
to learn from the experiences of other countries and 6.2.2 Expanded set of indicators
regions. Some governments may wish to track an additional set
of indicators to, for example, monitor equity and effi-
The proposed core set of indicators is outlined in Box ciency of HRH expenditure, detail labour expenditure
6.1. Six basic indicators are listed, referring to the total, for specific service areas or programmes, or identify
relative size and distribution of expenditure on health sources of cost escalation. Among a wide range of
workers. The total is proposed in absolute and per possibilities, often complemented with other types of
capita levels, specifying currency units useful for inter- data on health system performance, the most common
national comparisons (4, 5). might include the following:
Box 6.1 Proposed minimum set of indicators for monitoring expenditure on

human resources for health
1. HRH expenditure, total and per capita (in national currency units, in US dollars and in international dollars)a
2. Expenditure on HRH as a proportion of total expenditure on healthb
3. Expenditure on HRH as a proportion of gross domestic product or gross national income
4. Government expenditure on HRH as a proportion of general government expenditure on health
5. Government expenditure on HRH as a proportion of recurrent general government expenditure on health
6. Breakdown of HRH expenditure by:

a. place of work: hospitals, ambulatory centres, public health offices, etc.
b. sector: public, private for-profit, private not-for-profit
c. employment status: regular employees, self-employed workers
d. occupational function: health service providers (direct patient care), health system management and
support personnel.
a. Values for per capita expenditure are usually based on population estimates in the mid-year period. International dollars are
derived by dividing national currency units by an estimate of their purchasing power parity compared with the US dollar, i.e. a
measure that minimizes the consequences of differences in prices between countries. Definitions of selected health financing
terms can be found in the National health accounts section of the World Health Organization (WHO) Statistical Information
System (4).
b. For comparative purposes, data compiled by WHO by country on total expenditure on health, using health accounting figures
when available, can be freely accessed at WHO National health accounts (5).
64
Measuring expenditure on the health workforce: concepts, data sources and methods
t HRH expenditure by skill level and skill speciali- To capture such diversity, workforce size is often meas-
zation of health workers, for example physicians, ured both in terms of headcounts (physical persons)
nurses, midwives, pharmacists, community health and full-time equivalents (a measurement equal to one
workers, ambulance drivers; staff person working a full-time work schedule for one
t HRH expenditure by different service areas or types year) (see also Chapter 3 of this Handbook for more on
of health interventions, such as workers providing measuring workforce stock).
mental health services or attending deliveries;
t average earnings among health workers, i.e. hourly, Because country-specific and tool-specific data are
weekly or monthly income from wages, practice or often collected and classified in different ways, it is
business. useful to adopt an internationally standardized clas-
sification procedure to improve comparability. The
The last of these can be a particularly useful indicator relevant classifications for the purposes of harmonizing
for monitoring equity in the health workforce (for exam- data on human resources in health systems include the
ple gender equity). Its proposal here is consistent with International Standard Classification of Occupations
the International Labour Organization’s recommenda- (ISCO), the International Standard Classification of
tion that statistics on average earnings, as well as hours Education (ISCED), the International Standard Industrial
of work (useful for calculating full-time equivalents for Classification of All Economic Activities (ISIC), the
job positions), should be maintained and updated Central Product Classification (CPC), the Classification
regularly, covering all important categories of wage- of the Functions of Government (COFOG) and the
earners and salaried employees, including those in the Classification of the Outlays of Producers According to
health branch of the economy (6). Purpose (COPP) (7–12).1 For example, labour expend-
iture data by occupation should ideally be mapped
through the latest ISCO revision (with most health occu-
6.3 Approaches to measuring pations falling under sub-major groups 22, “health
professionals”, and 32, “health associate profession-
HRH expenditure als”). Health-care goods and services are classified
Information on total HRH expenditure, the most by CPC under group 931, “human health services”.
important component of the minimum set of indica- Relevant government expenditure or activity data
tors outlined in Box 6.1, is typically found in national are delineated in COFOG under division 07, “health”,
accounts and health accounts. These systems consist including services provided to individuals or on a col-
of an integrated set of macroeconomic accounts, bal- lective basis. COPP can be used to itemize expenditure
ance sheets and tables based on internationally agreed on human resource development, notably under class
concepts, definitions, classifications and account- 5.1, “outlays on education and training”, which includes
ing rules, which together provide a comprehensive vocational training and on-the-job training.
accounting framework within which data can be com-
piled and presented in a format that is designed for Standard descriptions of the concepts and methods of
purposes of analysis, decision-taking and policy-mak- HRH expenditure are available in the System of National
ing (1). This section describes the main models and Accounts (in particular, see SNA93 paragraphs 7.21–
ways data are compiled, in order to understand how 47 for approaches to the generation of original data)
they can be used and compared across contexts and and in the European System of Accounts (13, Chapter
over time. 8). The System of Health Accounts is another useful
resource, covering three core dimensions: health care
by function or service area, providers of health-care
6.3.1 Delineating expenditure
goods and services, and sources of funding (in par-
on the health workforce
ticular, Table 10 of SHA1.0 refers to “total employment
In general terms, expenditure on HRH is the prod- in health care industries”, covering both numbers of
uct of the number of health workers and their prices. employees and full-time equivalents). Expenditure on
Capturing the heterogeneity of the health labour market health workers is approached directly in the adaptation
requires consideration of many types of workers: people by WHO, World Bank and the United States Agency
who directly provide health services (including preven- for International Development (USAID ) (3) for low- and
tive, promotional, curative and rehabilitative services) middle-income countries – the economic classification
as well as administrators, suppliers and other support in that guide includes compensation of employees and
workers who help the health system function. The work- owners, and distinguishes resource costs on wages
force includes those who are salaried or self-employed, (code 1.1.1), social contributions (code 1.1.2) and
working full time or part time, having short-term or long-
term contracts, holding one job or multiple positions. 1 See Chapter 2 for further details on some of these
classifications.
65
non-wage labour income (code 1.1.3) – and then further that there is no unique data display format for HRH
expanded in an SHA data collection tool jointly devel- expenditures and, as such, analysts and decision-mak-
oped by the Organisation for Economic Co-operation ers need to carefully read the fine print attached to any
and Development (OECD), Eurostat and WHO (14). reported numbers (e.g. the metadata) to understand
Countries and stakeholders should find it useful to work how they can be used and interpreted.
with these complementary materials in the process of
building a specialized dataset on expenditure on the
6.3.2 Data requirements
health workforce.
and potential sources
The boundaries of “health” set the scope and content The estimation procedures used to derive country-level
of HRH expenditure data, and the results produced will figures in national accounts and health accounts inte-
differ somewhat depending on the measurement sys- grate volume and price data. Accounting involves the
tem used. Typically, in national accounts, the health use of a large mix of documentary sources and types
sector is defined as human health activities – an indus- of information, both monetary and non-monetary, and
trial division as classed by code 86 in the fourth or latest recurrent and one-off, including:
ISIC revision (or equivalent in national classifications) –
encapsulating only the people involved in the provision t surveys and censuses, for example labour force and
of health-care services. Other approaches focusing other household surveys, establishment surveys,
more on health accounting tend to adopt a broader economic and population censuses;
definition, also including other key actions related to t administrative records, for example budgetary
health, such as regulation and management of health records of government ministries, employment reg-
services delivery, provision of health-care goods and istries, social health insurance records, taxation
products, and, in some cases, complementary activi- files, earnings statistics, business and facility reg-
ties such as administration of health insurance. istries, registries of health professional regulatory
bodies, bookkeeping records of private facilities;
In this chapter, and as compatible with the System of t special administrative monitoring of labour and
Health Accounts, except when specified, the wider employment characteristics, such as sickness
range activities of the health system is referred to (Box absence, non-resident workers, seasonal workers;
6.2) (15). Although utilization of information available in t other information sources, such as ad hoc data col-
national health accounts or national accounts is recom- lection and processing activities, special analyses
mended to the extent possible (rather than a construct using complementary sources, extrapolation and
of own estimates based on the number of health work- other projection methods.
ers and their remuneration), it is important to remember
Box 6.2 Expenditure components under a health accounting approach

through classes of the International Standard Industrial Classification of All
Economic Activities (fourth revision)
a. Health services: division 86 “human health activities” (groups 861 “hospital activities”, 862 “medical
and dental practice” and 869 “other human health activities”); part of groups 871 “nursing care” and 881
“social work”; part of group 712 “laboratory testing and analysis”; and parts of divisions 49, 50 and 51
“transportation” (as related to patients)
b. Manufacturing and sale of medical goods: retail sale of pharmaceutical and medical goods (class
4772 – excluding toiletries); manufacture of medical and dental instruments and supplies (class 3250)
c. Other activities held by law or according to the culture and traditions of the country to contribute to the
restoration, maintenance or enhancement of human health, formal or informal, not specifically included
in ISIC, e.g. distribution of traditional, complementary and alternative medicines
d. Administration and planning, which are part of ISIC classes 8412 (public administration) and 8430
(compulsory social security).
Source: Poullier (15).
66
The collation, synthesis and analysis of these various of labour force surveys are usually too small to allow
types of data is generally the outcome of collaboration statistically valid inferences about specific branches of
among a wide range of stakeholders, including gov- economic activity. As such, labour force survey data
ernment ministries (health, labour, finance), central are usually complemented with data from administra-
statistical agencies, development partners, research tive and other sources (16).
and academic institutions, workers’ associations and
insurance agencies. New data collection should be Figure 6.1 displays a typical path the estimator of
undertaken only when the required information is not labour expenditure could consider (1). For illustra-
available elsewhere and sufficient resources have been tive purposes, the use of labour force surveys versus
secured to do it well. The initial challenge is to ensure other data sources in national accounting is described
all relevant records have been retrieved; for example, within a selected group of countries of OECD and the
remunerations and incentives for health workers paid European Union:
by special funds (for example international sources) or t Labour force surveys constitute the main source
entities as a secondary activity (for example occupa- of data in Australia, Canada, Cyprus, Estonia,
tional health services in industries) may be reported Hungary, Ireland, Lithuania, Switzerland and the
independently of government records on wages and United Kingdom.
salaries. t Labour force survey data are partially replaced with
administrative data and other recurrent surveys in
A standard estimation procedure for monitoring labour Bulgaria, Greece, Latvia, Portugal and Romania.
expenditure is recommended by SNA93. Relatively t Countries combining supply and demand data
minor adjustments have been developed for detailed include Austria, Denmark, Finland, Germany, Italy,
entries under health and according to country-spe- Malta, Norway, Slovakia, Spain and Sweden.
cific data characteristics. For example, sample sizes
Figure 6.1 Overview of the estimation process for measuring labour expenditure in the System of
National Accounts
Population
Employed Unemployed Not in the labour force

Week preceding Week preceding Week preceding
Annual average Annual average Annual average
Single job Multiple jobs
Add jobs of persons not registered but working in

resident enterprises. Subtract jobs of persons
included but working in non-resident enterprises
Jobs. Week preceding. Annual average.
Employee Self-employment
Total hours worked: annual total, excluding
sick and other leave and holidays
Compensation of
Divided by employees
Full time annual hours, excluding holidays and sick and

other leave, separately for each job group then summed Deflated by indices of
compensation per job
equals
Full-time equivalent employment = Employee labour input at

Number of full-time equivalent jobs, annual average constant compensation
Source: SNA93 (Figure 17.1: Population and labour concepts) (1).
67
Figure 6.2 Information on the labour market in the Netherlands’ national accounts
Self-employed income
Government
registries
Wage costs
Wages
Enterprises &
institutions Economic activity
Hours worked
Type of labour contract

Individuals
Data on persons
Source: van Polanen Petel (18).
t Only a minimal use of labour force surveys for order to ensure consistency across the various com-
national accounting is made by Belgium, the Czech ponents and types of data, HRH expenditure figures
Republic, France, Iceland, Japan, Luxembourg, should be verified against national accounts data,
the Netherlands, Mexico, Poland, Slovenia and the notably the earnings of workers in the supply and use
United States of America. tables (that is, tools used to check the consistency of
statistics on flows of goods and services on the princi-
Health accountants use the same techniques and data ple that the total supply of each product is equal to its
sources as national accountants, but limited to the total uses). A simplified overview is described in Figure
health labour force. Both share the same challenges of 6.4 as a “calculation square”, in which each box implies
multiple data sources and inconsistency across them. a specific process to identify the best data source and
When differences arise, a first step is to identify rea- the adjustments required (18).
sons for them and then to decide whether one is the
more suitable, or whether a composite is better. In Personnel in health services may also perform non-
contexts with many statistical sources, consistency is health activities; as such, an additional refinement may
obtained mainly by adjusting employment rather than be required if only the health-care component is to be
wages and salaries (17). The Netherlands illustrates the examined. Most general accounting rules should apply
diversity of aggregation levels and content of the data to these estimations, for example measurement based
sources required to estimate the expenditure on HRH on the accrual principle (i.e. payable and receivable),
(sketched in Figure 6.2) (18). not on a cash basis (i.e. received and paid) (see United
Nations Statistics Division (22) for an introduction to
Where countries move towards the expanded set of basic concepts in national accounting). Specifically,
indicators (as described in section 6.2.2 above), infor- the volume measurement of the workforce and its value
mation is needed on health workers’ characteristics should follow this rule. The original definition as pre-
(for example gender, education) as well as on labour sented in SNA93 is detailed in Box 6.4 (1).
activities, such as hours worked in the health system,
including overtime but subtracting hours of absen- Labour costs beyond remunerations of employees that
teeism due to vacation, sickness or other reason need to be considered include expenditure on recruit-
– dimensions for which routine data are rarely availa- ment, education and training, incentives for worker
ble at the national level. Two examples of the process retention and motivation, miscellaneous costs such as
to indirectly estimate hours worked in different contexts work clothes, and taxes on employment (23). Costs
are supplied in Figure 6.3 (19, 20) and in Box 6.3 (21). that are tax deductible can be measured through tax
records. With regard to (pre-service) education, direct
Additional efforts are typically required to treat special costs – including remunerations of educators – are
groups, such as interim labour and home-based per- treated outside the boundary of the health system in
sonal care workers, and to derive figures where there SHA1.0, but some guidelines are proposed for their
are no direct records (of hours worked, for example). In recording “below the line”. In-service training entails
68
Figure 6.3 Process to estimate hours worked in the Canadian national accounts
Industry accounts (SNA)
Benchmarks mainly from

labour force surveys
adjusted to SNA
Adjustments to industry accounts
SNA number of jobs x hours worked per job = volume of hours worked
Sources: Statistics Canada (19) and Maynard, Girard and Tanguay (20).
Figure 6.4 Calculation square

Box 6.3 Labour data estimation
criteria in the national accounts of
the Russian Federation Annual Number Total
earnings × of jobs = earnings
Hours worked = = =
Workplaces x average actual hours worked in
the accounting period Hourly Hourly
wage wage
Full-time equivalent employment =
Hours worked/average hours performed by full- × ×
time employees
Full-time equivalent = Paid hours Number

per job × of jobs = Paid hours
Number of workplaces in full-time equivalent
employment
Reference criteria: Source: van Polanen Petel (18).
40 work hours per week; 52 weeks per year –

4 weeks of leave
Lastly, in some contexts, complementary data collec-
Maximum total workable hours: 1920 tion or estimation procedures may be needed where
An adjustment is required to establish the main non-observed payments are likely to be significant:
job and the hours worked in additional jobs. unregistered (legal activities but deliberately concealed
from public authorities), informal (legal activities with a
Source: Surinov and Masakova (21). low level of organization with little or no division between
labour and capital as a factor of production) and ille-
gal (activities forbidden by law or which become illegal
when carried out by unauthorized persons). The labour
imputed method is frequently used to identify the need
both direct costs and indirect costs (for example train- for any such adjustment: a comparison of the volume
ees’ salaried work time). Clinical or hands-on training of work of the supply of labour (usually through labour
as part of health services delivery represents a joint force surveys) with estimates of demand obtained from
product that is also accounted for in HRH expenditure business or facility surveys. Other methods involve tri-
measures (at least in theory). angulation of various sources, the commodity flow
method and input–output comparisons (24).
69
Box 6.4 Defining remuneration of employees and self-employment income
Remuneration of employees
The remuneration of employees comprises the total compensation, in cash or in kind, payable by
enterprises to employees in return for work performed during the accounting period. Wages and salaries
as well as employer social contributions are included.
Wages and salaries of health employees include remuneration in-cash and an allowance for benefits
in-kind for health activities such as regular interval payments, piecework, overtime, night work, weekend
or other unsocial hours, allowances for working away from home, in disagreeable or hazardous
circumstances, as allowances linked to housing, travel or sickness benefits, ad hoc bonuses,
commissions, gratuities, and in-kind provision of goods and services not required to carry out their work,
such as meals and drinks, uniforms and transportation.
Social contributions paid for health employees involve actual or imputed payments to social schemes
to secure an employee’s entitlement to non-wage benefits. The valuation of social contributions includes
payments by employers to social security schemes or to private funded social insurance schemes
designed to secure social benefits for their employees; imputed social contributions by employers
providing unfunded social benefits are to be added. A statistical convention considers that employees
receive a gross compensation from which they pay their share of contributions to social protection
schemes, whether social security funds, private funded schemes or unfunded schemes.
Social contributions are monitored through administrative records. Imputed social contributions of
employers are estimated for unfunded social benefits paid by employers and correspond to the amounts
that would be needed to guarantee their right to social benefits. Taxes payable by an employer on the
wage and salary bill are excluded.
Self-employed income
Self-employment income refers to the independent health practitioners. After deducting compensation
of employees, taxes and subsidies from value added, the balancing item of the generation of income
account obtained is described as operating surplus or mixed income.
The operating surplus reflects the surplus or deficit accruing from production before taking account of any
interest, rent or similar charges payable on financial or tangible non-produced assets owned, borrowed
or rented by the enterprise. The gross operating surplus includes the returns of owned assets used in the
production process; these should be netted to isolate the remuneration component.
This component is called mixed income for unincorporated enterprises owned by members of households,
either individually or in partnership with others, in which the owners, or other members of their households,
may work without receiving any wage or salary. Their labour income is mostly an entrepreneurial income.
The mixed income contains an element of remuneration for work done by the owner, or other members
of the household, that cannot be separately identified from the return to the owner as entrepreneur. The
unincorporated enterprises owned by households that are not quasi-corporations are deemed to fall in
this category, except owner-occupiers in their capacity as producers of housing services for own final
consumption and households employing paid domestic staff, an activity that generates no surplus. The
mixed income is increasingly reported as an independent value.
t The concept of operating surplus or mixed income is not applied to measure the income of workers in
governments and not-for-profit enterprises.
t Household unincorporated enterprises who regularly sell most of their output should be treated as
market enterprises. Groups of households that engage in communal activities for their own individual
or community use should be treated as informal partnerships engaged in non-market production.
Households producing services are included when they occupy their own dwellings in their production,
and services produced by employed paid staff. The production of these services does not generate
mixed income. There is no labour input into the production of the services of owner-occupied dwellings
Continues…
70
so that any surplus arising is operating surplus. No labour input is assigned when measuring surplus
generated by employing paid staff (SNA93 4.148 to 4.150).
Employment relationship
The nature of the employment relationship has to be identified. An employer–employee relationship exists
when there is an agreement, formal or informal, between an enterprise and a person on a voluntary basis,
whereby work for an enterprise is contracted in exchange for a remuneration in cash or in kind, based on
time or product done. The self-employed, by definition, work for themselves (SNA93: paragraphs 7.23–24).
As applied, the concept excludes work not entitled to remuneration by members of a household within an
unincorporated enterprise owned by the same household.
Payments
The nature of payments should also be explicit and treated according to international agreements: wages
and salaries in cash should not include the reimbursement by employers of expenditures made by
employees in order to enable them to take up their jobs or to carry out their work, e.g.
t the reimbursement of travel, removal or related expenses made by employees when they take up
new jobs or are required by their employers to move their homes to different parts of the country or to
another country;
t the reimbursement of expenditures by employees on tools, equipment, special clothing or other items
that are needed exclusively, or primarily, to enable them to carry out their work.
The amounts reimbursed are treated as intermediate consumption by employers. For example, employees
required by their contract of employment to purchase tools, equipment, special clothing, etc., when they
are not fully reimbursed, the remaining expenses they incur should be deducted from the amounts they
receive in wages and salaries and the employers’ intermediate consumption increased accordingly.
Expenditures on items needed exclusively, or primarily, for work do not form part of household final
consumption expenditures, whether reimbursed or not.
Source: SNA93 (Chapter 7) (1).
6.4 Measurement frameworks means, whether labelled “health care” or not in national
statistics (2). The health accounting model comprises
and applications
a set of standard rules, tracking all resources enter-
As previously discussed, the most comprehensive ing the health system during a period, expressed in
and standardized measurement strategy for monitor- monetary terms. Health accounting uses existing data
ing HRH expenditure is within the systems of health structures, through data compilation and consolidation,
accounts and national accounts. These are built on var- to create new information, aiming to provide a consist-
ious types of data, both new and existing, and must ent picture (3).
typically undergo a consolidation and harmonization
effort as there are often at least some inconsistencies A health accounting system analysis involves three
across sources. A key consideration is to understand basic dimensions: financing, production and final use
the boundaries implied by different data sources, and – services purchased, services produced and serv-
what has been included and excluded. In this section, ices consumed, respectively. Data on HRH are meant
we outline the main frameworks used in HRH expend- to be included as a specific class in the “resource
iture monitoring and present some reality checks on cost” classification, intended to measure the cost of
their actual uses, along with illustrative examples from the resources involved in the production of health com-
selected countries. modities, as part of the production dimension (Figure
6.5) (25). Flows of resources are recorded in two-
dimensional tables showing the origin and the direction
6.4.1 Health accounts framework
of the transactions. The tables for HRH cross-classify
Health accounting is designed to measure all resource the flows by categories of health workers, by financ-
flows earmarked to provide health care or a substan- ing agent (for example government, private sector) or
tial amount of health status enhancement by medical
71
by type of health service or good that they contribute Figure 6.5 Resource flows in a health system: an
to producing (also called health functions). The data accounting representation
sources should usually link information on HRH by
place of work (hospital, health centre, etc.).
Financing
Pooling Purchasing/allocation
In reality, few countries collate and disseminate detailed
expenditure data on HRH, including who is paying for Sources and
burden sharing:
their work and in which services they are involved. purchasing
Health accounts rarely report expenditure on health schemes and
workers that allow the basic set of indicators of Box 6.1 payment flows
to be measured and monitored. As a result, the pol-

icy analysis of the aggregate spending flows and the
Provision
relative productivity of the system have been reduced. Cost of factors Providers flows
Tools such as the System of Health Accounts (2) and the
SHA guidelines project (26) invite countries to display
Delivery of health
the number of health workers, but typically no expendi- goods and
ture breakdown for HRH is spelled out. Other tools and services
resources also exist; for one, the Guide to producing
national health accounts (3) lists a resource cost clas-
sification detailing the uses made by providers of the Consumption/use
funds they capture, cross-classifiable with financing Health products Beneficiaries
agent purchases of these resources (paragraphs 5.19
Changes in level and distribution of health
and 5.20). A related manual on health spending meas-
Geopolitical subnational entities
urement by the Pan American Health Organization (27)
Demographic and socioeconomic characteristics
includes a cost structure, banking on SNA monitoring
Apparent health needs and interventions
of resource flows, intended to rely on available health
information systems and other (new and existing) data
sources to perform indirect calculations (for example Source: Adapted from Hernandez and Poullier (25).
paragraph 282). The joint data collection questionnaire
by OECD, Eurostat and WHO (14) includes HRH costs
used up by health-care delivery only as HRH by pro-
vider in a “memorandum” table.
employees and self-employed income (see the next
In a resource cost table listing the main expenditure section and also Box 6.4 for technical details on related
headings incurred in the production process, the accounting principles).
human factor emerges as an aggregate measure for
all employees (collated at institutional level), as well as
6.4.2 National accounts framework
non-salaried labour income. There is usually a greater
availability for public entities, allowing a display as the The system of national accounts describes the financial
sum of total income without a breakdown by type of flows across different components and dimensions of
worker. To date, few health accounts reports include a a market economy: production, income, consumption,
detailed cost table containing HRH, labelled by some accumulation and wealth and their interrelations. The
accountants as “line item” or “type of expenditure”. system is broken down into many different branches
Data are usually displayed cross-classified by provid- of which health is only one (identified under “human
ers or with financing agents.2 health and social work activities” as one of 21 branches
in ISIC). The boundaries and rules of the accounts
Illustrative examples of (simplified) data displays from framework allow the generation of a set of indica-
two national health accounts are presented in Tables tors, reported in core tables, showing the size of the
6.1 and 6.2 (28, 29). The first, from Mexico, shows dis- health branch compared to other social and economic
tribution of payments to providers under the column branches and to the economy as a whole. Income
“personal services”. The second, from Peru, presents accounts report the wages and salaries of employees
more information on the institutional breakdown across in each branch (primary income distribution tables) and
the various health system actors. The “value added” related payments for social contributions (secondary
components allow identification of remunerations of income distribution tables); income for self-employed
health workers is ideally recorded as gross operat-
2 In health accounting wording: provision x resource cost (HP x ing surplus and mixed income. Increasingly, national
RC) table and/or financing x resource cost (HF x RC) table. reports display an independent mixed income and also
72
Table 6.1 Percentage distribution of expenditure by type of health-care agent, Mexico health accounts, 1995
Institution Type of expenditure (%) Total

Personal General Supplies Infrastructurea Not specified
services services
Health social insurance 44.4 39.4 12.1 3.3 0.9 100
Tax funded services 65.8 9.5 9.3 6.1 9.3 100
Private medical insurance 22.6 25.7 43.9 7.9 – 100
Private medical services 19.3 22.0 51.9 6.7 – 100
a. Expenditure on infrastructure is included in this table but in search of a tri-axial accrual approach, a two-tier approach involves only
current spending.
Source: Adapted from Fundación Mexicana para la Salud (28).
Table 6.2 Percentage distribution of expenditure by main providers, Peru health accounts, 2000
Components Public (%) Private (%)

Ministry of Health social For profit Not-for-profit
Health insurance
Intermediate consumption 40.5 41.5 32.7 55.0
Medical inputs and pharmaceuticals 11.2 21.1 8.0 13.8
Non-medical goods and services 29.3 20.3 24.7 41.2
Value added 48.5 50.4 61.8 43.6
Remunerations 45.1 45.3 12.3 38
Professional services – – 13.6 –
Taxes 1.0 1.6 9.3 2.7
Depreciation 2.4 3.4 3.3 2.9
Operation surplus – – 23.3 0
Investment 8.4 8.1 5.5 1.4
Transfers to community bodies 2.6 – – –
Total 100 100 100 100
Source: Adapted from Ministerio de Salud del Perú (29).
net values (capital consumption is deducted to obtain agencies (as well as related research and methodologi-
net values). cal resources), notably OECD (31), Eurostat (32) and the
United Nations Statistics Division (33).3 Remuneration
The recommended display of the income components to government health employees is also increasingly
is presented in Table 6.3 (30). When available, com- reported in the component of national accounts deal-
pensation of employees and net mixed income can be ing with general government expenditures (mapped to
drawn directly from this display for decision-making COFOG division 07) (33, 35). Although SNA93 recom-
purposes. mends more complex breakdowns (tables 18.2–18.4),
they are rarely produced.
For a large number of countries, statistical informa-
tion on HRH expenditure as obtained through national 3 It may be noted that, as of mid-2008, the tables produced in
many countries still correspond to the previous 1968 edition
accounts is displayed as part of the primary income
of the SNA manual, and so do not necessarily reflect the
distribution tables disseminated by various international evolving standard for national accounts (34).
73
Table 6.3 Remuneration components in the “use” table of the income account, System of National Accounts
S11 S12 S13 S14 S15 S1

Code Transactions and Non- Financial General Households Non-profit Total
balancing items financial corporations government institutions economy
corporations serving
households
D1 Compensation of 549 15 142 39 24 769
employees
D2 Taxes on production 235
and imports
D3 Subsidies -44
B2g Operating surplus, 254 55 44 92 7 452
gross
B3g Mixed income, gross 442 442
P61 Consumption of 137 10 30 32 3 212
fixed capital on gross
operating surplus
P62 Consumption of 10 10
fixed capital on gross
mixed income
B2n Operating surplus, net 117 45 14 60 4 240
B3n Mixed income, net 432 432
Source: System of National Accounts 2008, Table 7.1: The generation of income account – concise form – uses (concise version) (30).
Table 6.4 presents a simplified example of the tables growing number of countries of the African and Eastern
displayed in the national accounts in South Africa, Mediterranean regions. Data for monitoring general
including a selected list of industries from both col- government accounts are drawn from the various pub-
umns and rows from the supply and use table (36). The lic sector institutions. Although some countries map
column “health and social work” allows tracking of the health-related data to COFOG or ISIC, there is no sin-
purchases of products; further refinement is possible gle classification method for HRH expenditure used
by subdividing into the health services versus social consistently across, or even within, all countries. For
services components (not shown here). example, in the case of Portugal, differences in the
health sector universe across accounting exercises
As previously mentioned, HRH data taken from national have meant that, in particular, activities of legal forensic
accounts are generally limited to workers in services institutions have been excluded from the Portuguese
provision and thus may undervalue the total expected health accounts but included in the 2008 national
from a health accounts framework. Where national accounts.
accounts are the only information source, these esti-
mates should be supplemented to ensure reflection An ongoing push by the International Monetary Fund
of the wider range of health system activities, tracked to implement a standard classification for government
through a detailed display at the provider or product finance statistics deals with a breakdown for health,
level and mapped through ISIC and CPC, respectively including compensation of employees defined similarly
(see section 6.3.1 above). to the SNA93 approach – that is, in terms of wages and
salaries (in cash and in kind) plus social contributions
(actual and imputed) (37, 38).4 While more and more
6.4.3 General government accounts
The expansion of public finance monitoring and other 4 See International Monetary Fund Government finance
good governance approaches yields relatively com- statistics manual 2001, paragraph 4.26 and table 6.1:
prehensive information on public sector expenditures Economic classification of expense (37). Companion
materials and research on government finance statistics,
in many Asian and Latin American countries, and in a including treatment of HRH data in the annual questionnaire,
are available on the web site of the International Monetary
74
Table 6.4 National accounts supply and use table, South Africa, 2002 (millions Rand, partial display)
Industry
Use of products Total Taxes on Subsidies Total Total
supply at products on Agriculture Coal Gold General Health and Activities/ industry economy
purchasers’ products government social work services
prices
I1 I2 I3 I92 I93 I94 I95 I96
Agricultural products 102 613 3 296 11 12 146 58 77 52 298
Coal and lignite products 38 543 3 2 28 81 80 5 23 742
Petroleum products 88 240 3 503 548 234 1 582 958 288 51 295
Pharmaceutical products 36 256 1 979 11 66 5 269 6 670 – 22 857
Optical instruments 22 594 – 93 132 2 736 3 512 550 12 590
Electricity 39 269 490 493 2 130 434 347 537 26 628
Buildings 64 294 168 8 10 874 334 242 25 501
Transport services 137 197 4 917 7 338 185 2 731 931 744 79 089
Communications 102 299 27 77 86 3 403 2 353 2 220 76 669
Insurance services 133 078 1 232 – 41 1 467 609 1 745 88 447
Real estate services 125 178 16 33 34 1 639 4 306 1 544 57 045
Other business services 111 811 68 759 1 550 7 709 5 153 2 792 102 245
General government services 241 233 – – – 21 139 2 832 – 24 755
Health and social work 63 153 975 – – 2 497 80 496 6 097
Purchases by residents 19 601 –
Purchases by non-residents – –
Total uses at purchasers’ prices 2 961 897 41 816 19 590 17 353 82 359 35 565 23 342 1 453 588 –
Total gross value added/GDP 109 660 (4 762) 44 179 17 464 26 915 157 391 24 664 37 966 1 063 879 1 168 777
Compensation of employees 10 730 6 420 14 255 136 085 12 059 31 693 520 501 520 501
Taxes less subsidies 109 660 (4 762) (749) 328 461 2 260 802 763 21 543 126 441
Taxes on products 109 660 – 109 660
Subsidies on products (4 762) – (4 762)
Other taxes less subsidies on production (749) 328 461 2 260 802 763 21 543 21 543
Gross operating surplus/mixed income 34 198 10 716 12 199 19 046 11 803 5 510 521 835 521 835
75
Total output at basic prices 85 995 37 054 44 268 239 750 60 229 61 308 2 517 467
Source: Adapted from Statistics South Africa (36 ).

countries are issuing reports complying with this stand- payments and others); and balancing adjustments
ard, any attempt at comparative analysis requires (for example between supply and demand for health
careful checking of the actual boundaries of HRH labour).
expenditure; variations may occur in the treatment of
some key items that may affect labour dynamics, such In addition, different policy concerns may require a
as allowances and incentives. specific breakdown of HRH expenditure estimates or
additional analyses beyond standard health or national
accounting methods. Guidelines for health sub-
6.5 Summary, conclusions accounts to produce additional estimations are being
developed and tested by WHO (40, 41) in the follow-
and further developments ing areas:
This chapter has focused on the tools, methods and t sub-accounts for specific diseases and pro-
usual means of measuring expenditure on the health grammes (including malaria, reproductive health,
workforce as a component of overall monitoring and HIV and tuberculosis services);
evaluation of HRH strategies. It has been argued that t sub-accounts on child health programmes;
there are many advantages of an integrated estimation t regional health sub-accounts (distributional tables
of HRH expenditure within routine accounts estimates for specific regions within a country – particularly
– either the system of national accounts or, prefera- relevant for decentralized health systems);
bly, health accounts. There are certainly economies t sub-accounts for specific population groups (dis-
of scale and quality gains from a comprehensive and tributional tables in terms of expenditure allocation
harmonized process in collating and processing the according to characteristics of health service cli-
required data, and in identifying and filling informa- ents, such as by age, sex or other socioeconomic
tion gaps through complementary data collection and characteristics).
analyses.
At the same time, it must be recognized that there are
To complement or refine available estimates on HRH presently no specific guidelines for disentangling the
expenditure, close collaboration between health whole set of health labour cost estimations, and this
accountants and national accountants is advisa- across the stages of the working lifespan. Based on
ble. When the data are taken from national accounts measurement results from several countries, the need
records, the most important adjustment required is an has been recognized for further development in the
expansion of the boundaries of the health system to following areas in order to reach a comprehensive
reach concordance with health accounts boundaries. assessment of expenditure on HRH:
As such, there is an advantage in generating the data t boundary problems to distinguish between labour
as a bottom-up exercise (estimating each of the various resources in the health system devoted to health-
components independently and then adding them up); care services delivery versus other functions and
this allows greater flexibility to use the data in different activities;
ways according to specific needs. When full records t challenges regarding completeness and coverage
and updates of HRH expenditure are not readily avail- of information on HRH stock and distribution (for
able, a series of progressive steps can be taken to example lack of centralized database, lack of suffi-
move towards a comprehensive assessment: from sim- cient detail for disaggregation);
ply persons working in health services delivery to those t problems with consistency of HRH information
across the whole spectrum of health system activities, across various data sources, such as lack of stand-
from measuring just wages and salaries to inclusion of ard practices to classify workers by occupation
non-wage contributions, or from government expendi- and education; differences in practices to esti-
ture to all (internal and external) financial sources. mate full-time equivalents across health worker
groups; potential double-counting of health work-
Ensuring the quality, coherence, consistency and rele- ers (for example due to multiple qualifications or job
vance of the data – which may be drawn from multiple positions); and, given such differences, resulting
sources – requires continuous verification during com- difficulties in interpreting and comparing statistical
pilation, integration, adjustment and modelling (39). findings across and within countries.
This may include data validation and adjustments (to
correct biases, errors, incompleteness and disconti- In sum, strengthening of national information systems,
nuities); conceptual adjustments (for example to bring better use of available data and intensified efforts for
figures using definitions from national accounts in line harmonizing definitions and measurement units rele-
with those from health accounts); comprehensive- vant for health labour accounting should ensure that
ness adjustments (to cover hidden activities, informal investments in the health workforce are appropriately
76
monitored and evaluated. This is critical information that 11. Classification of the Functions of Government:
could help address many important policy questions, COFOG. New York, United Nations Statistics
Division, 2000 (http://unstats.un.org/unsd/class/
such as the costs of scaling up health interventions or
family/family2.asp?Cl=4, accessed 14 January
providing incentives for improving staff retention and 2009).
motivation to ensure high quality and efficiency of
12. Classification of the Outlays of Producers According
services.
to Purpose: COPP. New York, United Nations
Statistics Division, 2000 (http://unstats.un.org/unsd/
cr/registry/regcst.asp?Cl=7&Lg=1&Top=1, accessed
14 January 2009).
References
13. European System of Accounts (ESA95).
1. United Nations, World Bank, International Monetary
Luxembourg, Eurostat, European Commission, 1996
Fund, Organisation for Economic Co-operation
(http://circa.europa.eu/irc/dsis/nfaccount/info/data/
and Development, Commission of the European
ESA95/en/titelen.htm, accessed 26 February 2009).
Communities. System of National Accounts 1993.
New York, United Nations Statistics Division, 14. Joint OECD-Eurostat-WHO SHA data questionnaire.
1994 (http://unstats.un.org/unsd/sna1993/toctop. Organisation for Economic Co-operation and
asp?L1=5, accessed 26 February 2009). Development, Eurostat and World Health
Organization, 2008 (http://www.oecd.org/health/sha/
2. A System of Health Accounts. Paris, Organisation
jointquestionnaire, accessed 14 January 2009).
for Economic Co-operation and Development, 2000
(http://www.oecd.org/dataoecd/41/4/1841456.pdf, 15. Poullier JP. National health accounts manual.
accessed 13 January 2009). Ramallah, Palestine Ministry of Health, 2007.
3. World Health Organization, World Bank, United 16. De la Fuente A, Lequellier F. Measuring employment
States Agency for International Development. Guide in national accounts. Presented at the joint UNECE/
to producing national health accounts: with special Eurostat/OECD meeting on national accounts and
applications for low-income and middle-income update of SNA, Geneva, 25–28 April 2006 (http://
countries. Geneva, World Health Organization, www.unece.org/stats/documents/2006.04.sna.htm,
2003 (http://www.who.int/nha/docs/English_PG.pdf, accessed 14 January 2009).
accessed 13 January 2009).
17. Human resources of European health systems.
4. National health accounts. World Health Organization Augsburg, BASYS, 2001 (http://circa.europa.eu/
Statistical Information System (http://www.who.int/ Public/irc/dsis/health/library?l=/reports/caresgroup/
whosis/indicators/2007NationalHealthAccounts1/en/, hla&vm=detailed&sb=Title, accessed 14 January
accessed 13 January 2009). 2009).
5. National health accounts (NHA). World Health 18. Van Polanen Petel V. Health labour accounts and the
Organization (http://www.who.int/nha/en/, accessed system of national accounts. Presented at the fifth
13 January 2009). meeting of the Eurostat Technical Group on Health
Care Statistics, Luxembourg, 19–20 June 2006.
6. Statistics on occupational wages and hours of work
and on food prices. Geneva, International Labour 19. Statistics Canada. Wages and salaries by branch.
Organization, 2000. Presented at the joint UNECE/Eurostat/OECD
meeting on national accounts and update of SNA,
7. International Standard Classification of Occupations.
Geneva, 25–28 April 2006 (http://www.unece.org/
International Labour Organization (http://www.ilo.org/
stats/documents/2006.04.sna.htm, accessed 14
public/english/bureau/stat/isco/index.htm, accessed
January 2009).
11 January 2009).
20. Maynard JP, Girard A, Tanguay M. Producing
8. International Standard Classification of Education:
hours worked for the SNA in order to measure
ISCED 1997. Paris, United Nations Educational,
productivity: the Canadian experience. Presented
Scientific and Cultural Organization, 1997 [http://
at the joint UNECE/Eurostat/OECD meeting on
www.uis.unesco.org/TEMPLATE/pdf/isced/
national accounts and update of SNA, Geneva,
ISCED_A.pdf, accessed 10 January 2009).
25–28 April 2006 (http://www.unece.org/stats/
9. International Standard Industrial Classification of All documents/2006.04.sna.htm, accessed 14 January
Economic Activities, fourth revision. Statistical Papers 2009).
Series M, No. 4/Rev.4. New York, United Nations
21. Surinov A, Masakova I. Using employment statistics
in the national accounts of the Russian Federation.
demographic/sources/census/2010_PHC/docs/
Presented at the joint UNECE/Eurostat/OECD
ISIC_rev4.pdf, accessed 11 January 2009).
meeting on national accounts and update of SNA,
10. Central Product Classification: CPC version 1.1. New Geneva, 25–28 April 2006 (http://www.unece.org/
York, United Nations Statistics Division, 2002 (http:// stats/documents/2006.04.sna.htm, accessed 14
unstats.un.org/unsd/class/family/family2.asp?Cl=16, January 2009).
22. National accounts: a practical introduction. Studies in
Methods Series F, No. 85. New York, United Nations
publication/SeriesF/seriesF_85.pdf, accessed 14
January 2009).
77
23. Main economic indicators: comparative 35. Cheung P. The System of National Accounts:
methodological analysis – wage related statistics, implementation status and implications for the
volume 2002, supplement 3. Paris, Organization for ICP. International Comparison Program Newsletter,
Economic Co-operation and Development, 2003 2007, 4(1):1–9 (http://siteresources.worldbank.org/
(http://www.oecd.org/dataoecd/20/18/16455130.pdf, ICPINT/Resources/ICPNewsletter_Feb07_Web.pdf,
accessed 14 January 2009). accessed 14 January 2009).
24. Non observed economy in national accounts: survey 36. National accounts: supply and use table. Pretoria,
of national practices. Geneva, United Nations Statistics South Africa, 2002.
Economic Commission for Europe, Statistical
37. Government finance statistics manual 2001 (GFSM
Division, 2008 (http://www.unece.org/stats/
2001). Washington, DC, International Monetary
publications/NOE2008.pdf, accessed 14 January
Fund, 2001 (http://www.imf.org/external/pubs/ft/gfs/
2009).
manual/pdf/all.pdf, accessed 14 January 2009).
25. Hernandez P, Poullier JP. Health systems, health
38. Government finance statistics (GFS): companion
financing and their measurement. Presented at
materials and research. Washington, DC,
the joint CPC/PAHO/WHO/FUNSALUD workshop
International Monetary Fund (http://www.imf.org/
on development of national health accounts in the
external/pubs/ft/gfs/manual/comp.htm, accessed 14
Caribbean, Bridgetown, 10–14 September 2001.
January 2009).
26. SHA guidelines project. Luxembourg, Eurostat,
39. Akritridis L. Accuracy assessment of national
2003 (http://circa.europa.eu/Public/irc/dsis/caretf/
accounts statistics. Economic Trends, 2002,
library?l=/feedback/projectsreports/, accessed 14
589:38–53.
January 2009).
40. National health accounts for specific diseases,
27. Satellite health account (SHA) manual, version 1.
programs, population groups, and regions. Geneva,
Washington, DC, Pan American Health Organization,
World Health Organization (http://www.who.int/nha/
2005 (http://www.paho.org/english/dpm/shd/hp/
developments/en/, accessed 14 January 2009).
satellite-health-account-manual-fin05.pdf, accessed
14 January 2009). 41. Guide to producing regional health accounts within
the national health accounts framework. Geneva,
28. Cuentas nacionales de salud. Mexico,
World Health Organization, 2008 (http://www.who.int/
Fundación Mexicana para la Salud, 1998 (http://
nha/developments/Guide%20to%20Producing%20
healthsystems2020.org/content/resource/
Regional%20Health%20Accounts.pdf, accessed 20
detail/1323/, accessed 14 January 2009).
February 2009).
29. Cuentas nacionales de salud 1995–2000. Lima,
Ministerio de Salud del Perú, 2003.
30. System of National Accounts 2008. Pre-edited
version of volume 1. Commission of the European
Communities, International Monetary Fund,
Organisation for Economic Co-operation and
Development, United Nations, World Bank, 2008
(http://unstats.un.org/unsd/sna1993/draftingphase/
WC-SNAvolume1.pdf, accessed 16 February 2009).
31. National accounts. Paris, Organisation for Economic
Co-operation and Development (http://www.oecd.
org/std/national-accounts, accessed 14 January
2009).
32. ESA95 core national accounts. Luxembourg,
Eurostat (http://epp.eurostat.ec.europa.eu/portal/
page?_pageid=2854,63497418,2854_63867997&_
dad=portal&_schema=PORTAL, accessed 14
January 2009).
33. National accounts statistics: main aggregates and
detailed tables, 2006. United Nations Publication
ST/ESA/STAT/SER.X/39, parts I, II and III. New York,
United Nations Statistics Division, 2008.
34. Joint UNSD-Eurostat Conference on
International Outreach and Coordination in
National Accounts for Sustainable Growth and
Development. Luxembourg, Eurostat, 6–8 May
2008 (http://epp.eurostat.ec.europa.eu/portal/
page?_pageid=2313,68881303&_dad=portal&_
schema=PORTAL, accessed 14 January 2009).
78
Use of facility-based assessments
7 in health workforce analysis

BOLAJI FAPOHUNDA, NANCY FRONCZAK, SHANTHI NORIEGA MINICHIELLO,
BATES BUCKNER, CATHERINE SCHENCK-YGLESIAS, PRIYA PATIL
7.1 Introduction and classifications of certain health worker catego-

ries; weak technical capacity to conduct in-country
The global health literature demonstrates that health- workforce analysis; lack of appropriate measurement
care service coverage and quality are directly tools; and underinvestment in national health informa-
correlated with health worker numbers and perform- tion systems (1–3). In particular, lack of standardized
ance. For instance, the World Health Organization approaches to HRH assessment limits the potential for
(WHO) has presented evidence showing that cover- comparative analysis over time and across countries
age of selected primary health-care services, including to better understand how different situations, policies
maternal, newborn and child health interventions, tend and interventions impact the performance of human
to rise with higher national health workforce densi- resources and health systems and, ultimately, popula-
ties (1). Using the Millennium Development Goals as tion health outcomes.
the benchmark, WHO reports that countries with the
highest shortfalls in numbers of physicians, nurses and Health facility assessments (HFAs), the focus of this
midwives are the ones most at risk of not meeting cov- chapter, are tools for gathering data that are a potentially
erage targets. The African region – home to only 3% of important source of information for health workforce
the estimated 59.2 million health workers in the world monitoring. A number of countries already conduct
but having 24% of the global burden of disease – is the such assessments, and demand for them is increas-
area hardest hit by health worker shortfalls and imbal- ing. Health facilities refer to service delivery points in
ances worldwide (1). the formal health sector, including hospitals, health
centres, dispensaries and health posts. HFA protocols
Securing improvements in the size and quality of the capture real-time information (i.e. at the moment of the
health workforce is important for achieving regional assessment) on a key component of the overall health
and country-specific Millennium Development Goals in system: facility-based service delivery. Depending on
health. Overcoming human resources for health (HRH) the nature of the data collection instruments, HFAs
shortages and imbalances requires strengthening edu- can provide detailed information on health workforce
cation and training programmes for health workers, availability, distribution, qualifications, skills mix, train-
improving health sector working conditions (including ing and performance. This information can be used
staff salaries and benefits) and forging cooperation to determine, for example, how existing staffing pat-
and collaboration in health workforce management terns relate to desired or planned staffing levels, how
within and across countries. Evidence-based moni- well staff members’ qualifications match their assigned
toring of health workforce dynamics is important for scope of work and the nature and extent of geographi-
ensuring that policy and programmatic inputs lead to cal or other staffing imbalances. HFAs can also provide
the expected outcomes. insight into the broader health labour market context,
including management practices and other features of
Often, a lack of comprehensive, timely and reliable data
the work environment (for example infrastructure and
on HRH results in poor knowledge of workforce status
availability of medical supplies and equipment), and
and curtails development of evidence-based policies
how these variables affect health worker supply and
among national and international stakeholders. Several
performance. In short, HFAs can inform workforce pol-
factors have contributed to the weak information and
icy by telling us what is happening on the ground, in the
evidence base on the health workforce in many low- and
real world of service delivery.
middle-income countries. These include lack of a com-
mon framework from which to understand HRH issues; The main objective of this chapter is to describe the
poor data availability and quality; imprecise definitions current and potential usefulness of HFAs as a source of
79
information for health workforce planning, management,

monitoring and policy-making. Illustrative examples are Box 7.1 Core health workforce
presented based on empirical data from HFAs con- indicators potentially measurable
ducted in Kenya, Nigeria and Zambia. Facility-based with HFA data
assessments cover a broad array of data collection
techniques, including facility audits, observations of
services delivered, interviews with service providers Entry stage of the working lifespan
and interviews with clients. These various methods,
t Number of new medical/health graduates
along with other practical considerations in planning
entering the facility-based health workforce
an HFA, are reviewed here. However, this chapter does
t Ratio of new medical/health graduates entering
not elaborate step-by-step instructions on designing
the facility-based workforce to the total facility-
and implementing facility assessments; for general
based health workforce
information and relevant resources see International
Health Facility Assessment Network tools (4, 5).
Active workforce stage
Supply
7.2 How facility-based t Stock or total number of facility-based health
assessments can be used for workers
health workforce monitoring t Number of facility-based workers relative to total
(catchment) population
As detailed in Chapter 1 of this Handbook, there are t Number of facility-based workers relative to
three interdependent stages in the lifespan of the health planned staffing norm
workforce: (i) pre-service or entry into the workforce; t Number of staff per health facility (by type of
(ii) active workforce; and (iii) exit from the workforce. facility or services offered)
Ongoing measurement and monitoring of perform-
ance indicators for each of the three stages is needed
Distribution
to determine the health system’s readiness and abil-
ity to maintain a sufficient stock of qualified workers. t Skills mix of facility-based staff
Because the stages are interactive and interdepend- t Geographical distribution of facility-based staff
ent, monitoring must be holistic rather than fragmentary, t Age and sex distribution of facility-based staff
focusing on the whole rather than little slices of the pie.
While detailed analysis of the entry and exit stages is Capacity, motivation and performance
beyond the scope of HFAs (for example health edu-
t Level and field of education among facility-
cation outputs and costs, and mortality and migration
based staff
among health workers, respectively), facility-based
t Years of professional experience among staff
sources can provide valuable information to comple-
working at the facility
ment data obtained using other methodologies (such
t Staff receiving in-service training during a
as special studies on education or migration).
reference period (by type of training)
t Services provided by staff during a reference
Box 7.1 shows a list of indicators that can potentially
period
be measured for each of the workforce stages using
t Proportion of staff working full time versus part
HFA data. Most population-based sources of health
time at the facility
workforce statistics (for example population censuses,
t Proportion of staff assigned to the facility (in
labour force surveys) tend to relate workforce data to
post) working at the facility on the day of the
the general population; while important in their own
assessment
right, they provide little insight into the service delivery
t Proportion of staff receiving (non-monetary)
environment, service quality or other operational fac-
incentives at their job
tors within the health system that can play a major role
in workforce performance. HFA data can help address
this information gap by describing health labour dynam- Exit stage
ics at facility-based service delivery points. t Facility-based health workforce attrition rate (by
reason for leaving workplace)
t Ratio of facility-based health workers entering
the workforce to those exiting the workforce
80
Use of facility-based assessments in health workforce analysis
7.3 Overview of key and other constraints. Usually, once the list frame has
been developed, a multistage sampling plan is followed
HFA methodologies
to ensure representation across various domains of
This section provides an overview of issues of impor- the universe of eligible facilities. The stages are deter-
tance when planning an HFA, and describes the mined by the different eligibility criteria (for example
methods and focus of data collection for several dif- administrative authority, type, geographical location).
ferent HFA tools developed by different private, public When a multistage sampling is used, sample weights
and international organizations. need to be applied when analysing the data to calibrate
for national representation. The weights are mathemat-
7.3.1 Issues when planning an HFA ically derived by sampling experts.
All HFAs collect data at the facility level, but methodol- A key weakness in generating a list frame of health
ogies and protocols may vary in relation to information facilities is that different authorities do not always
needs, costs and sources of funding, and local imple- have up-to-date records of facilities functioning in the
mentation capacity. Here the main practical and country. Experience shows that oftentimes facilities,
methodological issues are discussed, as they relate to particularly in the private sector (either for-profit or not-
HRH monitoring. for-profit), may have closed or changed addresses,
and there is no standard definition for a type of facil-
Selection of facilities to be covered in ity in the private sector. An initial list obtained from the
the HFA: census or survey sample ministry of health will usually need to be complemented
The design of an HFA requires careful attention to the with information from multiple other sources, such as
strategy for data collection from the initial planning private sector coordinating bodies, social ministries
stages, particularly the sampling method: census or where nongovernmental organizations register their
survey sample. The choice of method often depends activities, or directly from faith-based, private and par-
on trade-offs in scope and depth of information to be astatal organizations.
obtained with cost and time factors.
Where the HFA includes interviews with individual
Census method. A census is a full enumeration, or providers, the health worker to be interviewed is ran-
the collection of data from all facilities that meet eligi- domly selected from the list of those present on the
bility criteria. Examples of eligibility criteria that have day of survey. Although ensuring an unbiased (non-
been used for HFAs include: (i) managing authority, zero) chance of selection for each health worker is an
whether government, private for-profit, nongovernmen- important factor, practical considerations of availabil-
tal or faith-based organization, or other management or ity and relevance are also essential. In most situations,
funding criteria; (ii) facilities offering certain services (for health workers providing direct client services tend
example maternal and newborn services, child health to be prioritized over those performing administrative
services, HIV-related services); (iii) facilities of a given duties, such as maintaining health information records
type (from primary health-care centres to tertiary-level or other activities not directly entailing services deliv-
hospitals); or (iv) facilities within a certain geographical ery. A major advantage of a well-designed sampling
area. Often, a combination of several such criteria is plan is that in-depth data collection is feasible within
used. Advantages of a full enumeration, or facility cen- a reasonable time and cost. More information on sam-
sus, include having information specific to every facility pling methodologies to provide unbiased estimates of
and the potential for simpler data analysis and inter- facilities and their characteristics is available elsewhere
pretation of results (no need for sampling weights or (see, for example, the MEASURE Evaluation manual
calculating a statistical margin of error). Disadvantages (6)).
include difficulties in ensuring a complete enumera-
tion of all eligible facilities and higher costs, especially Data collection methods and tools
when the number of facilities to be enumerated is large. Facility-based assessments may employ one or more
techniques for data collection, including facility audit
Survey sample method. In survey approaches, prob-
(often referred to as an inventory), observation of serv-
ability sampling principles are used to draw a selection
ices delivered, interviews with clients, and interviews
of facilities for inclusion in the assessment. First, the
with service providers and other facility staff. Tested
eligibility criteria (see above) are developed; then,
data collection tools exist for each of these methodolo-
a number of facilities are selected based on a sam-
gies (see section 7.3.2 below).
pling frame or list of all eligible facilities. The larger the
sample size, the greater the precision of the results; Facility audit. This is the method used to collect infor-
however, the total size will often also depend on budget mation on infrastructure, availability of equipment and
81
supplies, staffing levels, services offered, and manage- highly skilled and specialized than the observed worker
ment and support systems in place. While structured revisits with the client afterwards to determine whether
questionnaires are always used in facility audits, there correct diagnosis, care and treatment were provided.
may be differences across tools in how the informa-
tion is collected. Protocols tend to differ with regard to Interviews with clients. Client or exit interviews are
choice of respondent to the questionnaire (for exam- often used to ascertain client perspective on the qual-
ple interviewing the person in charge of the facility’s ity and responsiveness of services received, which can
operations overall, versus selecting a range of persons be useful for assessing health worker performance. Exit
considered most knowledgeable for each information interviews may be conducted among a random sam-
domain); where within the facility physical data collec- ple of all attenders on the specific day, or include only
tion will occur (enumeration of all items regardless of those whose consultations were observed. A major
location in a facility, versus counting only those located weakness in exit interviews is the bias toward more
and functioning in the relevant service delivery area on recent and possibly more self-motivated clients whose
the day of the assessment); and whether there is vali- care-seeking behaviours may not be representative of
dation of reported responses (accepting as legitimate the target population, or the universe of those receiv-
any interview response, versus requiring additional ing such services. Additionally, exit interviews, by their
data collection by other approaches such as inspec- nature, provide a superficial rather than an in-depth
tion of equipment or review of administrative records). examination of the client’s perspectives on the serv-
Using multiple respondents, validating reported infor- ices received. An in-depth analysis is preferable, but
mation by observation and ensuring items are in the requires time inputs that may be unduly burdensome to
relevant service area each take time, particularly in clients who may have already spent many hours in the
large and complex facilities. In addition, validation exer- facility waiting for and obtaining services.
cises may require engaging more highly skilled data
collectors who are familiar with health services and Interviews with service providers. Interviews with
systems (see next subsection on selection of data col- service providers are used to collect information on
lectors). However, these techniques tend to provide the types of services provided, opinions related to the
most uniformly reliable and valid information, and allow working conditions, educational attainment, in-serv-
for more in-depth assessments of capacity to provide ice trainings and working experience. The method
quality services. For example, if blood pressure moni- may also be used to evaluate health workers’ knowl-
tors are found to be available somewhere in the facility, edge in specific topic areas. As with the exit interview,
but none are within the specific area where a health health worker interviews are, by design, short in order
worker who might require such equipment is providing to minimize disruption to service provision, since data
services, it is unlikely that clients will have their blood collection occurs during regular business hours. As
pressure measured. Past experience also indicates such, HFAs are not a source of detailed information
that when interview responses from key informants are for human resource development; additional informa-
not validated, respondents sometimes provide answers tion on workforce performance will have to be obtained
that describe the usual or even desired situation, rather from special studies, which are imperative in any HRH
than the actual situation on the day data are collected. monitoring and evaluation plan. Facility records of staff
This is especially true in larger facilities, where an over- training may provide an alternative source for some of
all administrator may not be closely familiar with the these data, provided such records are available, com-
day-to-day state of affairs in each service delivery area. plete and routinely updated.
Observation of key services. Client observations Selection of data collectors

are service specific, and may be based on a full cen-
A combination of persons with clinical and social sci-
sus (observing all eligible clients who receive services
ence backgrounds is recommended for data collection
during the period of data collection) or opportunistic
activities in health facilities. A mix of persons with skills
sample (clients served when the data collector was
in the science of data collection and those familiar with
available and present). Most of the data collection
the functioning of health facilities helps to assure the
tools are checklists that measure process, for exam-
quality of data collected. For instance, it is generally
ple information shared, examinations carried out and
agreed that interviews with providers and clients can
medicines or tests prescribed. In general, observation
be handled well by non-health personnel, but obser-
data are used to assess compliance of health worker
vations of provider–client interaction require observers
practice to established guidelines. Direct observation
with advanced training in a health-related field. Having
may be followed by special studies in order to improve
data collectors with a health background may be less
data credibility. These are sometimes referred to as
important for facility audits, depending on the complex-
“gold standard” observations, whereby someone more
ity of the tool being used, but knowledge of the health
82
system processes is likely to enhance the efficiency country (a subset of districts may be sampled in very
and reliability of data collection in any context. large countries). The health facility module captures
information on all public and private facilities within
districts. The capturing of geographical coordinates
7.3.2 Examples of tested
allows unique identification and charting of health
HFA instruments
facilities. Data collected on HRH include staff qualifica-
A number of different HFA tools have been developed tions, staff availability on the day of visit versus staffing
and implemented under the auspices of international norms – that is, approved staffing positions for the spe-
technical cooperation programmes for collecting data cific facility as assessed based on service utilization
relevant for HRH analysis. They include the Health or other workload indicator (see 12 for related tool) –
Facility Census (HFC) developed by the Japanese and working hours. Although essentially a district-level
International Cooperation Agency (JICA); Service monitoring tool, SAM is expandable to include more
Availability Mapping (SAM) developed by WHO; the HRH questions in order to provide detailed data on
Service Provision Assessment (SPA) developed by health personnel at the district level. The method can
Macro International; and the Situation Assessment of be combined with the HRH targeted approach, such
Human Resources in the Public Sector developed by as that developed in the PHRplus tool (described later
Partners for Health Reformplus (PHRplus) (7). Overall in this section), to produce data for more robust HRH
such methods can be tremendously useful for updat- analyses that take into account national sociodemo-
ing and validating national databases of health system graphic and epidemiological contexts.
statistics, including HRH, and should be considered
for this purpose by any country interested in investing
in its HRH information system. Customized modules Service Provision Assessment
can also be added to supplement these standardized SPA is designed to assess the quality of health serv-
approaches if they are needed for comprehensive HRH ices as measured through resources, systems and
assessment in a given context. The types of health some observed practices (13). Developed with fund-
workforce information collected are generally similar ing from the United States Agency for International
across sources, but results may not be directly compa- Development (USAID), SPA collects data on current
rable, given each tool’s specificity. workforce size versus staffing norms. Additional data
may be collected on each staff member’s qualifica-
tions and working hours and on seconded workers
JICA Health Facility Census
(for example numbers and salary source). An example
The JICA Health Facility Census tool is designed to of where this kind of assessment may be useful is the
provide detailed information on the status of physical Caribbean region, where physicians frequently rotate
assets at all health facilities within a country (8). The on a schedule between several facilities. Interviews
focus is on information useful for public health system based on a subsample of staff, including primary
infrastructure investment planning (9, 10). Designed providers of key services, are used to assess the work-
as a physical asset mapping, the HFC also conducts force skills mix in terms of workers’ levels and fields
a headcount of all health workers present on the day of pre-service education, types of in-service training
of visit. Data obtained include staff qualifications, the received during a reference period and years of expe-
number of staff present versus the number assigned, rience at the given facility. (The tool does not presently
demographic composition and staff qualifications. collect data on specialty education received after the
initial clinical qualification, although some information
Service Availability Mapping may be inferred from respondents’ reported duration
of schooling for their current occupation.) The propor-
Developed by WHO, SAM is designed to determine
tion of staff performing activities in the specific area
the availability of key programmes and resources, and
of in-service training can be measured by this instru-
map common services (11). While originally developed
ment, and can be utilized to assess whether training is
for implementation at the district level, the facility-
targeted to the appropriate staff and whether staffing
based data can be aggregated to provide evidence
assignments take training into account.
for decision-making at the national level when all dis-
tricts in a country are included. SAM consists of a suite
of tools, including district and health facility question- Partners for Health Reformplus
naires, each of which is administered to key informants; situation assessment
a data entry programme on personal digital assistant The PHRplus survey tool presents an example of how
(PDA); and a geographical information system software a facility-based assessment of HRH may be con-
(HealthMapper). The district questionnaire is meant strued. Developed by Partners for Health Reformplus
to be administered to all district medical officers in a with funding from USAID, this tool collects data on
83
workforce size, skills mix, distribution and turnover t The Quantitative Service Delivery Survey is an
rates in the public health sector (7). Data on individ- outgrow of the Public Expenditure Tracking Survey
ual workers are supplemented with a situation analysis of the World Bank, whose original purpose was to
of national health planning, information and strategy analyse the efficiency of financial resource flows and
documents, plus a modelling exercise to determine estimate the leakage of public resources from cen-
human resources needed for achieving targets under tral government to the front-line service providers
the United States President’s Emergency Plan for AIDS (17). The tool includes questionnaires addressing all
Relief and the health-related Millennium Development levels of service delivery: ministry of finance, min-
Goals. Estimates based on the modelling exercise istry of health, regional and district administrations,
developed for Nigeria – focusing on provision of serv- health facilities, health service providers and clients
ices for HIV, malaria, tuberculosis, maternal and of health services. The facility instrument, intended
child health, and family planning – illustrate how this for the head of facility, gathers financial data (both
approach could inform the development of a stand- revenue and expenditure sides) and information on
ardized version for cross-country analysis (14). Such institutional arrangements and governance, to name
adaptation should be made in the context of indicators a few. The providers survey can be used to study
that health managers need for programme monitor- worker morale, absenteeism and coping strate-
ing and should factor in country- and disease-specific gies such as informal payments. Some surveys also
contexts. include vignettes to assess the level of knowledge of
doctors and nurses, which serves as a proxy for the
quality of care provided.
Selected additional tools
Other HFA methodologies useful for assessing the
HRH situation have been developed, tested and imple- 7.3.3 Implications of HFA methodologies
mented at the national, subnational and programme and data collection issues
levels. A list of these tools, though not exhaustive, fol- When it comes to the sampling approach, in general,
lows below. censuses are more appropriate when facility-specific
t The Facility Audit of Service Quality is a rapid and information is needed, such as infrastructure, human
simple assessment tool developed by MEASURE resources, equipment, supplies and other essential
Evaluation to help district- and programme-level inputs. Survey samples are best when more in-depth
officials design and implement a tailored HFA (15). information is needed, including particular details
The audit employs a strategy that recommends a reflecting systems processes, services provision,
complete enumeration of all facilities in the target health information and record-keeping practices, pro-
districts. When implemented in the intended fashion vider productivity, management and supervision, and
– that is, with local district or programme staff serv- client perception of service quality. A mixture of the HFA
ing as data collectors and asking local stakeholders methods may be most appropriate when assessing
to tailor the protocol to their needs – results are not HRH at the facility level. Censuses (such as the HFC or
likely to provide the consistency required for aggre- SAM) can provide precise pictures of the numbers and
gation at the national level. As such, the approach is distributions of health workers, whereas sample sur-
not designed or recommended for use for national- veys (such as SPA) including provider interviews offer
level planning and evaluation. the means to identify systems issues that affect worker
motivation and satisfaction and the information base to
t The Assessment of Human Resources for Health design retention strategies. It is sometimes possible to
is a survey instrument developed by WHO for col- mix sampling methods within the same assessment,
lecting quantitative and qualitative HRH information collecting some (basic) information from all facilities
(16). Four questionnaires are included in the tool, and more in-depth data from a sample.
focusing on the following areas: health professional
regulation, training institutions, health facilities and Several variations of HFA data collection tools have
health-care providers. The questionnaire for health- been developed and used to meet a wide range of
care providers – which covers topics such as specific information needs. It is essential that any
professional qualifications, dual employment and generic tools are country-adapted before they are
occupational mobility – is meant to be implemented utilized so that precoded responses capture local ter-
among a representative sample of facility-based minologies and processes. This applies not only to
health workers, and can be merged with the data certain equipment and supplies (for example names of
collected at the facility level for collating information medicines) but notably also to occupational titles and
on the environment within which workers operate. qualifications of staff, which would then be harmonized
and mapped against standard definitions during data
84
processing and analyses for enhancing comparability relatively short timeframe. In countries where a large
across sources and over time. portion of health services are provided by the private
sector, government-only HFAs will underestimate the
Overall, there is no single HFA method that will meet all overall human resource stock and flow.
needs. Understanding the benefits and problems with
different methods and tools will help stakeholders to On the other hand, including the private sector could
select those most appropriate for their needs. increase the risk of double-counting health workers,
particularly in settings where dual employment across
both sectors is common but not sanctioned by regula-
7.4 Some limitations of HFA tory and legal frameworks. Dual practice (that is, where
a worker holds two or more jobs at different locations)
methodologies for HRH data is present in virtually all countries regardless of income
While HFAs present a number of advantages for HRH level and does not necessarily impact negatively on
analysis, there are also some limitations to this method- health system performance. Even in contexts with strict
ology that need to be considered. regulatory restrictions of health professional activities,
such as China and much of Latin America, physicians
often hold jobs in both the publicly funded and private
7.4.1 Assessment coverage
systems (18). The implication for HRH monitoring is that
As with all population- or establishment-based data careful consideration of ways to avoid double-count-
collection exercises, availability of an adequate sam- ing personnel should be an important component of
pling frame for selection (either universal selection for any health facility-based assessments. Methods that
census or sample for survey) is a key factor. Ideally, for have been used to address this issue include collect-
HFA approaches, a complete list of all service delivery ing information from health workers on the proportion of
points in a country (or targeted areas) can be obtained working hours during a week spent in a particular facil-
from registration or licensing offices, or from the coun- ity, or specifically asking about other facilities where
try’s routine health information system. Often however, they work.
these sources are non-existent, incomplete or out of date
when it comes to health facility information. Likewise, if
a representative sample of staff within selected facil- 7.4.2 Lack of standardization
ities needs to be drawn for interviews, this requires a of definitions and statistical
further step in sampling design, notably a complete classification of health workers
listing of all facility-based personnel (and even time- The lack of standardization of health worker definitions
specific duty rosters). Inadequately designed or poorly across the various HFA approaches, including consist-
implemented sampling at any level compromises the ent definitions for occupational categories, is a serious
validity of generalizations to the health workforce as hindrance to the usefulness of HFA data for compara-
a whole and severely compromises the utility of such tive analysis of HRH within and across countries and
data as evidence for decision-making. over time. Most HFA approaches use country-specific
or tool-specific occupational labels with no provi-
It is not uncommon for HFAs to cover only facilities in the sion for translation to an internationally comparable
government or public sector, with obvious implications set of categories. Many national occupational titles –
for applicability of ensuing analysis to reflect the true especially those referring to staff other than medical,
national situation. In some cases, this limitation reflects nursing and midwifery professionals – and the under-
weak enforcement of regulatory mechanisms oversee- lying data on staff qualifications are often not captured
ing service provision outside the government sector in a way that can easily be mapped to the International
(and thus poor data on the operations and locations Standard Classification of Occupations (ISCO) (19) or
of such providers from which to complete the sam- the International Standard Classification of Education
pling frame), and in others the difficulty of collecting (ISCED) (20), respectively.
data from privately operated facilities (where refusals
to participate are more common). When private sector Occupational categories that present special chal-
facilities are included, coverage tends to favour not- lenges include those for health workers providing
for-profit facilities (for example parastatal facilities or community or counselling services and those in the
those sponsored by recognized nongovernmental and assistant or aide category. The job criteria and respon-
faith-based organizations). In many low- and middle- sibilities assigned to these categories vary widely
income countries, private for-profit facilities tend to be from country to country, and even across the pub-
smaller, geographically concentrated in urban areas, lic and private sectors within a given country. These
less standardized in terms of staffing and breadth of categories of workers are numerous and, especially in
services, and more likely to cease operation within a
85
countries where highly educated medical practitioners the PHRplus tool. The SPA sample survey data are
are scarce, they often serve as the primary, on-the- weighted to correct for unequal selection rates across
ground providers of health services. Nationally defined sampling units. Data from the HFC and SAM methods,
occupations with labels such as “medical assistant” or which used census sampling (or complete enumeration
“clinical officer” can be mapped to ISCO only if infor- of all facilities), did not need to be weighted or adjusted.
mation on the level and specialization of educational
qualifications are known. In some countries, workers
7.5.1 Entry to the health workforce
with these labels work at an associate professional or
lower level, while in other countries workers with the HFAs can be used to estimate the number of new
same labels have education requirements and respon- entries or incoming staff to the facility-based workforce,
sibilities at a level higher than a professional nurse. and the ratio of new entries to the total facility-based
Reliable mapping of national to international occupa- stock of health workers in a particular reference period.
tional and educational classifications depends on the Table 7.1 presents an example of such data for selected
detail with which data on the national categories are categories of health workers in Nigeria’s public sector.
obtained, including, for example, data on staff training The data show the numbers who entered the public
and responsibilities. Existing HFA tools tend to vary on health service in 2005: 1214 physicians, 1331 nursing
this level of detail. and midwifery personnel, 501 laboratory technolo-
gists and technicians, 443 pharmaceutical staff (about
half of whom were pharmacists) and 2742 community
7.4.3 Lack of uniformity in health workers (including community health officers
definitions of facility types and community health education workers). New entries
To date there has been no uniformity across coun- constituted 7.7% of the existing stock of physicians,
tries and tools in definition of common categories of 1.1% of nurses and midwives, and 3.6% of pharmaceu-
health facilities. Even within a given country, facilities tical staff (14).
assigned the same label can vary greatly in size and
function. In Kenya, for example, HFA data have shown When compared to the number of graduates from
that almost one in three facilities officially classified as health professions education institutions in the same
health centres had no overnight or inpatient beds, but year, it can be observed that entries into the pub-
one in six had 20 or more beds. Standardized defi- lic sector accounted for 60.7% of all newly graduated
nitions for the most common types of health facilities physicians, but only 2.4% of newly graduated nurses
are needed to enable comparisons within and across and midwives. Such information, even if incomplete, is
data sources. The International Standard Industrial important for understanding health worker preferences
Classification of All Economic Activities offers some in choice of employment – in this case, public sector
guidance on this matter (21), but the definitions and service – and how this varies by cadre.
categories in this generalized resource are broader
than required in a typical HFA; even the lowest level In the absence of information from health education
of disaggregation classes just “hospital activities” and institutions on numbers of new graduates, an exami-
“medical and dental practice activities”. Recent efforts nation of the age profile of health workers can provide
towards establishing a common, detailed standard for some insight into workforce renewal patterns. Data
HFAs have experimented with using the number of from the Kenya 2004 SPA (22) indicate that the health
inpatient beds as a proxy for facility size and service workforce in the country is quite young, with over one
complexity. These issues are important for consider- quarter (26%) of all facility-based staff aged less than
ing, in particular, staffing norms and other indicators 30 years. As shown in Figure 7.1, the ratio of younger
pertinent to HRH analysis. workers to those closer to retirement age (45 years and
over) is relatively high, especially among medical per-
sonnel. On the other hand, the ratio of younger to older
personnel in the nursing and midwifery field is much
7.5 Empirical examples
lower. Public facilities tend to have younger health
based on HFA data workers than private facilities (results not shown). This
This section provides illustrative examples of the types type of information might indicate a need to further
of HRH data produced when using existing HFA tools, investigate the reasons younger health workers may
and how such facility-based data can be used in moni- be less likely to serve in certain facilities and develop
toring and analysis of the HRH situation. The illustrative appropriate incentives for their recruitment, or to inves-
analyses draw on microdata collected in Kenya and tigate the reasons older workers may be more likely to
Zambia using three different HFA techniques (HFC, leave this kind of service and improve incentives for
SAM and SPA) and published results for Nigeria using their retention.
86
Table 7.1 Estimated number of new graduates entering the public sector facility-based health
workforce, by cadre, Nigeria 2005
Cadre Number of new New incoming staff Number of New incoming

incoming facility- as % of existing graduates from staff as % of
based staff stock health professions annual education
education institution turnout
institutions
Physicians 1214 7.7 2000 60.7
Nursing & midwifery personnel 1331 1.1 5500 2.4
Laboratory personnel 501 3.4 n.a. n.a.
Pharmaceutical personnel 443 3.6 800 55.4
Community health workers 2742 3.3 n.a. n.a.
n.a. Not available (no information collected).
Source: Chankova et al. (14).
Figure 7.1 Ratio of facility-based health workers aged under 30 to those aged over 45, by cadre,
Kenya 2004 SPA
Ratio of younger to older health workers
2.0 1.9
1.5 1.4
1.0
0.7
0.5
0.0
Medical Nursing & All health
personnel midwifery workers
personnel
Source: Kenya 2004 SPA (22).
87
7.5.2 Active health workforce Figure 7.2 shows a total of 6985 health workers officially
assigned or hired (in post) in the sampled facilities at
Workforce supply the time of the Kenya SPA survey. Among these, med-
Table 7.2 and Figure 7.2 describe the supply of the ically trained personnel (including physicians and
active health workforce in two national contexts: paramedical practitioners) comprise 8% of the total;
Zambia and Kenya, respectively. The stock and density nursing and midwifery personnel comprise almost half
of facility-based health workers in Zambia is shown in (45%); laboratory personnel 6%; pharmaceutical per-
Table 7.2. The overall density, across all cadres, is 10.5 sonnel 2%; counsellors and social workers 5%; staff in
health workers per 10 000 population (23). In particular, other clinical areas such as nutrition and rehabilitation
for physicians the ratio is 0.8 per 10 000 and for nursing 1%; health information and medical records techni-
and midwifery personnel it is 6.9 per 10 000. Although cians 3%; and nearly a third fall into other categories,
these calculations do not account for health workers such as aides, clerks and community-based workers
who are not facility based, they present tracers of the who are also facility staff.
health system capacity and can serve as a proxy for
the overall health workforce density in those countries Tabulating staffing levels by the type of facility can pro-
where most service providers work at least part time vide a useful common reference for comparisons. The
in health facilities (for example through dual practice). number of staff generally increases with the size of
the facility, but the ratio of increased need for human
resources to facility size will vary and depends on many
different factors. An interesting feature observed from
the Kenya SPA data (Table 7.3) is the situation at stand-
Table 7.2 Number and density of facility-based alone sites for voluntary counselling and testing (VCT)
health workers, by cadre, Zambia 2006 HFC
for HIV. In Kenya, these specialized outpatient sites
are staffed by a generous average of 4.5 counsellors
Cadre Number Density of
health workers per site, and at a total staffing level higher than that for
(per 10 000 outpatient clinics and dispensaries with no overnight
population)a beds (7.2 versus 5.4). VCT service sites, representing
a small proportion of all facilities (9 of 440 facilities),
Physicians 908 0.78
usually offer one service, so they may rely more heav-
Nursing & midwifery 8068 6.91 ily on providers with counselling qualifications. It may
personnelb also be that counselling services are being provided
Paramedical practitionersc 1342 1.15 in other facilities by other categories of health workers,
such as nursing staff. Systems that train one service
Pharmacists 115 0.10
provider to offer multiple services are common, where
Physiotherapists 114 0.10
Environmental health 679 0.58
Figure 7.2 Number and percentage distribution
workers of staff currently in post at health facilities, by
Nutritionists 62 0.05 cadre, Kenya 2004 SPA
Hygienists 2 <0.01 Other health

workers: 2103
Laboratory technicians 480 0.41 (30%)
Medical
professionals:
Pharmaceutical technicians 116 0.10 567 (8%)
Health
Radiography technicians 139 0.12 information/
records
Orthopaedist technicians 14 0.01 technicians:
188 (3%)
Other technicians 35 0.03
Other health workers 145 0.12 Other clinical
staff: 66 (1%)
Total 12 219 10.46

Counselling/
a. Based on the estimated 2005 national population social workers:
(11 683 704). 382 (5%)
Nursing and
b. Cadre includes all levels of nursing and midwifery personnel, Pharmacy midwifery
personnel: Laboratory
including enrolled and registered nurses and midwives. personnel: personnel:
112 (2%) 3133 (45%)
c. Cadre includes all similar levels of paramedical practitioners, 433 (6%)
such as medical officers and clinical assistants.
Source: Herbst and Gijsbrechts (23). Source: Kenya 2004 SPA (22).
88
Table 7.3 Mean number of health workers by type of facility, according to cadre, Kenya 2004
Cadre Facility type/size

VCT stand- No 1–9 beds 10–19 20–49 50–99 100+ beds
alonea overnight/ beds beds beds
inpatient
beds
(n=9) (n=237) (n=93) (n=40) (n=39) (n=8) (n=14)
Medical practitioners 0.1 0.3 0.9 1.1 1.5 6.7 16.1
Nursing & midwifery 0.4 2.4 3.9 5.7 7.4 23.7 107.5
personnelb
Laboratory staff 0.1 0.4 0.9 1.2 1.2 3.5 9.6
Pharmaceutical staff <0.1 <0.1 0.3 0.2 0.2 1.6 3.3
Counselling/social work 4.5 0.3 0.6 1.1 0.8 4.4 6.5
Other clinical 0.2 <0.1 0.1 0.2 0.1 1.1 2.2
Health information/ 0.3 0.1 0.3 0.4 0.4 2.1 6.5
records techniciansc
Other health workers 1.6 1.8 3.1 4.2 4.0 23.9 70.1
Total 7.2 5.4 10.1 14.1 15.4 67.4 225.4
a. Service site that primarily offers voluntary counselling and testing for HIV.
b. Cadre includes all levels of nursing and midwifery personnel, including enrolled and registered nurses and midwives.
c. Cadre includes all staff with primary responsibility of implementing and managing administrative records-keeping processing, stor-
age and retrieval systems.
Source: Kenya 2004 SPA (22). Data weighted to reflect survey sampling procedures.
Figure 7.3 Ratio of health workers per inpatient bed, by facility size, Kenya 2004 SPA
1.5
1.5
Median number of health
1.0 0.9 Medical, nursing

workers per bed
0.8 0.8
0.7 & midwifery
0.7 0.7
0.5 0.5 All staff
0.5 0.4
0.3
0.3
0.0
1-9 10-19 20-29 50-59 100+ All facilities
Number of patient beds
scarce human resources do not allow a specialist for in Kenya, disaggregated by facility size, with facility
every service. size used as an indicator for complexity of services
offered and numbers of clients receiving services. This
Another useful indicator for monitoring workforce sup- information can be used to identify potential outliers
ply is the ratio of health workers per patient bed. Figure in levels of staffing and to monitor changes in staffing
7.3 graphs the median ratio of health workers per bed ratios over time.
89
Table 7.4 Number of health workers currently in post, number recommended by staffing norm, and
number requested by facility managers to meet the norm, by cadre, Kenya 2004
Cadre (1) (2) (3) (4) (5)

Number Number of Percentage Number Percentage of
of health health workers of norm requested by requested staff
workers recommended in currently in facility managers currently in post
currently in facility staffing post to meet work
post norm requirementsa =(1)/(4)
=(1)/(2)
Medical practitioners 567 1 114 51 1 852 31
Nursing & midwifery 3 133 4 343 72 6 295 49
personnel
Laboratory staff 433 742 58 1 175 37
Pharmaceutical staff 112 315 36 563 20
Counselling/social work 382 970 39 1 861 21
Other clinical staff 66 385 17 779 8
Health information/records 188 444 42 779 24
technicians
All other staff 2 103 2 183 96 2 740 77
Total 6985 10 495 16 044
a. Includes staff reported by facilities managers as being needed to meet their staffing norm, plus staff needed to meet work require-
ments as reported by managers of facilities without a staffing norm.
Source: Table produced by Macro International, based on Kenya 2004 SPA microdata.
Information on workforce supply is especially useful Among the total number of health workers that facil-
when placed in the context of staffing need. To sup- ity managers reported are needed, the percentage
plement information on existing levels of staffing, it is that are currently assigned to the facility is presented
important to have some idea of the extent of actual in Figure 7.4 for medical practitioners, nursing and mid-
or potential staff shortages. One way in which health wifery personnel, and all categories of health workers,
ministries attempt to address this issue is to develop by facility type. Only in VCT sites and the larger facil-
staffing norms. These norms are meant to take into ities (most of which are located in Nairobi and other
account the services offered and the numbers and major urban areas) are at least three quarters of the
types of clients to be served. The difference between overall staffing requirements met. Virtually none of the
the staff in post and the staffing norm is the vacancy outpatient clinics and dispensaries has the minimum
rate (Table 7.4, column 3), which provides information number of medical practitioners. Smaller facilities tend
on how well the managers are able to fulfil their planned to have only half of the nursing and midwifery personnel
staffing patterns. Additional information from the Kenya needed to meet staffing norms and work requirements.
SPA presented in Table 7.4 shows a more complex pic- This type of information gives us a quick assess-
ture that compares the number of staff recommended ment of the staffing situation in the surveyed facilities.
under current norms to the number reported by facil- However, interpreting the consequences associated
ity managers as required to meet the service utilization with unfilled positions requires additional knowledge of
at their facility. It appears that facility managers in how the assigned staffing relates to the actual human
Kenya have less than half of the staff members they resource needs at the facilities and for their catchment
require in every cadre, except in the “all other staff” cat- populations.
egory. Pharmaceutical, counselling and other clinical
staff are especially underrepresented. Further taking In addition to regular staff members, health facili-
into account the numbers of additional staff reported ties may have at their disposal seconded workers,
as being required to meet work requirements among those who are not on the facility’s payroll but who are
those facilities without a staffing norm, the perceived assigned to work there and are paid by another entity.
staffing shortage becomes even more apparent. Frequently, nongovernmental organizations or groups
90
Figure 7.4 Percentage of health workers requested by facility managers to meet staffing requirements
who are currently in post, by facility type, Kenya 2004 SPA
97 99
100 89
75 78 78
80 73
% of requested staff
67 65
63
currently in post
59 58 56
60 54
50 49 50
40 33 33 33
20
<1
0
on T
ds t
ds
ds
ds
ds
ds
be tien
al VC
e
be
be
be
be
be
a
p
19
+
in
1–
–4
–9
0
d-
10
o
10
20
50
an
N
st
Facility type/size
Medical Nursing/midwifery All staff
Table 7.5 Percentage of facilities reporting having at least one seconded health workera, by cadre,
according to management authority of the facility, Kenya 2004
Cadre Facility managing authority

Public Private Total
For-profit Nongovernmental Faith-based
organization organization
(n=246) (n=63) (n=21) (n=110) (n=440)
Medical practitioners 2 25 48 8 8
Nursing & midwifery personnel 3 17 5 13 6
Laboratory staff 1 14 0 1 2
Counselling/social work 1 2 15 1 2
a. Seconded worker defined as a person who works at the facility routinely (either full or part time) providing health-care services, but
who is paid by another entity.
dedicated to offering specialized services, such as lab- seconded counsellors. Government-operated facilities
oratory or HIV-related services, will second their staff to are least likely to have seconded workers. The survey
provide services in facilities where there is a perceived did not collect information on the source of second-
health worker shortage, and perhaps also to provide ments, but most are likely to be foreign health workers
on-the-job training to a regular staff member when the hired by international organizations and seconded to
service is being newly introduced at the facility. Table a local affiliate. Monitoring secondments should be
7.5 provides information from Kenya on seconded assessed not only in terms of perceived health worker
health workers according to the facility’s manage- shortages, but also in terms of strategies for maximiz-
ment authority. Facilities in the nongovernmental sector ing benefit and ensuring sustainability in the provision
report a widespread presence of seconded workers: and quality of services that currently depend on non-
nearly half (48%) of these facilities have at least one regular personnel.
medically trained seconded worker, and 15% have
91
Distribution personnel ratios are expected in facilities with more

The distribution of health workers helps to provide inpatient beds, as inpatient care is highly dependent
some indication of the accessibility of health services on 24-hour nursing services.
to the population, the quality of services offered and
the rational allocation of human resources. One indica- Table 7.7 presents data on the distribution of health
tor that is useful and measurable via HFAs is the skill workers by facility management authority in Kenya. The
mix, which provides information against which (regular) government is the largest provider of health services
staffing patterns can be analysed. In Kenya, the data and also the main employer of Kenyan health labour.
revealed the presence of a median ratio of 1.9 other staff About two thirds (62%) of facility-based health workers
to medical, nursing and midwifery personnel across all are in the government sector. Less than a fifth can be
facilities (Table 7.6). There was little marked difference found in either private for-profit facilities (17%) or faith-
in this ratio by the size of the facility. However, the ratio based organizations (18%), and only 3% in facilities
of nursing and midwifery personnel to medical practi- operated by nongovernmental organizations. This gen-
tioners does vary widely by facility size. The median eral pattern tends to hold for most categories of health
ratio is more than twice as high in very large facilities, workers, with the notable exceptions of pharmaceutical
those with 100 or more inpatient beds, compared to staff, who are found more often in the private for-profit
smaller facilities with 10–50 beds. sector (28%), and counsellors, who are more often in
the nongovernmental sector (16%), where many VCT
This may reflect staffing patterns that take into account sites are found.
client load and the higher costs associated with some
categories of personnel compared to others. On the In another example from Kenya, Table 7.8 shows the
one hand, skills substitution may be common in smaller distribution of health workers by geographical region, in
facilities, where human resources are less numerous this case the country’s eight provinces. The largest con-
and even highly skilled staff can be expected to per- centrations are in Nairobi and Rift Valley provinces: 25%
form tasks outside their initial field of specialization or and 23% respectively. While in the latter the number of
undertake administrative responsibilities such as com- health workers roughly corresponds with that of the pro-
pleting statistical records. On the other hand, larger vincial population, in Nairobi this is not the case. Here
facilities with higher client loads and more complex the concentration of health workers is roughly three
services may find it more practical to employ more times greater than the total population. Conversely,
specialized staff, such as pharmaceutical, counsel- North-Eastern and Western provinces appear likely to
ling and management staff. Higher nursing to medical be underserved, as these provinces’ share of the coun-
try’s health workers are disproportionately low.
Table 7.6 Skills mix of facility-based health Figure 7.5 Percentage of health workers and
workers, by type of facility, Kenya 2004 of the total population located in urban areas,
Zambia 2006 HFC
Facility Median ratio
type/size + "! !" 29
All other staff to Nursing &
medical, nursing midwifery + "! !" % 47
& midwifery personnel *"
personnel to medical
*" "
practitioners
*" ""
No beds 1.7 .. ) "
"!"
1–9 beds 1.9 ..
'( " "
10–19 beds 2.1 2.5 &"
"" "
20–49 beds 1.8 2.5
#$ "! !" %
50–99 beds 1.9 3.5 " "
"
100+ beds 1.5 5.5
!""
All facilities 1.9 .. !
.. Not calculated due to small number of surveyed facilities with

medical staff in post.

Source: Kenya 2004 SPA (22). Data weighted to reflect survey ,(
sampling procedures. VCT stand-alone sites are excluded due
to small number with medical or nursing staff in post. Source: Herbst and Gijsbrechts (23).
92
Table 7.7 Percentage distribution of health workers in post at health facilities by managing authority,
according to cadre, Kenya 2004
Cadre Facility managing authority

Public Private Total
a a
For-profit NGO FBO
(n=246) (n=63) (n=21) (n=110) (n=440)
Medical practitioners 60 22 4 14 100
Laboratory personnel 53 17 4 26 100
Pharmaceutical personnel 56 28 4 12 100
Other clinical staff 70 10 7 13 100
Health information/records 74 13 2 11 100
technicians
Other 56 22 2 20 100
Total 62 17 3 18 100
a. NGO = nongovernmental organization, FBO = faith-based organization.
Table 7.8 Percentage distribution of facility-based health workers and of the total population by
geographical region, Kenya 2004
Cadre Province Total

Nairobi Central Coast Eastern North- Nyanza Rift Western
Eastern Valley
Medical practitioners 22 10 10 15 1 10 26 6 100
Nursing & midwifery 22 14 8 17 1 10 22 6 100
personnel
Laboratory personnel 20 11 11 14 1 11 25 7 100
Pharmaceutical personnel 51 7 8 9 1 7 13 4 100
Counselling/social work 13 7 16 13 1 11 29 10 100
Other clinical staff 21 13 4 11 2 11 31 7 100
Health information/ 18 16 8 10 1 12 28 7 100
records technicians
Other 33 2 13 16 0 10 21 5 100
Total health workers 25 10 11 14 1 10 23 6 100
a
Total population 8.1 12.2 8.7 15.6 4.1 14.8 24.9 11.6 100.0
a. Based on the estimated 2004 national population (32 808 268).
Geographical imbalance of the health workforce can a third (29%) of the population, roughly half (47%) of
also be seen in another context, notably that of Zambia. all health workers are found in urban facilities. Three
Figure 7.5 compares the proportion of health workers quarters of physicians and half of nursing and mid-
and of the total population located in Zambia’s urban wifery personnel are urbanites. Among the different
areas. Although urban dwellers account for less than cadres of health workers examined, only environmental
93
health technicians are underrepresented in urban initiatives, where it is not uncommon to find wide dif-
areas compared to the general population. For some ferences in the training received among workers with
specializations, workers in higher-skilled categories the same occupational title, due to changes over time
are more likely to be situated in urban areas compared in the standards of both basic and advanced educa-
to their lower-skilled counterparts: physicians versus tion required for a health qualification. Information on
paramedical practitioners; pharmacists versus phar- educational attainment is also useful for enhancing
maceutical technicians. mapping of national occupational titles to the ISCO
standard, which categorizes occupations according to
Further analysis of the workforce distribution by district the skill level and skill specialization normally required
reveals that 2224 or 18% of Zambia’s 12 219 facility- for competent performance.
based health workers are located in the national capital,
Lusaka. Thirteen of the country’s 71 districts have no Figure 7.6 offers an illustrative example of the levels
physicians assigned to a health facility, while only three of HRH educational attainment as reported by inter-
districts (Lusaka, Kitwe and Ndola) are home to 67% viewed health workers in Kenya. As can be expected,
of all facility-based physicians (23). Such lopsided dis- while the level of basic education prior to training for
tributions of HRH may have serious implications for a health qualification remains similar for all types of
equitable accessibility, coverage and quality of health- service providers, the number of years of advanced
care services in rural and underserved areas. technical training varies across cadres. Medical doc-
tors (including specialist and generalist physicians)
have the highest overall educational attainment, while
Health workforce education,
nursing aides and counsellors have the lowest.
motivation and performance
Information on the levels and fields of education and Profiles of the number of years health workers have
training of the health workforce is critical for tailoring been employed at a particular facility help provide
health labour training needs and for understanding the insights related to professional experience, ongoing
overall technical capacity of various health worker cad- training needs and staff retention. Table 7.9 presents
res within a country. This is particularly true in contexts interview responses from the Kenya SPA on staff mem-
of rapid scaling up of human resources development bers’ years of service in their current location by the
Figure 7.6 Median number of years of education Figure 7.7 Percentage of health workers who
and training among health workers, by cadre, received in-service training in the past 12
Kenya 2004 SPA months, and who received personal supervision
in the last 6 months, by type of facility, Kenya
2004 SPA
All health workers 12.0 3.0
Other 12.0 3.0 68

All facilities
67
Counsellor 12.0 (1 month)
60
Hospital
Laboratory technician 12.0 3.0 71
12.0 1.0 Maternity 60

Nursing aide
centre 67
Enrolled midwife 11.0 3.0
Health 70
Enrolled nurse 12.0 3.0 centre 67
60
Registered midwife 12.0 3.0 Clinic
67
Registered nurse 12.0 3.0
72
Dispensary
Paramedical practitioner 12.0 3.0 65
VCT stand- 86
Physician – generalist 12.5 6.0
alone 83
Physician – specialist 13.0 6.5
0 20 40 60 80 100
0 5 10 15 20 % of health workers
Median number of years (excluding management staff)
Basic education Technical training Supervised Received training
Source: Kenya 2004 SPA (22). Source: Kenya 2004 SPA (22).
94
Table 7.9 Median number of years of service in current position among facility-based
health workers, by type and management authority of facility, Kenya 2004
Medical practitioners Nursing & midwifery All health workers

personnel
Facility type
VCT stand-alone .. 4.1 1.0
Dispensary 3.0 3.0 2.0
Clinic 4.6 4.3 3.0
Health centre 2.0 2.5 3.0
Maternity centre 1.1 7.2 2.0
Hospital 3.0 5.0 5.0
Managing authority
Public 3.0 4.0 3.0
Private 2.0 3.0 3.0
All facilities 3.0 4.0 3.0
.. Not calculated due to small number of surveyed facilities with medical staff in post.
type and management authority of the facility. Overall, personally supervised during the six-month period
half of the facility-based staff reported having been at preceding the survey. Service providers at VCT sites
their current employment for a short period of time, were most likely to have received supervision (86%).
about three years, with little variation in terms of the The concentration of health workers who had recently
facility’s managing authority. Nursing and midwifery received training or supervision in these centres prob-
personnel at hospitals and especially maternity centres ably reflects the high attention that HIV-related services
tended to have been in their current position for longer. are receiving worldwide. The survey did not collect
Conversely, the median number of years of service was information on the competencies acquired or used fol-
lowest among physicians at maternity centres (most of lowing the trainings or on the quality of supervision;
which are privately managed in Kenya) and staff at VCT many HFAs do gather at least some information on
stand-alone facilities, which are a relatively new cate- points discussed or activities conducted during the
gory of service and facility type. supervisory visit.
Training and supportive supervision may not only Among the four HFA survey and census tools consid-
improve the quality of health worker performance but ered here, none presently collects information on staff
may also act as incentives that motivate health work- incomes or wages, and as such these sources do not
ers in their jobs. Figure 7.7 presents data from Kenya allow analysis of financial incentives among health
on the proportion of facility-based health workers who workers. In the absence of data on wages and salaries,
received formal in-service training during the 12-month other incentives that might influence worker motivation
period before the survey, that is, structured training and performance and favour retention were exam-
sessions not including individual instruction received ined. The HRH literature suggests that the availability
during routine supervision. Overall, excluding manage- of a clear scope of work and other non-monetary ben-
ment staff, two thirds (67%) of health workers reported efits contribute to improved worker performance and
having received in-service training with little variation reduced attrition (24, 25). An example of this type of
in this proportion across the various types of facilities. information from the Kenyan context is provided in Table
The main deviation is noted with regard to VCT sites, 7.10. Only 8% of health workers were able to produce
where the proportion is by far the greatest (83%). a written scope of work at the time of interview, with lit-
tle difference in this percentage by cadre. Somewhat
Additional findings from Kenya on support supervision wider variations exist by facility type, but even where
received among health-care providers are presented the proportions were highest (at maternity centres and
in Figure 7.7. Excluding management staff, about two clinics) only about one of every ten workers was able
thirds (68%) of health workers reported having been to produce a written scope of work. These numbers
95
Table 7.10 Percentage of facility-based health workers with written job descriptions, perceived
promotion opportunities and other non-monetary incentives, by cadre and type of facility, Kenya 2004
Percentage of health workers

With written job Who perceive promotion Who receive incentives
description at the time opportunities other than salarya
of interview
Cadre
Physicians 8 50 40
Nursing & midwifery personnel 8 42 58
Laboratory personnel 7 35 35
Counselling/social work 8 43 60
Other 6 43 44
Facility type
VCT stand-alone 8 41 64
Dispensary 6 41 51
Clinic 10 37 56
Health centre 6 37 47
Maternity centre 13 13 63
Hospital 9 50 56
All health workers 8 42 54
a. Non-monetary incentives include subsidies for medicines or other goods, uniforms or other clothing, food and training.
may be underestimating the situation, however, as an One way to look at health worker performance is by
additional 30% of health workers reported that they had means of the absenteeism rate. This was examined
written scopes of work but were unable to show them to using SAM data for Kenya for selected urban areas
the interviewer (results not shown). and rural districts where the facility census was con-
ducted: Kilifi, Kisumu, Mombasa, Nairobi, Nakuru and
Further analysis of Kenya SPA data show that 42% of Thika (26). Overall, less than one third of all assigned
health workers perceive promotion opportunities in their health workers (that is, on the duty list) were actually
current job (Table 7.10). Physicians and hospital-based found at their post on the day of interview (Figure 7.8).
staff are most likely to say they perceive promotion Physicians were least likely to be present, followed by
opportunities. More than half (54%) of health workers nursing and midwifery personnel, with only about one
report receiving other non-monetary incentives from in four of the assigned staff at their post. Several rea-
their employer, including subsidies or discounts for sons could account for why health workers may not
medicines or other goods, uniforms or other clothing, be present at their duty post on a given day, including
food or training. Considerable variations are observed scheduled leave, unscheduled absence or unexplained
according to cadre (from 35% of laboratory staff to absence. A more detailed case study on assessment
60% of counsellors) and by facility type (with the high- of worker absenteeism can be found in Chapter 11 of
est percentage being at VCT sites, the workplace of this Handbook.
about a quarter of Kenya’s counsellors). Although the
inclusion of training opportunities as a form of incen- Information on working hours is important for HRH plan-
tive may have inflated the results for the latter indicator, ning and can be used to calculate, for example, health
this information offers some guidance on the potential system capacity in terms of full-time equivalents for
differences in compensation schemes by worker and job positions, and to support the development of flex-
facility characteristics and the related implications for ible management practices that could favour worker
staff performance and retention. retention. Table 7.11 presents HFA data for Kenya on
96
Table 7.11 Percentage distribution of facility-based health workers by number of hours normally
worked per week, according to cadre, Kenya 2004
Cadre Normal hours worked per week in facility Total

40+ 30–39 20–29 <20
Physicians 83 7 2 8 100
Laboratory personnel 98 1 1 0 100
Nursing aides 78 9 0 13 100
Other 74 6 9 11 100
All health workers 87 6 3 4 100
Figure 7.8 Percentage of assigned health workers outside the health sector or not working more hours
present on the day of the assessment, by cadre, due to voluntary or involuntary reasons).
Kenya 2005 SAM (selected districts)
7.5.3 Exits from the workforce

All health workers 29% Some HFAs have collected information to assess
workforce retention. Challenges with collecting infor-
Other facility staff 45% mation on HRH transitions and exits may make such
data incomplete, however, particularly for measuring
Health information/ international outflows. Figure 7.9 shows estimated attri-
49%
records technicians tion rates for Nigeria’s public sector health workforce
by cadre and for urban and rural areas. According
Pharmacists 47% to these results, overall attrition is highest for facility-
based physicians compared to other categories of
Laboratory
52% health workers. Attrition is higher for rural workers com-
technicians
pared to their urban counterparts, with the exception of
Nursing & midwifery pharmaceutical personnel and community health work-
personnel 28%
ers. The latter are recruited and trained specifically to
work in rural areas, so these staff may be less likely to
Physicians 23% have many tempting alternative employment opportuni-
ties beyond their area of residence.
0% 20% 40% 60% 80%
In the same context, resignation is by far the most com-
% present on the day of visit
mon reason for workforce attrition, the reason cited
Source: Kenya 2005 SAM (26). for nearly half (46%) of measured health worker exits
(Figure 7.10). Although the underlying causes of resig-
nation were not determined in the present assessment,
experience suggests that poor service conditions are
the number of hours normally worked per week in the often at the root. Twenty-three per cent of worker exits
facilities where health workers were interviewed. Most are attributable to involuntary termination of employ-
(91%) of the facility-based staff are full-time employees, ment. Retirement and death account for about one in
usually working at least 40 hours per week. Laboratory seven exits each. While the available data do not offer
and nursing and midwifery personnel are least likely to a complete picture of the dynamics of health worker
work part time, while counselling staff are most likely to attrition, they can be used to highlight areas where pro-
be part time. Almost two in every ten physicians work gramme managers and policy decision-makers can
in the facility part time. Areas for further exploration in further invest in research and retention interventions.
future HFA tools could include the nature of work activ-
ities outside the facility among part-timers (for example Lastly, as a crude estimate of the health workforce
whether in another facility, in research or teaching, regeneration ratio, the ratio of health workforce increase
97
Figure 7.9 Attrition rates for health workers in public sector facilities, by cadre, Nigeria 2005
5 4.7
4
3.4
Attrition rate
3
2.3 2.3 2.3 2.2
2.0 2.0
2
1.4 1.4 1.3 1.4
1.0 1.1
1
<0.1
0
Physicians Nursing & Laboratory Pharmaceutical Community health
midwifery personnel personnel personnel workers
Urban Rural Total
Figure 7.10 Percentage distribution of outgoing 7.6 Summary and conclusions

health workers by reason for leaving the
workforce, Nigeria 2005 There is worldwide consensus that the HRH situation
is in a state of crisis and that low- and middle-income
countries are most at risk. It is also widely accepted that
overcoming the crisis requires effective monitoring of
the three stages of the workforce lifespan (entry, active
Resigned: Retired:
46% workforce and exit), which in turn requires access to
16%
measurement methodologies and analytical tools that
can be used to collect and analyse workforce data in a
timely and credible manner.
This chapter has documented use of HFAs as one such

Terminated:
23% methodology and has presented examples of the types
Death: of HRH information it can provide. Potential statistics
15%
that can be produced from HFA sources include health
workforce stock and supply, education, skills mix, geo-
Source: Chankova et al. (14). graphical distribution, productivity and other contextual
practices supportive of efficient job performance (for
example adequate resources and infrastructure,
in-service training, management and supervision prac-
tices, and incentives) and estimates of staff entry and
retention. When analysed in conjunction with national
due to new graduates from health professions educa- staffing norms (when they are in place), HFA data can
tion institutions joining the facility-based workforce can be used to ascertain the degree to which perceived
be compared to the attrition rate at these facilities. The staffing needs are being met, and to support effective
results in Table 7.12 suggest that for every physician planning for HRH education and training, recruitment,
departing Nigeria’s public sector health workforce, management and eventual retirement.
three new ones enter. The estimated ratios of incoming
to outgoing staff are well above 2:1 among laboratory A number of limitations of currently available data from
staff and community-based workers. However, fewer HFA sources were also discussed. Among these is a
new graduates of nursing and midwifery training dearth of information on migration of health workers,
schools seem to be entering Nigeria’s public workforce especially international outflows. By nature, HFAs do
compared to the number of nurses and midwives who not directly collect data on the pre-service stage, or on
are leaving it. workers outside facility-based service delivery points.
As discussed elsewhere in this Handbook, the need
to improve methods for monitoring the pre-service and
98
Table 7.12 Comparison of the increase in stock of the public sector health workforce from new
graduates with the attrition rate, Nigeria 2005
Cadre New incoming graduates Attrition rate Ratio of new incoming

from health professions staff to outgoing staff
education institutions as
% of existing stock
Physicians 7.7 2.3 3.3
Nursing & midwifery personnel 1.1 1.4 0.8
Laboratory personnel 3.4 1.3 2.6
Pharmaceutical personnel 3.6 2.2 1.6
Community health workers 3.3 1.4 2.3
exit stages is not limited to HFA sources. However, exist- any) examples of countries or stakeholders using this
ing HFA tools could be strengthened by capturing the information for HRH policy and planning.
numbers of newly recruited staff that are fresh gradu-
ates, geographical in-migrants (internal or international) Since the overall objective of most HFAs is to assess
or returning to the workforce after an extended period health services, the tools tend to cover a broad range of
of absence, and potentially perceptions among current topics, including not only facility staffing but also infra-
staff of the main reasons their former colleagues left. structure, services delivery, supplies and equipment,
protocols and client satisfaction. As such, they do not
Other basic information that tended to be lacking in generally allow for detailed analysis of any specific
the HFA tools examined in this chapter (although not component; expanding the information collected on
an exhaustive review) included workforce remunera- HRH would enable more detailed analyses but would
tion and sex distribution. Data on workers’ wages and also increase the complexity of the assessments, with
salaries are important both for costing and budgeting implications for interviewer training, fieldwork logistics,
of strategies for scaling up services delivery and HRH average length of interview, data processing and anal-
development initiatives, and for evaluation of mon- ysis, and of course budgeting. It is possible to conduct
etary incentives influencing provider performance. stand-alone surveys among health-care providers, but
Only limited information was collected in the reviewed the sustainability of this approach, especially in low-
instruments on non-monetary benefits. In addition, income countries, needs to be examined.
understanding the gender dimension can help frame
actions that may minimize the deleterious impact of Ideally, a comprehensive HRH analysis would optimize
gender imbalance on the workforce and service deliv- the use of periodically collected HFA data integrated
ery, particularly when such imbalances are not detected in the national health information system with routinely
early (see also 27, 28). At the time of this report, the (continuously) collected data drawing on administrative
SPA and SAM data collection tools were already being sources (the latter is further elaborated in Chapter 9 of
updated to incorporate gender in future assessments. this Handbook). Facility-based assessments should be
carried out every seven to ten years in order to validate
One important constraint that is inherent in HFAs is the and augment the information provided through routine
potential for sampling bias where a complete and accu- data sources (for example new deployments, support
rate sampling frame, or listing of facilities for selection, supervision, worker absenteeism, job vacancies and
is not available, as is the case in many low- and mid- workforce exits). This periodicity is recommended to
dle-income countries, especially for the private sector. help keep costs low and allow sufficient time between
Sampling variations, and differences in questionnaire exercises so that changes can be observed. Most
design and other non-sampling issues, have often countries can develop the technical capacity to carry
hindered comparability of HFA data across sources, out HFAs after one round with external technical assist-
hampering monitoring of trends across regions and ance. However, ongoing international financial support
over time. Despite the availability of rich, time-specific is often needed, as well as technical cooperation to
information on HRH from HFA data, there are limited (if ensure the quality (and cross-national comparability)
of the data collected. Further investment is frequently
99
needed in data analysis and dissemination strate- 10. Suzuki Y et al. Capital investment planning using
gies that put the information retrieved within the reach the geographic information systems. Journal of
International Health, 2006, 21(S):89.
of policy-makers (29). Institutionalizing HFA capacity,
whether within the ministry of health or another auton- 11. Service Availability Mapping (SAM). World Health
omous or semi-autonomous agency (such as private Organization (http://www.who.int/healthinfo/systems/
serviceavailabilitymapping/en/, accessed 20
survey and research firms), should be prioritized if HRH
January 2009).
information from HFAs is to become a fully integrated
12. Workload indicators of staffing need (WISN): a
component of the health information system and widely
manual for implementation. Geneva, World Health
used to bridge information gaps for HRH planning, Organization, 1998 (http://www.who.int/hrh/tools/
monitoring and development. workload_indicators.pdf, accessed 20 January 2009).
13. Service Provision Assessment (SPA) surveys.
Macro International (http://www.measuredhs.com/
References aboutsurveys/spa/start.cfm, accessed 12 February
2009).
1. The world health report 2006: working together for
health. Geneva, World Health Organization, 2006 14. Chankova S et al. A situation assessment of human
(http://www.who.int/whr/2006, accessed 10 January resources in the public health sector in Nigeria.
2009). Bethesda, MD, Partners for Health Reformplus
Project, Abt Associates Inc., 2006 (http://www.
2. Kinfu Y. Assembling global health workforce healthsystems2020.org/content/resource/
statistics: challenges and implications. Presented detail/1704/, accessed 20 January 2009).
at the World Health Organization/Health Metrics
Network Technical Working Group Meeting on 15. MEASURE Evaluation. Facility Audit of Service
Health Workforce Statistics, Geneva, 11–13 July Quality. Chapel Hill, NC, Carolina Population Centre,
2006. 2000.
3. Annell A, Willis M. International comparison of health 16. Assessment of Human Resources for Health: survey
care systems using resource profiles. Bulletin of the instruments and guide to administration. Geneva,
World Health Organization, 2000, 78(6):770–778 World Health Organization, 2002 (http://www.who.
(http://www.who.int/bulletin/archives/78(6)770.pdf, int/hrh/tools/hrh_assessment_guide.pdf, accessed
accessed 20 January 2009). 12 February 2009).
4. International Health Facility Assessment Network. 17. Dehn J, Reinikka R, Svensson J. Survey tools for
Health facility assessment: relevant resources/ assessing performance in service delivery. In:
supporting documents and mapping resources Bourguignon F, da Silva LP, eds. Evaluating the
– annotated bibliography. MEASURE Evaluation poverty and distributional impact of economic
Working Paper No. WP-08–107. Chapel Hill, NC, policies. Washington, DC, Oxford University Press
MEASURE Evaluation, 2008 (http://www.cpc.unc. and World Bank, 2003.
edu/measure/publications/pdf/wp-08–107.pdf, 18. Ferrinho P et al. Dual practice in the health sector:
accessed 20 January 2009). review of the evidence. Human Resources for
5. International Health Facility Assessment Network. Health, 2004, 2(14) (http://www.human-resources-
Flow chart of steps to conduct a health facility health.com/content/2/1/14, accessed 20 January
assessment. Chapel Hill, NC, MEASURE Evaluation, 2009).
2008 (http://www.cpc.unc.edu/measure/publications/ 19. International Standard Classification of Occupations.
pdf/ms-08–28.pdf, accessed 20 January 2009). International Labour Organization (http://www.ilo.org/
6. MEASURE Evaluation. Sampling manual for facility public/english/bureau/stat/isco/index.htm, accessed
surveys for population, maternal health, child 11 January 2009).
health and STD programs in developing countries. 20. International Standard Classification of Education:
MEASURE Evaluation Manual Series No. 3. Chapel ISCED 1997. Paris, United Nations Educational,
Hill, NC, Carolina Population Centre, 2001. Scientific and Cultural Organization, 1997 (http://
7. International Health Facility Assessment Network. www.uis.unesco.org/TEMPLATE/pdf/isced/
Profiles of health facility assessment methods. ISCED_A.pdf, accessed 10 January 2009).
Arlington, VA, MEASURE Evaluation, USAID and 21. International Standard Industrial Classification of All
World Health Organization, 2008 (http://www.cpc. Economic Activities, fourth revision. Statistical Papers
unc.edu/measure/publications/pdf/tr-06–36.pdf, Series M, No. 4/Rev.4. New York, United Nations
accessed 20 January 2009). Statistics Division, 2008 (http://unstats.un.org/unsd/
8. Japan International Cooperation Agency. Health demographic/sources/census/2010_PHC/docs/
Facility Census. Kurume, Japan, St Mary’s Hospital, ISIC_rev4.pdf, accessed 11 January 2009).
2004. 22. National Coordination Agency for Population and
9. Suzuki Y, Hozumi D. Health Facility Census: Development (NCAPD), Ministry of Health, Central
background, meaning, and its methodology. Journal Bureau of Statistics, ORC Macro. Kenya Service
of International Health, 2005, 20(S):110. Provision Assessment Survey 2004. Nairobi, Kenya,
NCAPD, 2005 (http://www.measuredhs.com/pubs/
pdf/SR139/SR139.pdf, accessed 20 January 2009).
100
23. Herbst CH, Gijsbrechts D. Information on stock,

profiles, and distribution of health workers in Zambia:
analysis of the Health Facility Census data. World
Bank Human Resources for Health Program Paper.
Washington, DC, World Bank, 2007.
24. Hornby P, Forte P. Human resource indicators and
health service performance. Staffordshire, United
Kingdom, Keele University Centre for Health
Planning and Management, 2003 (http://www.who.
int/hrh/en/HRDJ_1_2_03.pdf, accessed 21 January
2009).
25. Diallo K et al. Monitoring and evaluation of human
resources for health: an international perspective.
Human Resources for Health, 2003, 1:3 (http://
26. World Health Organization and Kenya Ministry
of Health. Service Availability Mapping: Kenya
report. Geneva, World Health Organization and
Kenya Ministry of Health, 2005 (http://www.who.int/
healthinfo/systems/samreportkenya.pdf, accessed
21 January 2009).
27. Hojat M, Gonnella J, Zu G. Gender comparisons
of young physicians perceptions of their medical
education, professional life, and practice: a follow
up study of Jefferson Medical College graduates.
Academic Medicine, 1995, 70(4):305–312.
28. Reamy J, Pong R. Physician workforce composition
by gender: the implications for managing physician
supply. Paper presented at the 4th FICOSSER
General Conference, Cuernavaca, Mexico, 28–30
July, 1998.
29. AbouZahr C, Adjei S, Kanchanachitra C. From data
to policy: good practices and cautionary tales.
Lancet, 2007, 369:1039–1046.
101
102
Use of population census data
8 for gender analysis of the

health workforce
RENÉ LAVALLÉE, PIYA HANVORAVONGCHAI, NEERU GUPTA
8.1 Introduction 8.2 Importance of gender

Despite the undoubted importance of monitoring the considerations in health
health workforce and impacts on health systems per- workforce analysis
formance, the empirical evidence to support policy
Women make up about 40% of the estimated global
formulation is often fragmented. Many sources that can
working population (3). Within the health sector, in
potentially produce information relevant to this issue
many countries women comprise over 75% of the work-
remain underused in health research, especially among
force (4), making them indispensable as contributors to
low- and middle-income countries. Although a range of
the delivery of health services. However, gender issues
standard statistical sources can be exploited for con-
remain a neglected area in most approaches to HRH
ducting human resources for health (HRH) assessments
policy and planning (5).
– including national population censuses – their poten-
tial for HRH monitoring has generally not been met.
In many countries, women tend to be concentrated
in the lower-status health occupations, and to be a
Population censuses can be a key source for statis-
minority among more highly trained professionals. In
tics describing HRH, providing precise information on
particular, the distribution of women by occupational
the stock and composition of the health workforce and
category tends to be skewed in favour of nursing and
on distribution by spatial units and sociodemographic
midwifery personnel and other “caring” cadres such as
characteristics (1, 2). One of the main strengths of
community health workers (6). Women are often poorly
census data for HRH analysis is the possibility of disag-
represented in other categories, for example physi-
gregating individual-level information by sex. Censuses
cians, dentists, pharmacists and managers.
offer an advantage over survey sources in that they do
not suffer from problems of sample sizes too small to
The underrepresentation of women in managerial and
allow estimates for specific subgroups. Moreover, as
decision-making positions may lead to less attention to
noted in Chapter 7 of this Handbook, health facility
and poorer understanding of both the particular fea-
assessments have tended to be gender blind when it
tures of working conditions that characterize much of
comes to monitoring the staffing situation.
women’s employment, and the health-care needs spe-
cific to women. In many contexts, access to female
The objective of this chapter is to present selected
providers is an important determinant of women’s
findings from multicountry analyses exploring gender
health service utilization patterns (7, 8).
dimensions of the health workforce using census data.
Including this introduction, the chapter is divided into
Omission of gender considerations may also lead to
five parts. In the next section, an argument is presented
inadequate health system responsiveness to the needs
for the importance of gender mainstreaming in health
of men: for example, reproductive health services are
workforce analysis. Next, the sources of census data
often not set up so as to encourage male involvement
used here are presented. Fourth is the empirical anal-
(7). Better recognition of and information on the role of
ysis, with attention paid to each of the three stages of
men as caregivers may help to “de-gender” gender
the working lifespan framework for HRH assessment.
norms in health service provision (9).
The chapter concludes with some recommendations
for promoting the use of sex-disaggregated data, nota- Gender analysis of the health workforce may reveal that
bly from census sources, as a step towards monitoring health systems themselves can reflect or even exac-
and evaluation of gender-sensitive human resources erbate many of the social inequalities they are meant
policy planning and management. to address and be immune from (6). Understanding
103
and addressing the gender aspect of the health work- transitional economies where a variable for occupa-
force require better measures of women and men in tion was included that allowed identification of health
the health workforce to help identify and prioritize HRH occupations.
planning and management interventions. t For one country, Thailand, access to census micro-
data was obtained from the National Statistical
Office with support from the International Health
8.3 Using census data for Policy Program, Thailand (14), a national research
programme on health priorities related to health sys-
health workforce analysis tems and policy.
Strengthening the evidence base on gender and the t For another country, Uganda, data were drawn from
health workforce in low- and middle-income coun- a special census volume on health workers pro-
tries is especially critical. Most of the (scant) available duced in collaboration with the Developing Human
research on gender and HRH refers to countries with Resources for Heath Project (15). The publication
developed market economies, especially the United included tabulations and maps of the stock and
States of America (6). In particular, while census micro- distribution of health occupations, as well as infor-
data archives exist for most countries, access to and mation on the variables used in the compilations.
use of these data for health systems research has gen-
erally been limited. For some countries, analyses of In the present analysis, in order to enhance cross-
census data can be facilitated through collaborative national comparability, health workers were defined
research projects aiming to disseminate microdata for according to the self-reported main occupation among
public use. In only a very limited number of low- and the economically active population as recorded in
middle-income countries have national census-based the census, with titles mapped where possible to the
HRH analyses been conducted and disseminated. International Standard Classification of Occupations,
1988 revision (ISCO-88) (16). This includes the follow-
A gender-based analysis of the health workforce was ing broad groupings:
conducted in 13 countries across different regions and t health professionals (except nursing and midwifery):
contexts (Table 8.1). As discussed in Chapter 2 of this physicians, dentists, pharmacists and other profes-
Handbook, censuses with questions designed to col- sional-level health occupations;
lect data on the nature of work activity can provide t nursing and midwifery personnel: nursing profes-
valuable information for HRH analysis (Box 8.1). The sionals, midwifery professionals, nursing associate
application of international standards for data collec- professionals, midwifery associate professionals;
tion and processing facilitates production of statistics t other health service providers: health technicians
that are comparable across countries and over time on and associate professionals (for example medical
many aspects of labour dynamics (10). Cross-national assistants, dental assistants, laboratory technicians,
comparisons of HRH indicators can provide valuable therapeutic equipment technicians), traditional
opportunities for gaining insights into workforce issues and complementary medicine practitioners, per-
that are of major concern to many countries, and learn- sonal care workers (including institution-based and
ing how countries have dealt successfully or otherwise home-based nursing aides) and others not identi-
with these issues (11). fied elsewhere;
t health management workers: administrators
The data used for the analysis were drawn from three and supervisors in health services and similar
sources: occupations.
t For 11 of the countries – Argentina, Brazil, Cambodia,
Chile, Ecuador, Hungary, the Philippines, Romania, Other efforts were also undertaken to enhance compa-
Rwanda, South Africa and Viet Nam – micro- rability given the particularities of individual censuses;
data samples were obtained from the Integrated in particular, the analysis excludes workers abroad,
Public Use Microdata Series (IPUMS), a collabo- who were counted in the Philippines but not in the other
rative project dedicated to collecting, preserving, countries.
harmonizing and disseminating census data and
documentation from around the world for social and The number and density of workers with a self-reported
economic research (12). As of late 2007, IPUMS- health occupation at the time of the census, as per the
International had created a unique census microdata boundaries retained here, can be found by country in
collection consisting of 80 censuses from 26 coun- Table 8.1. As could be expected, across countries, the
tries (13). For the present analysis, data from the density of health workers tends to increase with the
last census round (covering the period from 1995 to national income level. The highest densities are found
2004) were used for countries with developing and in the upper-middle-income countries of Hungary and
104
Use of population census data for gender analysis of the health workforce
Table 8.1 Countries and sources of census data
Country Income Regionb Census implementing agency Year Workers with a

categorya health occupation
Number Per 10 000
population
Argentina Upper Americas National Institute of Statistics 2001 418 530 115.4
middle and Censuses
Brazil Upper Americas Institute of Geography and 2000 1 463 001 86.1
middle Statistics
Cambodia Low Western Pacific National Institute of Statistics 1998 26 940 23.6
Chile Upper Americas National Institute of Statistics 2002 133 580 88.2
middle
Ecuador Lower Americas National Institute of Statistics 2001 79 290 65.3
middle and Censuses
Hungary Upper Europe Central Statistical Office 2001 169 960 166.5
middle
Philippines Lower Western Pacific National Statistics Office 2000 360 217 48.0
middle
Romania Upper Europe National Institute of Statistics 2002 295 880 138.4
middle
Rwanda Low Africa National Institute of Statistics 2002 10 230 12.1
South Africa Upper Africa Statistics South Africa 2001 256 393 60.0
middle
Thailand Lower South-East Asia National Statistical Office 2000 294 905 48.4
middle
Uganda Low Africa Bureau of Statistics 2002 57 508 23.5
Viet Nam Low Western Pacific General Statistics Office 1999 253 500 33.2
a. Income category as classified by the World Bank according to gross national income per capita.
b. Regions as classified by the World Health Organization.
Box 8.1 Questions used in the population census to determine main

occupation, selected countries
Brazil (2000): How many jobs did you have during the week of July 23–29, 2000? (response options: “one”,
“two or more”); What was the principal job done during [this] week? (open question).
Chile (2002): In which of the following situations did you find yourself during the last week? (choice of
10 response options); What occupation or type of work do you perform, or did you perform in the past if
unemployed? (open question).
Hungary (2001): What is your source of livelihood? (choice of up to 3 of 13 response options); What is the
name of the main occupation and what activities characterize it? (open question).
Philippines (2000): What was [respondent]’s usual activity/occupation during the past 12 months? (open
question).
Rwanda (2002): During the month from July 15 to August 15, 2002, was [respondent] employed? (choice of
8 response options); What is (was) the main occupation of [respondent]? (open question).
Uganda (2002): What kind of work did [respondent] do in the last 7 days? (open question).
Viet Nam (1999): What was the main job that [respondent] performed during the last 12 months and what
position did [respondent] hold (if any)? (open question).
105
Romania, and the lowest densities in the four low- Further examination of the evidence points to horizon-
income countries included in the analysis (Cambodia, tal and vertical clustering within the health workforce. In
Rwanda, Uganda and Viet Nam). all countries where data are available, women form the
majority of nursing and midwifery personnel – in some
cases over 90% – but this is not necessarily the case
8.4 Empirical analysis for other occupational categories (Table 8.2). Among
health professionals (except nursing), women are con-
In this section selected findings are presented on gen- sistently underrepresented relative to their share in the
der dimensions of the health workforce drawing on total health workforce. Women are likewise underrepre-
census data for 13 countries. The section is divided sented among health management compared to their
into three parts, according to the three stages of the overall share in the health labour market, as observed in
working lifespan that form the framework for HRH anal- the two countries where the national occupational clas-
ysis introduced in Chapter 1. A gender analysis of the sification allows their identification, namely Argentina
active workforce is first presented, followed by exami- and Brazil.
nation of each of the stages that directly influence its
size and distribution, namely entry and exit. Gender imbalances can also be found for certain
specific occupations. As illustrated in Figure 8.1, the
8.4.1 Gender and the active physician workforce is mostly male in those countries
health workforce where data are available. On the other hand, the per-
sonal care workforce – one that is generally less skilled
Censuses with properly designed questions on labour
– tends to be numerically dominated by women (Figure
force activity allow identification of workers with a
health-related occupation. They also offer the advan-
tage of allowing disaggregation of all key variables by Figure 8.1 Sex distribution of the physician
sex. This offers the possibility of examining for occupa- workforce, according to census data for selected
tional segregation by sex: a framework for assessing countries (around 2000)
gender equity that can correspond to either vertical
clustering (differentials in the sex ratio according to 100%
relative job status) or horizontal clustering (sex differen-

tials according to specialization) (17). 34.8
Percentage of physicians
75%
63.8 62.7
In most of the 13 countries, women form the major- 75.1
ity of the health workforce – a pattern contrasting with 50%

that observed for the rest of the national labour force,
where men tend to be more numerous (Table 8.2). 65.2
25%
Exceptions to this trend are Cambodia, where women 37.3
36.2
comprise a minority (40%) of the health workforce but a 24.9
small majority (51%) of the rest of the labour force, and 0%
Rwanda, where the percentage of female workers is Brazil Romania Thailand Uganda
similar for both health occupations and all other occu-
Female Male
pations combined (some 55%).
Figure 8.2 Sex distribution of the personal care workforce, according to census data for selected
countries (around 2000)
100%
Percentage of personal care workers
8.1 10.6
13.4 21.0
19.7
32.7 32.4
40.1
75% 50.9
Female
50% Male
86.6 91.9 89.4
80.3 79.0
67.3 67.6
59.9
25% 49.1
0%
Argentina Cambodia Ecuador Philippines Romania South Thailand Uganda Viet Nam
Africa
106
8.2). A notable exception is Romania, the country with It has been speculated that, given the longer history
the highest proportion of women in the health work- of high female labour force participation in transitional
force, and where female physicians are more common. countries of eastern Europe, along with social policies
Table 8.2 Percentage distribution of the health workforce by sex, by occupation, according to census
data for 13 countries (around 2000)
Health workforce
Country Sex Health Nursing & Other health Health Total Rest of
professionals midwifery service management labour
(except personnel providers occupations force
nursing &
midwifery)
Male n.a. n.a. 32.9 35.5 33.0 63.6
Argentina
Female n.a. n.a. 67.1 64.5 67.0 36.4
Male 55.1 18.7 25.9 44.2 32.7 62.8
Brazil
Female 44.9 81.3 74.1 55.8 67.3 37.2
Male 74.6 45.4 60.1 n.a. 59.8 48.6
Cambodia
Female 25.4 54.6 39.9 n.a. 40.2 51.4
Male 43.6 n.a. 32.7 n.a. 38.8 66.0
Chile
Female 56.4 n.a. 67.3 n.a. 61.2 34.0
Male 57.7 8.5 18.2 n.a. 33.9 67.6
Ecuador
Female 42.3 91.5 81.8 n.a. 66.1 32.4
Male 43.3 n.a. 9.8 n.a. 21.3 55.8
Hungary
Female 56.7 n.a. 90.2 n.a. 78.7 44.2
Male 39.4 14.6 31.7 n.a. 27.6 50.2
Philippines
Female 60.6 85.4 68.3 n.a. 72.4 49.8
Male 31.9 5.6 13.4 n.a. 17.2 57.5
Romania
Female 68.1 94.4 86.6 n.a. 82.8 42.5
Male 61.9 35.4 50.5 n.a. 43.7 45.0
Rwanda
Female 38.1 64.6 49.5 n.a. 56.3 55.0
Male 62.8 8.2 27.4 n.a. 22.1 59.3
South Africa
Female 37.2 91.8 72.6 n.a. 77.9 40.7
Male 48.4 6.7 30.8 n.a. 25.1 52.0
Thailand
Female 51.6 93.3 69.2 n.a. 74.9 48.0
Male 68.8 13.0 54.9 n.a. 41.7 n.a.
Uganda
Female 31.2 87.0 45.1 n.a. 58.3 n.a.
Male 55.2 22.2 31.4 n.a. 34.3 51.7
Viet Nam
Female 44.8 77.8 68.6 n.a. 65.7 48.3
Note: Health professionals include physicians, pharmacists, dentists and other professional-level health occupations as reported at
the time of the census. Other health service providers include – depending on the country – health technicians and associate profes-
sionals, traditional medicine practitioners, personal care workers and others not identified elsewhere. Health management occupations
include administrators and supervisors in health services and similar occupations. In Argentina, all health service occupations (health
professionals, nursing and midwifery personnel) are grouped under “other health service providers”. In Chile and Hungary, nursing
and midwifery personnel are assimilated under either health professionals or other health service providers. In Uganda, health manag-
ers and planners are included under health professionals.
n.a. Not available (based on available data source or occupational classification).
107
emphasizing equality and supporting working women secondary and tertiary education is one of the core indi-
and their families, vertical gender imbalances may be cators of the Millennium Development Goals (19).
less pronounced than in other regions (18).
An illustrative analysis of gender imbalance in educa-
In addition to providing opportunities for monitoring tional attainment drawing on census data is presented
health occupations, censuses with data on place of in Table 8.3. In almost all countries, women health
work allow identification of those with other (non-health) workers outnumber their male counterparts in the
occupations working in the health services industry, or lower education categories, i.e. at primary level and
health systems support staff. As seen in the illustrative secondary level. Moreover, at these levels the imbal-
example from Uganda (Figure 8.3), gender imbalance tends to be more pronounced within the health
ances can also be seen among support staff. Women workforce compared to the general adult population.
are underrepresented compared to men in non-health It is especially pronounced in the two countries of the
technical specializations (for example accounts and eastern European region. The exception is Cambodia,
finance, engineering, information technology); moreo- where there are fewer less-educated female health
ver, within these fields they are less numerous at the workers than males; this may be related to the fact that
professional level compared to their counterparts at the Cambodia is one of the countries with the lowest over-
associate professional level. Conversely, women are all level of education (98% of the adult population with
overrepresented as clerks. at most primary schooling).
While a high female-to-male ratio among HRH can be

8.4.2 Gender and entry into
seen across all education levels in most countries – a
the health workforce trend reflective of the overall feminization of the health
It is possible that occupational clustering by sex is a workforce – the ratio is generally much lower at the ter-
reflection of differential access among women and tiary level of educational attainment than at the primary
men to education and training leading to a skilled pro- level. When it comes to the highest education cate-
fession. One of the strengths of using censuses for gory, the sex ratio among health workers tends to more
monitoring the gender dimension of entry to the health closely follow that of the general population.
workforce is the availability of nationally representative
data on the pool of eligible candidates for advanced To further the analysis, as seen in Figure 8.4, coun-
health education and training. Monitoring the pool of tries with greater gender inequities in access to higher
eligible candidates is an important HRH policy ques- education tend to be those with greater gender imbal-
tion, but one that is often overlooked in health workforce ances in the health professional workforce. Among the
research and planning (see Chapter 4). Monitoring 11 countries with comparable census data, a close and
gender differences in access to education is also of par- direct relationship is revealed between the sex ratio in
ticular importance: the ratio of girls to boys in primary,
Figure 8.3 Sex distribution of health systems Figure 8.4 Relationship between sex ratio in
support staff, by occupation, Uganda, 2002 census tertiary-level educational attainment and health
professional work activity, according to census
100% data for selected countries (around 2000)
21.2
Percentage of health support staff
75% 53.4 Romania

69.0 71.8 2.0
Ratio of women to men among health professionals
50%
Philippines 0.66
y = 1.16x
78.8
(except nursing & midwifery)
1.5 R2 = 0.54
Chile Hungary
25% 46.6
31.0 28.2 Thailand
1.0
Viet Nam
0% Brazil
Rwanda Ecuador
Non-health Non-health Clerks Other support
professionals technicians & staff South Africa
0.5
associate
professionals Cambodia
Female Male 0.0

0.0 0.5 1.0 1.5 2.0
Ratio of women to men with tertiary education
Source: Ssennono, Petit and Leadbeter (15). (population 15 years and over)
108
Table 8.3 Ratio of women to men by level of educational 8.4.3 Gender and exit from
attainment, health workforce and total population, according the health workforce
to census data for selected countries (around 2000)
Although a census is cross-sec-
Country Group Educational attainment tional, i.e. time specific, it is possible
to use a series of censuses to evalu-
At most Secondary Tertiary
ate exits from the health workforce. In
primary
Thailand, as in many other countries,
Health workers 4.5 3.0 1.2 censuses are carried out once every
Argentina
Total population 1.0 1.4 0.9 decade. A cohort analysis approach
was used across two successive
Health workers 2.2 3.4 1.3
Brazil censuses to measure net workforce
Total population 1.0 1.3 1.1 attrition by sex, that is, using age-
specific data for following the same
Cambodia cohort from one census to the next.
Total population 1.1 0.4 0.3 For example, the number of nurses
Health workers 1.1 2.4 1.1 aged 35–39 in 1990 was compared
Chile with the number aged 45–49 10 years
Total population 1.0 1.1 0.8
later to estimate net attrition (or net
Health workers 6.9 5.3 0.9 effect of workforce gains and losses)
Ecuador for this cohort. This entailed additional
challenges for ensuring data com-
Health workers 12.4 9.4 1.3 parability. In particular, the system
Hungary
Total population 1.3 0.9 1.0 of occupational coding was different
across the two census rounds: in the
Philippines 1990 census occupations had been
Total population 1.0 1.0 1.4 coded according to the 1968 version
Health workers 9.0 5.8 2.1 of ISCO, while for the 2000 census
Romania they were mapped to the 1988 ver-
Total population 1.3 0.8 0.9 sion. The analyses of Thai census
Health workers 1.3 1.7 0.3 data presented here were done using
Rwanda ISCO-88 as the reference.
Health workers 3.7 5.5 1.3 Table 8.4 shows the change over the
South Africa decade in workforce size for selected
cohorts for two cadres, namely physi-
Health workers 1.9 2.6 1.7 cians and nurses. As almost all newly
Thailand
Total population 1.1 0.9 1.1 educated health workers enter the
profession under the age of 35, and
Viet Nam assuming there is no appreciable
Total population 1.1 0.8 0.5 international in-migration, the number
Note: Data on educational attainment refer to the population aged 15 years
of workers aged 35 and over can be
or over. Secondary attainment includes post-secondary education at the non- used to calculate the rate of workforce
university level. Health workers are defined as persons active in the labour force attrition, or percentage reduction in
and reporting a health occupation at the time of the census.
total health professionals over the pre-
vious 10 years.
tertiary-level educational attainment among the general Three key trends can be ascertained from the cen-
population and the sex ratio among active health pro- sus data: increasing levels of attrition with age; higher
fessionals. Notably, the correlation coefficient, which attrition among male health professionals compared to
gauges the strength of a relationship between two var- their female counterparts; and, in the older age groups,
iables, is found to be relatively high by social science higher attrition among nurses than physicians. The cul-
standards (0.54 for the trend analysis including all 11 mination is the observation of the highest rate of attrition
countries, or 0.67 when excluding the outlier data point among older male nurses. The reasons for leaving the
for Romania). health workforce can include retirement, out-migra-
tion, career change, work-limiting illness, death or
other. Such findings underscore the importance for
109
Table 8.4 Estimates of intercensal health workforce attrition by sex, Thailand, 1990 and 2000 censuses
Males Females
Age group Number % loss Number % loss
in 1990 in 2000 in 1990 in 2000 in 1990 in 2000
Physicians
35–39 45–49 1872 1543 18 661 583 12
40–44 50–54 1368 1114 19 527 466 12
45–49 55–59 732 543 26 332 272 18
Nurses
35–39 45–49 694 596 14 7619 6755 11
40–44 50–54 562 407 28 5352 4541 15
45–49 55–59 459 270 41 3042 2217 27
HRH analysts, planners and policy-makers to consider third of all countries (17 out of 53) did not participate,
workers’ gender, equally with their occupation and age, resulting in nearly half of the continent’s population not
among the core variables in national workforce supply being enumerated. With intensified national, regional
and requirement planning and projection efforts. and international efforts and support, as of mid-2008,
censuses have been undertaken or are being planned
in 51 African countries for the 2010 round (20).
8.5 Summary and conclusions
One of the main constraints to census sources for HRH
This chapter has focused on uses of national popula- analysis (in addition to relatively lengthy periodicity)
tion census data for gender-specific HRH assessments has been limited access to and use of microdata for
as a basis for formulation of evidence-based policy health research, especially in many low- and middle-
options. Population censuses can be a useful source income countries. In order to optimize census uses,
of information for health workforce monitoring and eval- collaborations should be planned in advance between
uation. In particular, appropriate census data allow ministries of health, census bureaus and other stake-
calculation of a number of health workforce indicators holders for developing a strategy for data collection
and their disaggregation by sex, a critical requirement and processing and for dissemination and use. The lat-
for assessing gender equity in the workforce. ter should ideally include some combination of each of
the dissemination tools used here: (i) public access to
For a census to be useful for health workforce analy- microdata samples for scholarly research (anonymized
sis, the most important requirement is that it contains to protect confidentiality); (ii) limited access to full cen-
adequately defined and classified labour force varia- sus databanks (such as secured access for approved
bles that allow distinction of health occupations from researchers); and (iii) specialized health workforce pro-
other occupations. As previously noted in Chapter 2, files as part of the national series of census publications.
ideally, occupational data gathered in a census should In practical terms, the cost of collecting, processing
be processed to the lowest level of disaggregation as and tabulating nationally representative data on HRH
classified in ISCO (or its national equivalent). The case will be marginal for census exercises already including
studies presented here have shown how censuses can questions on labour force activity.
provide relevant information for many aspects of work-
force planning and policy development, notably from a In general, the analysis supported the notion that the
gender perspective. situation of human resources in health systems is often
a reflection, or even exacerbation, of broader societal
In order to be of greater value for research and pol- gender norms and inequalities, particularly with regard
icy, population censuses should be taken at regular to access to the education and training required to
intervals, at least every 10 years (10). Despite the rec- become a skilled health professional. Most HRH anal-
ognized importance of the census, many countries did yses are approached using occupational lens, with
not participate in the previous 2000 round; in Africa, a little explicit attention to gender dimensions within and
110
across occupation groups. Gender mainstreaming in 9. Reichenbach L. The overlooked dimension: gender
health workforce monitoring and evaluation strategies and the global health workforce. In: Reichenbach
L, ed. Exploring the gender dimensions of the
is needed to ensure that evidence-based gender-sen-
global health workforce. Cambridge, MA, Harvard
sitive approaches are undertaken to health workforce University, 2007.
planning and management, relevant to the needs
10. United Nations Statistics Division. Principles and
and interests of both women and men health workers.
recommendations for population and housing
Research, policy and programme efforts to address censuses, revision 2. Statistical Papers Series M, No.
gender equality in the health workforce should lead to 67/Rev. 2. New York, United Nations, 2008 (http://
strengthened health systems more broadly. Access to unstats.un.org/unsd/demographic/sources/census/
reliable, comprehensive information on gender and the docs/P&R_Rev2.pdf, accessed 11 January 2009).
health workforce – such as through sex-disaggregated 11. Dubois CA, McKee M. Cross-national comparisons
data from a population census – can inform the steps of human resources for health: what can we learn?
Health Economics, Policy and Law, 2006, 1:59–78.
needed to achieve gender equity.
12. McCaa R, Ruggles S. IPUMS International: a global
project to preserve machine-readable census
microdata and make them usable. In: Hall PK, McCaa
References R, Thorvaldsen G, eds. Handbook of international
historical microdata for population research.
1. Diallo K et al. Monitoring and evaluation of human
Minnesota, MN, Minnesota Population Center, 2000.
resources for health: an international perspective.
Human Resources for Health, 2003, 1:3 (http:// 13. Minnesota Population Center. Integrated Public Use
www.human-resources-health.com/content/1/1/3, Microdata Series (IPUMS)-International: version 3.0.
accessed 10 January 2009). Minneapolis, MN, University of Minnesota, 2007
(http://www.ipums.umn.edu, accessed 23 January
2. Gupta N et al. Uses of population census data for
2009).
monitoring geographical imbalance in the health
workforce: snapshots from three developing 14. About IHPP. International Health Policy Program-
countries. International Journal for Equity in Thailand (http://www.ihpp.thaigov.net/about.html,
Health, 2003, 2:11 (http://www.equityhealthj.com/ accessed 23 January 2009).
15. Ssennono V, Petit P, Leadbeter D. Uganda 2002
3. Global employment trends for women. Geneva, population and housing census: special analysis
International Labour Office, 2008 (http://www. on health workers commissioned by the Developing
ilo.org/wcmsp5/groups/public/ – -dgreports/ – Human Resources for Heath Project. Kampala,
-dcomm/documents/publication/wcms_091225.pdf, Uganda Bureau of Statistics and Developing Human
accessed 23 January 2009). Resources for Heath Project, 2005.
4. Gupta N et al. Assessing human resources for 16. International Standard Classification of Occupations:
health: what can be learned from labour force ISCO-88. Geneva, International Labour Office, 1988
surveys? Human Resources for Health, 2003, (http://www.ilo.org/public/english/bureau/stat/isco/
1:5 (http://www.human-resources-health.com/ isco88/index.htm, accessed 23 January 2009).
17. Gornick JC. Gender equality in the labour market:
5. Standing H. Gender: a missing dimension in human a comparison of the public and private sectors
resource policy and planning for health reforms. in six countries. Luxembourg Employment Study
Human Resources Development Journal, 2000, Working Paper Series No. 7. Syracuse, NY,
4:1 (http://www.who.int/hrh/en/HRDJ_4_1_04.pdf, Syracuse University, 1996 (http://www.lisproject.
accessed 23 January 2009). org/publications/leswps/leswp7.pdf, accessed 23
January 2009).
6. George A. Human resources for health: a gender
analysis. Background paper prepared for the WHO 18. Anker R. Gender and jobs: sex segregation of
Commission on Social Determinants of Health, 2007 occupations in the world. Geneva, International
(http://www.who.int/social_determinants/resources/ Labour Office, 1998.
human_resources_for_health_wgkn_2007.pdf,
19. Millennium Development Goals. United Nations
Development Programme (http://www.undp.org/
7. Sen G, Ostlin P, George A. Gender inequity in mdg, accessed 23 January 2009).
health: why it exists and how we can change
20. African Centre for Statistics. Status of
it. Report to the WHO Commission on Social
implementation of 2010 round of population and
Determinants of Health, 2007 (http://www.who.int/
housing censuses in Africa. African Statistical
social_determinants/resources/csdh_media/wgekn_
Newsletter, 2008, 2(2):7–9 (http://www.uneca.org/
final_report_07.pdf, accessed 23 January 2009).
statistics/statnews/ACSNewsletterJune2008.pdf,
8. Shengelia B et al. Access, utilization, quality, and accessed 23 January 2009).
effective coverage: an integrated conceptual
framework and measurement strategy. Social
Science and Medicine, 2005, 61(1):97–109.
111
112
Use of administrative data
9 sources for health workforce

analysis: multicountry experience
in implementation of human
resources information systems
PAMELA MCQUIDE, DYKKI SETTLE, WALID ABUBAKER, GHANIM MUSTAFA ALSHEIKH,
CÉLIA REGINA PIERANTONI, SHANNON TURLINGTON, DANNY DE VRIES
9.1 Introduction across various records, and issues of data quality, such
as incomplete records, timeliness and inconsistencies.
The most efficient and immediate way to track changes As a result, decision-makers may not have access, in a
to a health workforce is to use data from a routine timely and accurate fashion, to critical pieces of infor-
administrative information system. Censuses and sur- mation necessary to developing an effective human
veys, both population based and facility based, are key resources for health (HRH) strategy. For example:
tools to provide an accurate snapshot of a country’s t How many physicians and nurses are being trained
health workforce, but must be fully reconducted peri- and in what specialties?
odically to look at a changing situation. As these tools t How are health workers distributed across urban
are prohibitively expensive to implement on a regular and rural areas?
basis, data from such sources should be considered t Why are health professionals leaving the health-care
as a basis (albeit an essential one) for complement- services industry?
ing the national routine human resources information t How many health professionals are currently not
system (HRIS), which provides a continuous record employed?
of changes in the health workforce and serves as the
timeliest source of information available. These sys- Understanding the answers to these and other key pol-
tems are typically used by administrative organizations icy questions will help decision-makers effectively:
in the country, such as ministries of health, professional t ensure a steady supply of trained health professionals;
councils and professional associations, to qualify, man- t deploy human resources with the right skills to the
age and plan the health workforce. It is in the interest of right positions and places to meet health-care needs;
these organizations to maintain updated and accurate t retain health worker skills and experience in the
information in the HRIS. country.
An HRIS can be as simple as a filing cabinet of paper A mature, comprehensive HRIS links all human
personnel files or as complex as a multi-database sys- resources data from the time health profession-
tem with the capacity to analyse workforce problems als enter pre-service training to when they leave the
and assist in identifying possible solutions. The strength health workforce. Using such a system, decision-mak-
of an HRIS does not depend on technology but on its ers can quickly find the answers they need to assess
ability to be adapted to address current workforce HRH problems, plan effective interventions and evalu-
issues and generate accurate and timely information. ate those interventions. If well designed, managed and
In most low- or middle-income countries HRIS data maintained, HRIS data can provide a cost-effective yet
are routinely collected in paper form. While such a sys- extremely useful source of information with which to
tem can represent a functioning HRIS, there are often monitor and evaluate the impacts of changes in social
serious limitations to the use of these data. Although policy at the national and subnational levels.
records representing a single role or individual can be
located and accessed (with varying degrees of ease), Country experience indicates that, while often less
often these records cannot be used efficiently because accurate than census or survey data, particularly in the
of the intensive effort involved in updating or aggregat- initial phases, routine health data systems improve in
ing data, difficulties linking data on an individual level accuracy over time (1). A well-functioning HRIS has an
113
advantage over a survey because it allows for ongo- contributes to the success of health workers and health
ing monitoring of detailed information in large sample workers influence the functionality of the system, there
sizes, including subnational geographical analyses. is also a critical interdependence of the HRIS and the
In addition, it is longitudinal in nature with the capa- other components of the health information system. In a
bility to record health workforce dynamics. Continuous mature and comprehensive health information system,
and current information on the same individual can be the HRIS will be the authoritative source of HRH data
tracked over long periods of time at less cost per data for each of the other building blocks. This underscores
point. Due to advances in computer technology, linking the critical nature of the HRIS to the integrity and suc-
various administrative databases is easier, less expen- cess of the whole health system.
sive and more reliable than ever before (2).
Ideally, the HRIS development and strengthening pro-
At the same time, many low-income countries need to cess comprises five key stages using a participatory
build capacity, both human and technical, in order to approach (Figure 9.1). First, a stakeholder leadership
improve the use of HRIS data, including planning, soft- group is established, or a national task force involving
ware design, infrastructure support and management, all key stakeholders and led by the ministry of health,
as well as strategies for data use and strengthening. and the key HRH policy questions are identified. Next,
The implementation of a comprehensive country strat- HRIS technical staff conduct a needs assessment to
egy should include the following underlying principles: focus on the current infrastructure (for example exist-
t a participatory approach that involves stakeholders ing networks, Internet connectivity, software) and data
from various ministries and sectors from the outset already being collected by the different ministries,
and increases the ownership of the system; councils and other organizations. After the questions
t an iterative development methodology that incor- and infrastructure are agreed upon by the stakeholder
porates existing systems, tools and processes as group, HRIS software solutions should be custom-
much as possible to lower costs and speed up ized to answer the key health workforce policy and
implementation; management questions for a respective country. The
t a mature software solution designed for the country end result could either be a step solution or a mature
context and to answer key HRH policy questions for (multiple component) system able to support manag-
that country; ers and decision-makers in their efforts to effectively
t an emphasis on building capacity, ensuring sus- use and analyse data for informed and confident deci-
tainability and continuously improving the system sion-making. At the stage of data use, attention is paid
through training and technical support; to how data are actually used for decision-making.
t a continuous effort to train decision-makers to ana- Throughout the process, sustainability and continu-
lyse and use the data that the HRIS provides to ous improvement of the HRIS can be ensured through
make sound HRH decisions. training and building of capacity in the country team
to independently support, use and improve the system
In this chapter, an overview is provided of the essential into the future.
elements and lessons learnt to date in the implementa-
tion of a comprehensive HRIS strengthening initiative.
9.2.1 Building a stakeholder
A series of critical stages in the development and
strengthening of a complete HRIS is first discussed,
leadership group
followed by presentation of a number of case studies A key to the success of implementing an HRIS is the
drawing on experiences in selected low- and middle- ability to respond to in-country needs. Developing a
income countries. leadership group of all the essential stakeholders that
produce and use HRH information assists in develop-
ing a country-focused HRIS. This group will initiate, lead
and monitor all subsequent activities in HRIS strength-
9.2 Recommended first steps
ening. The stakeholder leadership group empowers
to develop a human resources stakeholders to develop an HRIS that meets their
information system needs, ensures ownership of the system and builds the
necessary capacity to support, use and improve the
The first and most important thing to keep in mind when
HRIS. Another benefit of including the key stakehold-
developing an HRIS is that just as human resources are
ers is that it opens communication channels between
a building block of a health system (3), so the HRIS is
groups of individuals that typically do not meet together,
a part of the comprehensive health information system.
thereby facilitating collaboration and sharing of data
And just as every aspect of the health system (including
across groups.
facility- and community-based service delivery, availa-
bility of medical products, financing and governance)
114
Use of administrative data sources for health workforce analysis
Figure 9.1 Framework for institutionalizing a human resources information system
Intermediate
Establish
step solution
stakeholder
group
Define key HRH
policy questions
Data for:
Assess and t QMBOOJOH Result
strengthen IT HRIS approach?
t NBOBHFNFOU
infrastructure
t QPMJDZ
Provide Mature, comprehensive

data-driven solution
decision-making
training to
stakeholders
Ongoing stakeholder involvement, training & ownership
Source: Capacity Project (1).
The stakeholder leadership group should include together to develop a mature and complete HRIS that
experts in health workforce planning and produc- tracks health professionals from the time they enter
tion, and in information systems. The outcome of the training until they leave the health workforce. Depending
first meeting should be the terms of reference for the upon the scope of the HRIS development, there may
group (such as its leadership, membership, account- be a need for employees of facility-based service
ability, mission, function and duties) and the principles delivery points to be included in the stakeholder lead-
of operation for the meetings (such as the equality of ership group to ensure that individual-level data remain
all members, the need to hear from all participants and accurate and that they have access to necessary infor-
the need to reach consensus for decisions to occur). mation (for example, direct salary deposits into their
bank account). At every juncture, the most important
After deciding on the terms of reference and princi- outcome is that ownership of the HRIS is being built
ples of operations, the group is ready to begin defining and capacity is being developed.
the key HRH questions that need to be answered and
the indicators that will be used to monitor the status
9.2.2 Infrastructure development
of the health workforce via the HRIS, and considering
issues pertaining to data confidentiality, ownership and Despite the momentum that is often generated by the
policies for data sharing. It is critical that the country stakeholder leadership group, experience suggests
identifies and owns these questions to ensure the suc- that no country is ever ready to deploy a complete and
cess and usability of their HRIS. Once the stakeholder mature HRIS in the first instance. Strengthening the
group identifies these questions, the group will continue information system in planned steps provides quick
to meet regularly to provide direction for the infrastruc- gains without overwhelming the infrastructure needed
ture development, programming, data inputting, report to support the system. Data collectors and managers,
development and use of data for decision-making. It is technical support staff and decision-makers should
particularly important that issues of data confidential- receive training at each step, become comfortable with
ity and data ownership are addressed from the onset the new system and then take the next step when they
since these are not typical considerations with a paper- are ready. In particular, training should occur in updat-
based system and could have serious consequences if ing new information at each stage of the workforce life
safeguards are not considered and instituted from the span (for example pre-service training, new deploy-
very beginning. ment, redeployment, migration, retirement, death).
Many stakeholders are involved. Ministries, licensing Improvements to existing information technology infra-
and registration or certification bodies, private sec- structure can generally be implemented quickly and
tor organizations and other stakeholders must work often result in increased efficiency and productivity.
115
Proposed infrastructure improvements should be t Qualification. This component is used to collect

based on a technical assessment and consider low- and aggregate data on skilled health service provid-
cost solutions that can rapidly but significantly enhance ers and speeds up the process of generating routine
existing systems and processes. Recommended forms, such as registrations and licences or records
improvements might include procuring computers for of professional examination results, that were previ-
workers who need to enter or access data in the sys- ously handwritten. It should reside with the licensing
tem, improving software applications and tools that are or certification authority for a health worker cadre,
currently being used, increasing data storage capacity, such as physicians or nurses. In most countries
upgrading network connections for transmitting data or these authorities are boards or councils. These data
expanding technical support services. are critical for hiring authorities throughout the coun-
try to ensure that only qualified professional health
workers are hired, since forging paper documents is
9.2.3 HRIS development steps
reported in many countries.
Depending on the current form, existing HRH informa- t Management. This component tracks detailed
tion systems may be strengthened in iterative steps. A information about health workers who are employed
step solution is any interim solution for managing HRH by the ministry of health, a public sector health-care
data that is deployed while a mature system is being facility such as a hospital, or a private sector serv-
developed. Step solutions can be deployed to ena- ice delivery point within the country. In addition to
ble the ministry or another organization to quickly start individual deployment information, other pertinent
entering and managing its HRH data. The data can information may include performance appraisal, dis-
then easily be migrated to the mature solution when it ciplinary action, retirement, change of occupation
is available and the end user is ready to implement it. and payroll information.
t Education and training. There are two main train-
Where there is only a paper-based system, an electronic
ing components: (i) pre-service education, which
register can be implemented. Where an electronic reg-
tracks the level and field of education leading to
ister is already in place, a simple database can be built.
qualification for a health occupation; and (ii) continu-
Where there is a simple database, that database can
ing education and in-service training, which update
be progressively strengthened or expanded. For exam-
the professional knowledge, practice and skills of
ple, a situation may arise where the routine database
individual health service providers. Continuing edu-
system is so out of date and unreliable that policy-mak-
cation may be required to maintain an active licence
ers cannot manage the results. One particularly useful
to practise in certain countries. These two com-
step solution is to develop a short survey form to collect
ponents of training may be included in either the
minimal accurate information on health workers at the
qualification or management systems or they may
facility level and enter these data into a register. These
be stand-alone systems, depending on the needs
data can be used to update databases and make the
of a country.
system operational again. Another important step solu-
t Planning. This component uses data from each of
tion could be the specification of a gender-sensitive
the other systems and statistical modelling appli-
HRIS.
cations to form a complete picture of the health
workforce in the country and allow projections on
The goal of such an iterative strengthening methodol-
how that workforce will change in the future (based
ogy is to ensure that every country or organizational
on known influences such as retirement age,
programme can quickly benefit from an HRIS strength-
number of trained health workers annually entering
ening process regardless of resources. Proceeding
the workforce, attrition rates, changes in population,
in iterative steps also lessens the impact of too much
staffing norms, disease patterns and other factors).
change too quickly, while ensuring that each step
The planning and modelling component of the HRIS
progresses towards the goal of a mature and complete
can help decision-makers assess their health work-
solution.
force needs and make effective policy decisions to
meet those needs.
When a country or organization is ready and the appro-
priate infrastructure and supporting systems are in
Together, these four components can provide a power-
place, a complete set of mature software solutions can
ful feedback loop for analysing, planning and managing
be implemented to fully track skilled health service pro-
health workforce resources and needs. If all four com-
viders from the time they enter training until they exit
ponents use the same core programming, database
the health workforce. Four components are needed to
architecture and supporting hardware and software
address the most critical health workforce planning,
systems, once one is fully deployed, the others can
production and management issues:
be added at a significantly lower cost. Each of the four
116
systems may be deployed independently or integrated quality and integrity. Data quality is of primary impor-
with software products already in place to provide a tance and should be emphasized at every step of the
customized, contextual solution for the country, filling in process, from initial data collection to data analysis and
any gaps that existing systems may have left (1). interpretation.
Data accuracy and completeness are necessary to

9.2.4 Supporting use of data for
inform decision-making. Standardization of data col-
effective decision-making lection forms and data coding can facilitate ease of use
The primary aim of any HRIS should be to promote bet- and internal validity. In addition, procedures to minimize
ter use of data to drive effective decision-making for data entry errors when transferring data from paper to
addressing daily challenges and positively impacting electronic forms can result in improved data quality.
HRH policies and practice. However, it would be very Dual data entry, in which a record is entered into an
difficult for a few people poring over data reports to electronic database at two separate times by two sepa-
make sound and binding decisions. The ways in which rate staff, can reduce data entry errors by ensuring that
data are used for effective decision-making rely on the discrepancies in data are compared against the origi-
active involvement of a broad range of stakeholders nal document. Should dual data entry be prohibitively
working together. Understanding the context in which expensive or time intensive, a system of spot-checking,
data are used to make decisions is also essential. in which a randomly selected list of electronic records
is checked against the original record, can be used.
Rather than having an external consultant supply a Both dual data entry and spot-checking permit data
packaged training programme for using data for deci- managers to identify, record and correct data errors.
sion-making, for instance, all national stakeholders A data error log can also serve as a starting-point for
should be engaged in understanding how they use later improvements in training methods, data collection
data now, both individually and organizationally, and forms or software modifications.
what factors are important in their context for using
data effectively. Examples of approaches that have Routine, reliable updates can also improve data qual-
been employed in countries to improve the use of data ity by ensuring that data remain relevant for planning
for decision-making include: and decision-making. Data reports at the central level
t mapping how data are used to support a decision, can be sent (either electronically or via paper forms)
resulting in case studies of practical data use; to provincial or district representatives for review and
t providing opportunities for decision-makers to expe- updating on a regular schedule. Similarly, processes
rience critical decision-making moments so they should be put in place to facilitate the flow of data and
can develop their skills using real data in real-life reports between the central and district levels, and
situations; between hospitals, health centres and other service
t improving communication among users of data; delivery points. This exchange of information allows the
t identifying and leveraging opportunities for improved HRIS to better account for changes in the workforce,
data sharing across different levels of the organiza- such as new deployments, transfers, specific in-serv-
tion and with other stakeholders. ice trainings and workforce exits. Sharing data not only
improves accuracy but also enables health planners at
9.2.5 Methodology for sustainability all levels to gain access to information valuable for pol-
and continuous improvement icy and administrative decisions. Furthermore, enabling
public access to aggregated HRIS data may facilitate
An information system requires ongoing support and
HRH planning and research across sectors, including
improvement to ensure maximum utility and sustain-
nongovernmental organizations, academic research-
ability. Depending on the context and needs of the
ers and policy-makers.
country, sustainability strategies include continuous
collection of feedback from stakeholders about chang- Although sharing data is essential in order to improve
ing data needs and rolling out of improvements that data quality and encourage evidence-based decision-
align with those needs. The early involvement of stake- making, maintaining data security is equally critical.
holders with the design and eventual implementation of HRIS data includes personal information that must
the HRIS encourages their sense of ownership. As part remain secure in order to build confidence and trust
of this, the HRIS strengthening process should include in the system. Implementing and adhering to a data
the training of decision-makers and stakeholders to security policy starting from the initial stages of devel-
effectively use and analyse data for informed and con- opment of the information system can help ensure
fident decision-making. In addition, the training of data data confidentiality and integrity (Box 9.1) (4). A sys-
collectors, system support staff and system manag- tem can be built with several levels of access based on
ers helps to improve the technology infrastructure, data
117
Box 9.1 Confidentiality and security of HRH information
With increasing attention to human resources opportunities and constraints to achieving health systems
objectives, greater emphasis is being placed on the collection of information to improve HRH development
and monitoring. Having longitudinal data, or information gathered at different points over time, allows
individual health workers to be tracked in their labour market activities and other significant events,
supporting evidence-based decision-making for policies and programmes at critical junctures along the
working lifespan.
Provider-level information becomes even more important when used for human resources development
strategies or health services delivery monitoring. This will require information systems, whether paper
based or electronic, that ensure health worker confidentiality yet allow relatively easy access to the
information at both the individual and aggregate level. Implemented systems must also address issues of
system availability.
Using personal data for health systems goals must be balanced against individuals’ rights to privacy and
confidentiality, and should be based on human rights principles.
When developing approaches to protecting data, a distinction should be made between providing for the
physical protection of data to guard against environmental threats, and the protection needed to guard
against inappropriate use of sensitive information, whether due to inadvertent or deliberate activities.
Three interrelated concepts, each implemented in a different manner, have an impact on the development
and implementation of protection of sensitive data: privacy, confidentiality and security.
Privacy is both a legal and an ethical concept. The legal concept refers to the legal protection that has
been accorded to an individual to control both access to and use of personal information and provides the
overall framework within which both confidentiality and security are implemented.
Confidentiality relates to the right of individuals to protection of their data during storage, transfer and
use, in order to prevent unauthorized disclosure of that information to third parties. Development of
confidentiality policies and procedures should include discussion of the appropriate use and dissemination
of health worker data with systematic consideration of ethical and legal issues as defined by privacy laws
and regulations.
Security is a collection of technical approaches that address issues covering physical, electronic and
procedural aspects of protecting information collected as part of the HRIS. It must address both protection
of data from inadvertent or malicious inappropriate disclosure, and non-availability of data due to system
failure and user errors.
Source: Adapted from UNAIDS (4).
user roles. For example, such a system could enable systems. This part of the strategy could include the
some users to enter records without accessing reports, development of regional user communities to facilitate
other users to see reports without the ability to enter problem solving and share system improvements, pos-
or modify data, and a third group of users to access sibly in collaboration with local educational institutions.
only aggregate reports, without the ability to view or
edit individual records. All HRIS users should be sup- Taking full advantage of HRIS results to improve HRH
plied with a password that is regularly changed, and planning and management in developing countries
a system should be put in place to back up data on a requires a concerted supportive process. When pol-
regular schedule. icy-makers and other key stakeholders obtain access
to extensive information about the health workforce,
The development of skills is necessary for supporting, it may be difficult for them to see uses for these data
maintaining and developing computerized information beyond the usual reports that they formerly generated
118
with paper records. Additionally, the HRIS data are strengthening step could focus on generating regular
being produced in a policy-making context that is reports to answer those questions.
highly political, and appointments to positions may
change frequently (5). The data for the decision- Improvements were made to the network and hardware
making process involves the key stakeholders in a infrastructure at the Uganda Nurses and Midwives
practical, participatory procedure of using, interpreting Council, the regulatory body that licenses profes-
and applying the new HRIS information while consider- sional nurses and midwives working in the country. An
ing what capacity needs to be developed to implement open source software application was installed: iHRIS
changes on an individual, organizational and policy Qualify, a training, certification and licensure-tracking
level. This may involve sharing information and reports database (7). “Open source” refers to computer soft-
that are now available and training on different ways to ware distributed under a licence that allows anyone
interpret the information and present evidence effec- to study, copy and modify the source code (the set of
tively to respond to key policy questions. Having the instructions that creates a piece of software) and redis-
key stakeholders present their own data to their peers tribute the software in modified or unmodified form,
helps to build the ownership and confidence in using without restriction and without the need to pay a licens-
these data to inform management and policy deci- ing fee. This means that products can be distributed
sions. The final phase involves developing a plan for at minimal cost, and users can continue to use and
disseminating these HRIS data regularly, based upon improve their systems without paying onerous licens-
the reporting cycle within the country. ing or upgrade fees. Open source software does have
some disadvantages, most importantly the potential
for poor support for users in countries with a shortage
9.3 Country case studies of skilled personnel in new information and commu-
nications technologies, and a lack of accountability
Examples are now provided from HRIS development, if software glitches or unauthorized access occurs.
strengthening and evaluation efforts in three low- and However, using open source software has the advan-
middle-income countries: Uganda, Sudan and Brazil. tage of encouraging software development in context,
and often represents the least-cost alternative where
9.3.1 Building a health professional there is a foundation of computer skills in the coun-
licensure information system in Uganda try, or better yet, the health system. It is also possible
to draw on the global open source support commu-
In Uganda, the Ministry of Health (6) and four health
nity that has developed around these technologies to
professional regulatory councils, including the Nurses
voluntarily support and improve the systems. This can
and Midwives Council, needed updated and reliable
be quite advantageous in helping users answer ques-
information on how many health professionals by cadre
tions, fix problems with software and even develop new
were licensed to work in the country, what training they
modules. The result is a completely tailored (but still
had received, if they were leaving the health workforce
low-cost) system that can grow and change over time.
and if so, why. Until recently, although a complex sys-
tem of paper forms was in place, there was no way to
In the Ugandan context, implementation teams were
aggregate or analyse the information, and it was diffi-
formed to programme and deploy the software. Entry
cult even to track down a given nurse’s current address
of historical registry data from the Uganda Nurses and
or licensing information.
Midwives Council was completed first, followed by the
other three licensing bodies – for medical and dental
A stakeholder leadership group was formed, includ-
practitioners, pharmaceutical practitioners and allied
ing representatives from several departments in the
health professionals. Quality controls were incorporated
Ministry of Health, the four professional licensing asso-
into the data entry and processing procedures, nota-
ciations, training institutions and nongovernmental
bly the assignment of a unique identification number
organizations, as well as consultants in health workforce
within and across cadres to address potential biases
planning and information systems from the Capacity
such as double-counting (for example, when a health
Project (1), a global HRH initiative funded by the United
professional has more than one type of training), and a
States Agency for International Development. The goal
dual data entry system to ensure accuracy and permit
was to develop a registration and licensing informa-
tracking of data entry problems. Another method used
tion system that would track all health professionals
for validating the data was to request each individual
from the time they entered pre-service training until
health worker, upon entry to the appropriate council, to
they left the health workforce. One of the first activi-
verify the contents and update any information that may
ties of the stakeholder group was to identify the key
have changed based on a printout of their electronic
policy questions that it wanted addressed regarding
record. This process keeps the database updated and
nursing and midwifery personnel, so that the first HRIS
119
gives the health professionals an appreciation of the Figure 9.2 Number of students entering nursing
need for data accuracy. training programmes (leading to qualification
as registered or enrolled nurses) in Uganda,
1980–2004
The next three figures offer examples of the type of
information that can be examined with the database,
drawing on administrative records for all student nurses 5000
and midwives that entered training between 1980 and
Number of student nurses

4000
2004. Figure 9.2 shows the increase in enrolment in
nursing and midwifery training programmes during this 3000
4425
timeframe. 2000
3579
2463
1000 2089
Figure 9.3 presents findings on completion of training
835
programmes and professional qualification among stu- 0
1980-1984 1985-1989 1990-1994 1995-1999 2000-2004
dent nurses and midwives in Uganda. Of the 21 888
student nurses and midwives that entered training from Source: Capacity Project (1).
1980 to 2004, only 17 297 completed the training pro-
gramme and sat for the exam. Of those who passed
the examination, 16 658 qualified to register with the
Nurses and Midwives Council and 14 637 eventually Figure 9.3 Number of student nurses who
registered – approximately two thirds of those who entered training between 1980 and 2004, passed
the professional licensing exam, and qualified,
originally entered training. One way in which these registered and became licensed with the Uganda
data are being used is to ensure that nurses and mid- Nurses and Midwives Council
wives hired are registered with the Uganda Nurses and
Midwives Council by giving central and district author- 25 000
ities restricted access to view the HRIS to see if the 20 000
applicants for professional positions are in good stand- 15 000
ing with the Council. 10 000 21 888

17 297 16 658
14 637
5000
7022
Figure 9.4 identifies the location of nursing schools by 0
district, showing where students are most successful in ntered tr n n t for t e e m u f ed to

re ster
e stered t
re u tor bod
ensed t
re u tor bod
passing the licensing examination and becoming reg-
istered, and where they are having more difficulty. For Source: Capacity Project (1).
example, 18% of nurses and midwives who attended
school in Kampala District and 11% of those who
attended school in Kisoro District passed the exam
but did not become registered by the Council. In con- important to observed positive outcomes from which
trast, all nurses and midwives who attended school stakeholders can learn.
and passed the exam in Bushenyi District became
Based on the results discussed above, in Uganda, a
registered.
recommendation was developed to further explore the
These data have important policy implications about causes for the high failure rates among student nurses
expenditure of scarce national resources for educating and midwives and propose ways to tackle the problem
health professionals that are wasted if individuals do (8). An interministerial stakeholder meeting was set up
not complete training and become registered with the to address some of the issues about examining nurses
appropriate regulatory body. They also raise questions for registration with the Nurses and Midwives Council.
about selection of students and the examination pro-
cess. Without knowing how many of these students do 9.3.2 Institutionalizing a
not complete training, policy-makers and planners do health workforce monitoring
not know how to budget or plan for training or deploy- framework in Sudan
ment of health workers to meet the needs in the country.
A number of strategies and initiatives have been imple-
Also, these data help educators, professional councils
mented in Sudan in recent years to improve health
and various ministries to identify potential problems in
system performance, including strategies to achieve
the early stages, prioritize where to target interventions,
the Millennium Development Goals and other inter-
offer an indication on where further investigation and
national, regional and national goals for reducing
monitoring are needed, and point to factors that were
poverty and improving population health. Accordingly,
Sudan’s Federal Ministry of Health made a decision
120
Figure 9.4 Number and percentage of student nurses who passed the professional licensing exam and
registered with the Uganda Nurses and Midwives Council, by school district (entrants between 1980
and 2001)
ssed e m not re stered ssed e m nd re stered

100
1178 56 653 1074 1077 6962 467 959 305 400 935 862 14 368 423 470 306
50
25
310 1281 188 109 95 68

115 107 157 78 82 35 32 162 1 39
0
uu e se ro r to r r t
ru en n b m mu se so um s r r oro de un oro
us b u
Source: Capacity Project (1).
to closely monitor the institutional performance of the pertaining to HRH variables, including those related to
key departments and programmes at all levels (9, 10). entry, active labour force and exit. The objective was
In collaboration with the World Health Organization, the development of a sustainable information and mon-
the Ministry developed a comprehensive monitoring itoring system for the timely and continuous updating,
and evaluation matrix tailored to the national context verification and analysis of data on health workforce
for health system performance. The matrix drew on dynamics. The plan drew on the national HRH 10-year
baseline information and set periodic targets for each strategy and other relevant initiatives. The monitoring
agreed upon indicator, against which the information and evaluation plan was thoroughly reviewed, dis-
system would be used to monitor progress over time. cussed and endorsed by a wide range of staff from the
Three main sections were included: Department of Human Resources Development and by
t global indicators, including those related to health the Ministry of Health undersecretary.
system coverage and health workforce density;
t disease-specific indicators, such as those pertain- Special HRH monitoring and evaluation reports were
ing to HIV, malaria and tuberculosis; presented at the monthly departmental performance
t department- or programme-specific indicators. review meetings, including progress made and trans-
parent identification of areas needing improvement.
In order to ensure national ownership, the Ministry of While the reports did not tackle such global issues as
Health assumed the leadership role, particularly during retention and migration, they did monitor recruitment,
the launching and institutionalization of the informa- deployment, transfer, exits, opportunities for pre-
tion system. In addition to international recognized service and in-service training and compliance with
expertise, a number of other partners and community performance-based incentive schemes. Once cleared
representatives were asked to participate through- after the departmental review, wallcharts were dis-
out the development and evaluation process. While played in each unit within the Department with monthly
the undersecretary of the Ministry of Health served as statistical variations. Inputs of monitoring and evalua-
the champion of the whole process, the Department of tion activities related to HRH were further used for the
Health Planning was commissioned to establish a new annual Ministry-wide statistical report.
institutional monitoring and evaluation unit with full-time
professional and support staff. A detailed profile was A number of practical steps to HRH information and
drafted for the unit, including operational terms of refer- monitoring system development and institutionalization
ence for both central and subnational teams. were part of the Sudan experience, including:
t critical mass strategy: advocacy to raise awareness
Each main department and programme selected focal of the crucial value of the HRIS and obtain political
points to assume responsibility of the implementation commitment;
of their respective monitoring and evaluation plans in t institutional ownership and leadership: multidisci-
accordance with the appropriate indicators, baselines plinary team approaches under the guidance of a
and targets. In particular, the Department of Human national leader;
Resources Development put in place a specific annual
monitoring and evaluation plan to monitor key dynamics
121
t consensus on a limited set of core indicators 9.3.3 Evaluating an HRH information

adapted to the country context, including explicit and management system in Brazil
baselines and benchmarks;
In Brazil, the HRH information and management sys-
t going electronic: deployment of user-friendly soft-
tem (sistema de informação e gestão de recursos
ware for monitoring progress;
humanos em saúde, or SIGRHS) was conceived in the
t methodology refinement and local adaptation: allow-
context of health systems reform. It was designed as an
ing flexibility for refining and adapting or readapting
instrument for the collection, processing and use of pri-
the information system to the changing local context
mary data relevant to planning and management of the
for the health workforce;
health workforce. The system was developed to inform
t training in data analysis, dissemination and use, and
regional and local health managers to make better
in basic information technology skills for national
decisions related to their personnel. Initiated in 1987 by
human resources development staff and monitoring
the Department of Health of the State of Rio de Janeiro
and evaluation coordinators and focal points;
and implemented by the University of the State of Rio
t transparent dissemination of information generated
de Janeiro’s Institute of Social Medicine, the HRIS has
by the system, including periodic dissemination ses-
been continuously updated, modified and evaluated
sions among key stakeholders aimed at gaining
in response to local demands and taking advantage
inputs and generating maximum consensus building.
of innovations in information technologies. In partic-
ular, the development of new, user-friendly software
Selected results from the HRH information and monitor-
applications facilitated data entry operations at the
ing system in Sudan are presented in Table 9.1. Critical
decentralized level and allowed inclusion of both quan-
to these results is the linking of the monitoring and eval-
titative and qualitative variables on HRH management.
uation analysis to problem solving and HRH strategic
enhancement: there is no point in investing in a com-
The HRIS includes data drawing on administrative
plex information system if the results are only used
records from different levels and types of institutions.
to monitor implementation. The system must include
Data collection and entry are conducted using stand-
mechanisms to identify and rectify areas that need
ardized templates. The basic elements captured in the
improvement.
system include sociodemographic characteristics of
health workers, professional qualifications and work-
related variables, such as job position, employment
sector and working hours. These items of information
enable profiling of the health labour force, for example
Table 9.1 Selected indicators and benchmarks from the health workforce information and monitoring
system in Sudan, 2006–2007
Indicator Period Target Result Gap Compliance Suggested rectification measures

rate
1. Number of Q1/2007 750 750 0 100% The deployment system has been
newly graduated substantially improved since utilization
physicians entering of the newly developed electronic
the service for system/web application (2006), with a
internship on time maximum waiting time of three months.
2. Number Q1/2007 10 3 7 30% Compliance rate is too low, to be
of internship discussed in next Department of
supervisory visits Human Resources Development staff/
at three identified performance review meetings.
hospitals using the
quality assurance Effort should be given to report
approved checklist qualitative results and not merely
quantitative.
3. Number of Q1–Q4, 12 7 5 58% Revisit newly developed incentive
medical specialists 2006 system and recommend modifications
deployed to to attract more specialists to work
underserved areas outside the capital.
Source: Federal Ministry of Health of the Republic of the Sudan.
122
the distribution by age, sex and educational attainment. The process of health systems reform and decentraliza-
They can be used to assess, among other things, skills tion, and the availability of new information technologies
mix of the health workforce and the deployment of staff at the level of local health organizations, were a cata-
across different types of professional functions, health lyst to the implementation of the HRIS in Brazil. In order
facilities and regions. Regular updating of the data also to evaluate the success of maintenance and consoli-
allows for estimating worker attrition rates. The main dation of the system, and the use of information as a
data source is staffing rosters and payroll for public planning and managerial tool, a survey was conducted
health services. Given the confidential nature of these in collaboration with the University of the State of Rio de
data, a technical team oversees the level of access, Janeiro for appraising the implementation of the HRIS
whether partial or total, to the information contained in and its performance. An appraisal tool was developed
the system for research purposes. drawing on a framework for evaluation of the imple-
mentation process covering three key components
The use of administrative data offers many advan- of accountability of health policies and programmes:
tages over other types of standard statistical sources. administrative, political and community levels (13, 14).
Traditionally, information on HRH in Brazil came from The survey was administered among various stake-
large-scale databases, compiled by different national holders across the country by means of site visits and
agencies for a range of objectives. They include the semi-structured interviews. The specific objectives of
decennial population census, household sample the survey included:
surveys including the monthly labour force survey, t identification of the different types and degrees of
and health facility surveys carried out by the Central involvement of relevant actors in the implementation
Statistical Office (11). Other sources of information on and utilization of the HRIS;
the formal labour market include various registries on t identification of the opportunities and constraints for
jobs, wages, recruitments and dismissals of the Ministry integrating additional information in the area of HRH
of Labour, and on specific occupations through the reg- management, notably qualitative information;
istration systems of professional associations, unions t assessment of the level of satisfaction among health
and councils (12). However, these sources tend to be service managers regarding the utilization of the
fragmented and are often not readily usable for inform- information contained in the HRIS;
ing managerial practices and strategies at the local t profile of the utilization of the information generated
level, particularly in the health sector. by the HRIS among other organizations both within
and outside the health sector.
Table 9.2 Selected indicators and corresponding criteria used for evaluating the performance of the
HRH information and management system in Brazil
Indicator Response options for performance level

High Average Low
Coverage of implementation At least 90% of registered Between 70 and 90% of Less than 70% of registered
of the HRIS among registered servers registered servers servers
servers in the public health
services network
Completeness of information Collection of data from Secondary data sources Incomplete data collection/
sources compiled in the HRIS primary sources and primarily used with at least compilation
complete fields one review of primary
sources
Productivity of implementation Emission of first Emission of first Emission of first
of the HRIS dissemination report less dissemination report within dissemination report more
than three months after data three to six months after than six months after data
collection data collection collection or not at all
Regularity of implementation Data collection routinely Data collection updated Data collection not updated
of the HRIS updated occasionally or on demand
Regularity of utilization of the Dissemination reports Dissemination reports not Dissemination reports not
HRIS issued regularly, up to four issued regularly, but at least issued
in last four months or on once in last six months
demand
Source: Pierantoni (15).
123
The performance of the HRIS was evaluated using a utilization of information on the health workforce across
number of criteria, including coverage, completeness, different situations. The need for comprehensive infor-
quality, timeliness and regularity of utilization (Table mation for the management processes of HRH and
9.2) (15). Forty-five per cent of the surveyed sites were health systems outweighs conditions of even the lowest
ranked as having a high level of performance of the performance of the implementation of the information
HRIS, 33% average performance and the remaining system itself.
22% insufficient. While an evaluation of the technical
characteristics and operation of the information sys- More recently, as requested by the Brazilian Ministry of
tem’s computer software programme was not explicitly Health (16), the information system has been updated to
outlined, the results of the survey did gather informa- incorporate more detailed information on education and
tion on usability of the instrument that guided future training for health workers, including numbers of gradu-
upgrades. ates of advanced health education programmes. Both
a new software (sistema de informação sobre a grad-
Table 9.3 presents selected findings on the institutional uação em saúde, or SIGRAS) and the latest version of
impact of the implementation of the HRIS in Brazil (15). the SIGRHS package have been made available on the
Given the decentralized management of the health sys- Internet (17) (see also Box 9.2). The freely accessible,
tem in this context, homogeneity was not expected in adaptable, web-based or network-based programmes
the information needs pertaining to HRH, expectations should help facilitate the integration of data from local
for the HRIS, capacities to implement and use the sys- health facilities with other information on HRH produc-
tem, or collected responses to its evaluation. However, tion and management processes, reduce operational
the results do indicate an overall positive direction in costs, strengthen data dissemination and use, and
Table 9.3 Selected indicators and results from the evaluation of the institutional impact of the HRH
information and management system in Brazil
Indicator Expected outcome Response options for institutional impact Survey

results
Yes No (n=9 sites)
Utilization of the Improvements The implementation of the The implementation of Yes=100%
information in in the processes HRIS influenced at least three the HRIS did not alter
the HRIS in the of planning, of the following processes: the HRH management
processes of HRH management and processes.
planning and capacity of HRH – professional qualification;
management – workforce planning;
– team building;
– continuing education;
– plans for staffing,

career development and
remuneration;
– decentralization of HRH
management.
Utilization of the Induction to The implementation of the The implementation of Yes=78%
information in innovations/ HRIS influenced at least one the HRIS did not alter
the HRIS in the reformulations of the following processes: the management of the
processes of health of management health system.
system management processes in – restructuring of local health
at the local level decentralized health services;
systems – integration of information
on HRH with other health
system databases;
– implementation of
programmes for
performance improvement.
Source: Pierantoni (15).
124
expand the evidence base for informing decision-mak- monitor the complex dynamics of the health workforce.
ing for management change within organizations. A generic and conceptually sound framework for HRIS
implementation is neither useful nor sufficient. Field
experience shows that a “one framework fits all” rec-
9.4 Summary and conclusions ipe is not the solution, given the unique context of the
health workforce in each country.
Due to the growing interest in health system strength-
ening and its critical human resources function in At the same time, this chapter has revealed a number
particular, the need for documenting and analysing of commonalities across various country-specific
trends and results pertaining to the health workforce experiences and lessons learnt that are crucial for
is becoming part of the global and national agenda in the development of an operational HRIS that is viable,
monitoring health system performance. Although the effectively functioning and sustainable within the rou-
instalment and maintenance of a functioning HRIS at tine health management system establishment.
the national, regional and global levels is not an easy
endeavour, it is doable, given the political commit- Firstly, while other partners and local representatives
ment and evolving quest in many countries to generate should participate throughout the process, the minis-
evidence-based policies to make progress in tack- try of health has to be the owner in establishing and
ling HRH-related challenges. The development and institutionalizing the information system. Moreover, insti-
sustainability of a comprehensive HRIS to inform deci- tutional readiness is one of the most crucial conditions
sion-making is a leadership, financial, educational, to making an HRIS both functional and sustainable.
partnership and management issue (18). Bringing together all HRIS stakeholders, often for the
first time in the same room, ensures that information is
Well-defined indicators and high-quality data to shared and helps in identifying the data that do exist
continuously monitor the status of the health work- and reaching consensus for selection of appropri-
force and evaluate outcomes are being increasingly ate indicators and benchmarks. Training and capacity
sought in many countries, particularly by ministries of building among stakeholders of the system is critical
health. However, prospective countries are not look- from the early planning stages, not only in data analy-
ing for merely cutting-edge technology and technical sis and basic information technology skills, but also in
assistance with attractive manuals and guidelines, data dissemination and use to inform decision-making.
but practical options and easy-to-use methods to
Box 9.2 Implementation and use of the SIGRHS information system for HRH
management in Guinea Bissau
The SIGRHS HRH information and management system was adopted and implemented by the Ministry of
Public Health of Guinea Bissau as part of a partnership between the Brazilian Ministry of Health and the
World Health Organization to support the use of information systems for health workforce development in
Member States of the PALOP organization (Países Africanos de Língua Oficial Portuguesa).
To implement the SIGRHS software for data collection, storage and analysis in Guinea Bissau, it
was necessary to adapt the system to the local specificities. A new configuration was based on the
administrative organization of the country’s health system, the organizational structure of the Ministry of
Public Health, the identification and distribution of health facilities and services at the district level, the
positions and location of health workers, and the structure of the education system for producing skilled
health service providers.
It was also necessary to upload about 2400 pre-existing facility staffing return records from their previous
form in a simple spreadsheet (Excel file). Incorporating these initial data into the adapted SIGRHS system
enabled preliminary reports to be run and an initial profile of the health workforce to be drawn. The results
showed that the health workforce in the public sector of Guinea Bissau is largely constituted by workers
with lower levels of education; only 11% were reported to have university-level education. The majority of
health workers (63%) were to be found in primary care facilities, 25% in secondary-level facilities and the
remaining 12% in large regional-level facilities.
125
A number of practicalities also need to be considered. References

The relevance of the HRIS for review of the national
1. Global HRIS strengthening. Capacity Project (http://
HRH situation critically depends on the quality of the www.capacityproject.org/hris, accessed 24 January
data used to feed it. Dimensions against which to meas- 2009).
ure data quality include the validity, reliability, integrity,
2. Hotz VJ et al,, eds. Administrative data for policy-
precision, truthfulness and timeliness of the system relevant research: an assessment of current
(19). As an example, data validity can be compromised utility and recommendations for development. In:
when individual records are not updated upon retire- Report of the Advisory Panel on Research Uses of
ment, emigration or death. While the development of Administrative Data of the Northwestern University/
University of Chicago Joint Center for Poverty
the HRIS is ongoing, a formal evaluation plan for data
Research, 1998.
quality would include comparing routine administrative
3. Everybody’s business: strengthening health systems
data used for continuous monitoring to periodic sur-
to improve health outcomes – WHO’s framework
vey or census data for purposes of validation. Another for action. Geneva, World Health Organization,
issue closely related to data quality is that of privacy 2007 (http://www.who.int/healthsystems/strategy/
and confidentiality. No data that can specifically iden- everybodys_business.pdf, accessed 10 January
tify individuals should be made accessible for public 2009).
use. Even aggregate data sharing should only come 4. Confidentiality and security of HIV information. Joint
through the permission of the responsible stakeholder. United Nations Programme on HIV/AIDS (UNAIDS)
(http://www.unaids.org/en/KnowledgeCentre/
HIVData/Confidentiality/default.asp, accessed 24
While an idealized planned progression towards a
January 2009).
complete and mature software solution was identi-
fied in this chapter, experience suggests that in some 5. Bowen S, Zwi AB. Pathways to “evidence-informed”
policy and practice: a framework for action. PLoS
cases less is better, and patience is paramount to leave Medicine, 2005, 2(7) (http://medicine.plosjournals.
space and time for understanding, ownership and org/archive/1549–1676/2/7/pdf/10.1371_journal.
engagement before a solution is rushed in. In partic- pmed.0020166-L.pdf, accessed 24 January 2009).
ular, although there are proprietary software solutions 6. Ministry of Health, Republic of Uganda (http://www.
for HRIS, the choice of free and open source software health.go.ug/, accessed 24 January 2009).
offers significant advantages for sustainability at the 7. Global HRIS strengthening: iHRIS software suite.
country level where cost-effective solutions are essen- Capacity Project, 2008 (http://www.capacityproject.
tial (1). Open source solutions decrease the costs of org/hris/suite, accessed 24 January 2009).
implementation and ongoing maintenance of the HRIS, 8. Mshana E. Addressing the human resources for
and provide a global community for support and con- health crisis in Uganda. Presented at the National
tinuous improvement. Many governments and health Symposium on Developing Human Resources for
systems have elected to standardize on open source Health hosted by the Uganda Ministry of Health,
Kampala, 2–4 April 2006.
technologies and build the necessary capacity to sup-
port them in order to realize the benefits. 9. Ahmed HS. Monitoring and evaluation system of
the Federal Ministry of Health: the current situation
and future perspective. Sudanese Journal of Public
Unfortunately, some countries remain reluctant to make Health, 2008, 3(1):39–40.
data publicly available for further analysis, notably in
10. Sudan National Health Policy, 2006. Federal Ministry
contexts of severe HRH shortages. It is highly rec-
of Health of the Republic of the Sudan, 2006.
ommended to link results of the information system to
problem solving among experts in health workforce 11. Instituto Brasileiro de Geografia e Estatística,
Ministério do Planejamento, Orçamento e Gestão
planning and management. Sharing consistent infor- (http://www.ibge.gov.br/home/, accessed 24 January
mation is important so that improved human resources 2009).
strategies can be compared and used by others.
12. Pierantoni CR. A informação para a gestão local
Intercountry knowledge sharing as part of the HRIS de recursos humanos da saúde. In: Ministério da
strengthening process provides models that help avoid Saúde, Observatório de recursos humanos em
repeating mistakes and standardizes HRH information saúde no Brasil. Rio de Janeiro, Editora Fiocruz,
and evidence across regions and countries. As will be 2003.
further discussed in Chapter 12 of this Handbook, HRH 13. Perez JRP. Avaliação do processo de
observatories are one valuable mechanism that can be implementação: algumas questões metodológicas.
used for widely disseminating information and evidence In: Rico EM, ed. Avaliação das políticas sociais: uma
questão em debate. São Paulo, Cortez, 1999.
for effective practices at the national, regional and glo-
bal levels. 14. Pierantoni CR, Viana AL. Avaliação de processo na
implementação de políticas públicas: a implantação
do Sistema de Informação e Gestão de Recursos
Humanos em Saúde (SIG-RHS) no contexto das
126
reformas setoriais. Physis: Revista de Saúde

Coletiva, 2003, 13(1):59–92.
15. Pierantoni CR. Reformas da saúde e recursos
humanos: novos desafios x velhos problemas. Uma
estudo sobre recursos humanos e as reformas
recentes da política nacional de saúde [doctoral
thesis]. Rio de Janeiro, Universidade do Estado do
Rio de Janeiro, 2000.
16. Programa de Qualificação e Estruturação da Gestão
do Trabalho e da Educação no Sistema Único de
Saúde: ProgeSUS. Ministry of Health of Brazil, 2007
(http://www.saude.gov.br, accessed 24 January
2009).
17. Department of Health Planning and Administration:
Workstation of the Health Human Resources
Observatory Network, Social Medicine Institute,
University of the State of Rio de Janeiro (http://www.
obsnetims.org.br, accessed 24 January 2009).
18. World Health Organization. Strengthening national
and subnational departments for human resources
development. Health Workforce Development
Series, Issue 1. Cairo, World Health Organization
Regional Office for the Eastern Mediterranean, 2008
(http://www.emro.who.int/dsaf/dsa954.pdf, accessed
24 February 2009).
19. Selvaggio MP. Producing good data: data quality
management. Presented at the professional
development workshops prior to the South African
Monitoring and Evaluation Association (SAMEA)
Conference, Gauteng, 26–28 March 2007.
127
128
Understanding health workforce
10 issues: a selective guide to the use

of qualitative methods
TOMAS LIEVENS, MAGNUS LINDELOW, PIETER SERNEELS
10.1 Introduction structure and precise measurement in favour of more

open-ended approaches to examining, analysing and
Data and evidence are critical ingredients in the design interpreting behaviours and phenomena. Although it is
of effective health workforce policies and strategies. often hard to generalize from qualitative analyses, the
Policy-makers need information on the size and dis- rich information they generate can be of considerable
tribution of the health workforce, inflows and outflows, value. Consider, for instance, an effort to understand
absolute and relative earnings of health workers, and the impacts of current or planned health labour market
so forth. They also need to understand why particu- reforms. These impacts are likely to depend on a wide
lar labour market outcomes are being observed, and range of factors, many of which are difficult to measure
how different policies or reforms are likely to impact on (for example preferences and expectations of health
the outcomes of interest. For example, if a country is workers, social and cultural conditions, and implemen-
having problems deploying staff to rural facilities, why tation arrangements). Moreover, some of the effects of
is that the case? Would the introduction of contractual reforms on outcomes such as job satisfaction, team-
obligations for rural service, perhaps combined with work or workplace behaviour are themselves difficult to
financial incentives, resolve the problem, or are other measure, and may not even be anticipated.
approaches needed? Or, if the problem is that too few
health workers opt for a career in the public sector, It is clear, then, that qualitative methods can be useful to
favouring private employment instead, then policy- understand a context in which policies or interventions
makers need to understand the appeal of the private are introduced, anticipating behavioural responses and
sector, and how health workers are likely to respond impacts, identifying implementation issues and shed-
to different forms of government regulations to restrict ding light on how the policy is actually perceived and
such movements. understood by different stakeholders. The aim of this
chapter is to provide some guidance on how to design
Other chapters in this Handbook have highlighted how and implement qualitative studies. In doing so, it draws
quantitative data from administrative sources, health on a rich methodological literature and a wealth of
facility assessments and population-based sources applied work, but it does not seek to be a comprehen-
can help meet policy-makers’ needs for information sive guide to qualitative methods. The text is focused
and evidence. Such data can shed light on patterns on practical tips and guidelines for qualitative health
and trends in the health workforce, the determinants of workforce research. For more fundamental issues,
health worker behaviour and choices in the labour mar- including epistemological issues, the reader is referred
ket, and the impact of health workforce policies and to the work that addresses these issues more com-
regulations. However, quantitative data have impor- prehensively than is possible within the scope of this
tant limitations, in particular when it comes to issues chapter (1–11).1 To illustrate both methodological points
or phenomena that are poorly understood or difficult
and the potential value of qualitative work, the chapter
to classify and measure. Recognizing these limitations,
draws on two studies from Ethiopia and Rwanda.
this chapter focuses on how qualitative methods can
be used to support and complement quantitative anal- The remainder of the chapter is organized as follows.
ysis in understanding health workforce issues. The next section describes qualitative methods and
Qualitative methods include a broad array of

1 For general background on methodological and interpretive
techniques and approaches for describing and under- issues in qualitative research, see for example Becker (1)
standing social phenomena. In contrast to quantitative Bryman (2), Bryman and Burgess (3), Denzin and Lincoln (4),
Flick (5), Greene (6), Seale (7), Silverman (8, 9), Walker (10)
methods, qualitative methods eschew the focus on and Wolcott (11).
129
approaches and looks into their usefulness for health also have a number of strengths and weaknesses in
worker studies. The following section discusses selected common.
issues in the design and implementation of health worker
research. The last section concludes with some sugges-
10.2.2 Strengths and weaknesses
tions on how qualitative work on health workforce issues
can be expanded and strengthened in the future. In thinking about the strengths and weaknesses of qual-
itative methods, it is helpful to start by contrasting them
with quantitative methods. The latter rely on structured
instruments to collect quantitative data that correspond
10.2 Qualitative methods: to predetermined categories and classifications. The
their value and potential resultant data can easily be described and summa-
rized. Moreover, provided the data are representative
10.2.1 Qualitative research defined of a broader population, statistical techniques can be
The most frequently used definition of qualitative used to make inferences about that population, and to
research is “research that is not of a quantitative explore multivariate relationships.
nature”. An alternative approach is to concentrate on
what all qualitative research has in common, namely Qualitative methods, in contrast, tend to impose
the use of data in the form of text, as this in itself has less structure on the data and this has a number of
profound implications for the techniques used in both advantages:
data collection and analysis (1, 12, 13). Another dis- t It allows for open-ended responses and interactive
tinct characteristic of qualitative research is that the exploration of issues with the respondent. This can
researcher tends to be heavily involved in the collection shed light on issues that are difficult to measure,
of the data. More than in quantitative research, on-the- such as beliefs, feelings, values and perceptions. A
spot judgement is needed regarding what to register, welcome side-effect of this approach is that it also
where to focus and where to go deeper. At the level registers the participants’ vocabulary, which is often
of analysis there is a similar challenge, as the results useful for designing quantitative surveys, and for
depend to a certain extent on the interpretation of the communicating with stakeholders about the issues
data, more so than in quantitative research. under study.
t An open-ended interaction with the respondent(s)
In practice, “qualitative research” applies to a wide range helps create the necessary trust and rapport to
of diverse research techniques. In this chapter, three explore difficult or sensitive issues, or to engage
main methodological approaches are distinguished: with marginalized or difficult-to-reach groups.
(i) individual interview; (ii) focus group discussions; For example, issues such as corruption and other
and (iii) ethnography or participant observation (Box forms of illicit behaviour are notoriously difficult to
10.1). Each of the three techniques discussed here2 capture through surveys, but can be effectively
has its own advantages and disadvantages, but they explored – although not measured – using qualita-
tive approaches.
2 This chapter does not consider other qualitative techniques, t Qualitative methods can help build an understand-
such as language-based approaches (for example discourse
conversation analysis), projective techniques, case studies ing of behaviours and of the relationship between
and text analysis. Tesch (14) distinguishes 26 different types different variables. For example, they can be used
of qualitative approaches in social research. For an overview, to explore patterns of behaviour (such as the uptake
see the sources listed in footnote 1 of this chapter.
Box 10.1 Three important techniques in qualitative research
Individual interview. This is a generic term to describe interviews of a qualitative nature. The interviewer
has a choice to follow a structured, semi-structured or unstructured approach. A qualitative individual
interview differs from a quantitative interview in that it leaves the answers open and registers them as text,
whereas in quantitative surveys the response options are mostly precoded.
Focus group discussions. Group discussions where the participants are asked to discuss specific topics
openly. They tend to be semi-structured or unstructured in nature.
Ethnography and participant observation. Data collection in which the researcher is immersed in a
social setting for some time in order to observe and listen with a view to gaining an appreciation of the
culture and processes of the group.
130
Understanding health workforce issues: a selective guide to the use of qualitative methods
Table 10.1 Contrasting qualitative and quantitative approaches
Qualitative Quantitative
Words Numbers
Point of views of participants Point of view of researcher
Researcher close Researcher distant
Theory emergent Theory testing
Unstructured Structured
Contextual understanding Generalization
Rich, deep data Hard, reliable data
Meaning Behaviour
Tends to follow an inductive approach Tends to follow a deductive approach
Source: Adapted from Hammersley (15), Bryman (16) and Halfpenny (17).
of rural service of health workers with different socio- easy-to-understand images and concepts. These fea-
economic and demographic profiles) or the nature tures, as reported in Table 10.1, are generalizations and
of a relationship between different measurable var- do not necessarily hold in every case; nevertheless
iables (such as education and external migration). they are helpful and underline that qualitative research
t Through purposive selection of either individual can, for example, be very useful where theory is weak
interviewees or focus group participants, qualitative or inexistent to fine-tune hypotheses that can be tested
methods can elucidate differences in views and per- with quantitative data or to better understand findings
spectives between and within groups, and explore from quantitative research (15–17). Moreover, by dem-
the reasons for these differences. onstrating how stakeholders perceive and discuss a
policy or phenomenon, and by identifying responses
Inevitably, the merits of qualitative methods come at a and behaviours that a policy is likely to engender,
cost. Clearly, one of the limitations of qualitative data qualitative methods may support the development of
is that they cannot be easily described and summa- questionnaires for quantitative surveys (2, 12, 15, 18,
rized. As a result, it is difficult to make statements about 19).3
the magnitude and relative importance of phenomena
addressed in qualitative work. More importantly, per- Reflecting the strengths and weaknesses outlined
haps, respondents in qualitative work tend to be small above, qualitative methods have been used exten-
in number and purposively selected. This means that sively in health systems research, including in efforts
findings cannot be meaningfully thought of as repre- to understand health-seeking behaviour, identify com-
sentative of or generalizable to a broader population. munity health needs, assess health-care quality and
Another potential limitation is the relatively weak objec- client satisfaction, and develop programmes for hard-
tivity that results from the close involvement of the to-reach groups (for example sex workers or people
researcher in the collection of the data and the interac- who use injecting drugs). It is not surprising, therefore,
tive nature of data collection. This feature of qualitative to find a substantial and rapidly growing body of quali-
methods also means that qualitative work is difficult to tative work on health workforce issues.
replicate, either by a different researcher or in a dif-
ferent context. While these weaknesses of qualitative
10.2.3 Qualitative health worker research
methods are at least to some extent inherent in the
approach, they can be partially overcome through The field of health worker research is relatively young.
careful design and implementation, and through trans- An important consequence of this is that there is little
parent documentation of methods and approaches theoretical framework to study health worker behav-
(see section 10.3 below). iour. Existing quantitative studies, although booming
in number, therefore remain largely descriptive in
The strengths and weaknesses of qualitative methods
3 For more on combing qualitative and quantitative
explain the appeal of combining qualitative with quanti- approaches, see Bryman (2), Holland and Campbell (12),
tative research, as the two complement each other very Hammersley (15) and Morgan (18). For resources and
well. Several authors have tried to capture the differ- ongoing discussion on the differences, complementarities
and tensions between the two approaches in poverty
entiating features of the two methods in powerful and research, see Centre for International Studies (19).
131
nature and can substantially benefit from a qualita- example of what can be obtained from this method is
tive approach. Qualitative research can for example offered by Jaffré and Olivier de Sardan (20), who com-
help to understand and categorize the complex envi- bine participant observation with site visits, structured
ronment in which the health worker functions. It can interviews and focus group discussions in their sem-
also assist in unravelling the motivation behind cer- inal study of doctor–patient relationships in five West
tain behaviours, and it can provide inputs on how to African countries.
improve the measurement of health worker perform-
ance. Researchers studying health worker behaviour In some cases, qualitative studies are implemented as
are increasingly aware of these and other potential con- a complement to quantitative research, either to under-
tributions of a qualitative approach, as is clear from the stand puzzles thrown up by the quantitative research or
increasing number of studies on health workers using in preparation of quantitative research. Box 10.2 gives
this approach. an illustration of how qualitative research can help to
prepare quantitative analysis.
Table 10.2 illustrates how qualitative research has been
used in a variety of ways to address a broad set of Although qualitative work on health workforce issues
health workforce issues, ranging from broad diagnos- has generated a lot of insights already, there is plenty
tic exercises that have sought to explore the motivation of scope for further work. This includes the exploration
and behaviour of health professionals in the workplace of similar issues to those highlighted in Table 10.2 in
and labour market, to more focused studies of specific different contexts, and the use of qualitative methods
phenomena such as dual practice or migration. Most to better understand a broader set of labour market
studies have focused on issues that are poorly under- issues in the health sector. This issue will be revisited in
stood or difficult to measure in a quantitative way. This the concluding section of the chapter.
includes, for example, health worker performance and
motivation, corruption and the relationship between
providers and clients of health services. The major- 10.3 Issues in the design
ity of studies use individual qualitative interviews or
and implementation of
focus group discussions, or both. Participant obser-
vation has had more limited use, possibly because of a qualitative study
the private nature of the relationship between health- While the former section outlined the value and potential
care professionals and their patients. However, a good uses of qualitative methods in health worker research,
Box 10.2 Qualitative research to inform quantitative work: absenteeism in

Ethiopia and Rwanda
To date, most of the work on absenteeism has been quantitative in nature, testing specifications derived
from standard economic theory. When applied to health workers, the work is usually explorative in
nature. An important reason is that the existing theoretical frameworks on absenteeism seem to be less
appropriate to study human resources in the health sector, especially in developing countries.
To address this issue, two qualitative studies were implemented in Ethiopia and Rwanda, using focus
group discussions sharing a similar design. These studies were labelled pre-research, as their main
purpose was to provide inputs for future quantitative work. Financial support for their implementation was
provided by the World Bank, the Bill & Melinda Gates Foundation and the Norwegian Government.
The studies confirmed that existing theory falls short of providing an appropriate framework for analysing
absenteeism among health workers in these two countries. The focus group discussions suggest that the
theoretically predicted correlates of absenteeism (wages, contracted working hours and expected cost of
detection) affect absenteeism as indicated by theory, but that their relative importance seem to depend
on the country context. The studies also indicate that additional factors, such as access to a second
job, intrinsic motivation, job mobility and perceived health risks, play an important role in explaining
absenteeism levels. This then argues for revisiting theory and empirical estimation to take this broader set
of determinants into account.
Source: Adapted from Serneels, Lindelow and Lievens (42).
132
Table 10.2 Illustrations of qualitative research on health workers
Topic Focus of studies Approach Reference

Coping strategies Study of informal economic activities Combined quantitative and McPake et al. (21)
of health workers in Uganda, including qualitative approach
the leakage of medicine supply, informal
charging of patients, and mismanagement
of revenues raised from the formal
charging of patients
Assessment of extent of and reasons Mix of open- and closed-ended Ferrinho et al. (22)
for drug pilfering by health staff in questions in self-administered
Mozambique and Cape Verde survey to small, purposively
selected sample
Study of role of professional identity, Qualitative in-depth interviews and Kayaddondo and
motivation and other factors in focus group discussions in four Whyte (23)
understanding survival strategies in purposively selected facilities
response to health system reforms in
Uganda
Exploration of how health workers in Key informant interviews and focus Israr et al. (24)
Cameroon cope with salary cuts, and of group discussions and interviews
the impact of coping strategies on service with service users
quality
Dual practice Exploration of how financial and non- Quantitative survey combined Gruen et al. (25)
financial incentives shape job preferences with in-depth interviews for a
of doctors in Bangladesh who work both in subsample
government health services and in private
practice
Exploration of supplementary income- Mix of open- and closed-ended Ferrinho et al. (26)
generating activities of public sector questions in self-administered
doctors in the Portuguese-speaking African survey to small, purposively
countries, and of reasons why they have selected sample
not made a complete move out of public
sector
Exploration of nature of dual practice in Semi-structured interviews with Jumpa et al. (27)
Peru, including key factors that influence purposively selected health
individual decisions to engage in dual workers
practice
Informal charging Investigation of out-of-pocket payments In-depth interviews and focus Belli, Gotsadze
and corruption for health services, formal and informal, in group discussions with users and and Shahriari (28)
Georgia providers
Motivation Exploration of a broad range of motivational Contextual analysis, qualitative Franco et al. (29)
determinants and outcomes in two assessment and quantitative
hospitals in Jordan and two in Georgia analysis
Assessment of key motivating factors for Focus group discussions, in-depth Dieleman et al.
health workers and managers in Viet Nam interviews and exit interviews (30)
Assessment of the role of non-financial Semi-structured interviews with Mathauer and
incentives and professional ethos in doctors and nurses from public, Imhoff (31)
motivating health workers in Benin and private and nongovernmental
Kenya facilities in rural areas
Deployment and Exploration of barriers to recruitment and Focus group discussions Brewer et al. (32)
retention retention of nurses in New York State,
United States of America
Continues…
133
Table 10.2 continued from previous page…
Topic Focus of studies Approach Reference

International Understanding the reasons for out- In-depth interviews and focus Muula and
migration migration from the health sector in Malawi group discussions Maseko (33)
Understanding the motives and experiences In-depth interviews over a Robinson and
of Indian migrant health workers to six-month period Carey (34)
hospitals in the United Kingdom
Experiences of migrant nurses in hospitals In-depth interviews Hardill and
in the United Kingdom MacDonald (35)
Perceptions and opinions of those In-depth interviews with hospital Troy, Wyness and
involved in recruitment and migration from directors and overseas nurses McAuliffe (36)
developing countries
Quality/practice Exploration of how organizational issues, Minimally structured in-depth Jewkes,
patterns professional insecurities, and other factors interviews and focus group Abrahams and
contribute to patient neglect and poor discussions held with patients and Mvo (37)
clinical management in South Africa staff
Stocktaking and analysis of poor quality Participant observation combined Jaffré and Olivier
of care in five capital cities of West African with in-depth interviews, site visits de Sardan (20)
countries and focus group discussions
Assessment of changes in prescription In-depth interviews and focus Asenso-Okyere et
patterns and quality of care associated with group discussions al. (38)
introduction of user fees in Ghana
Broad diagnostic Exploration of experiences of health Focus group discussions and Manongi,
exercises workers in primary health care in Tanzania in-depth interviews Marchant and
in terms of motivation, satisfaction, Bygbjerg (39)
frustration
Exploration of performance and labour Key informant interviews and focus Lindelow and
market issues for doctors and nurses in group discussions with nurses, Serneels (40)
Ethiopia doctors and service users
Exploration of performance and labour Focus group discussions with Lievens and
market issues for doctors and nurses in providers and users of health Serneels (41)
Rwanda, including a focus on career and services and with people living
performance in the face of HIV with HIV
this section addresses the major practical challenges individual and group interviews lies in the interaction
in design and implementation. Its focus is on interviews between group members. Group discussions, which
and discussions (terms that are used interchangeably typically take place in groups ranging from four to nine
in this chapter, abstracting from the concept whereby people, allow researchers to elicit a multitude of views
free interaction is allowed in each). A third technique on a topic and explore and contrast the opinions of dif-
previously described, ethnography and participant ferent participants. They also allow the researcher to
observation, is rarely applied in health workforce collect data on a large range of behaviours in a relatively
research and so is not elaborated here as it requires short timespan. One of the risks with group discussions
a more specific methodology and different skills. The is that some individuals dominate the discussion, while
references cited in footnote 1 of this chapter provide other participants refrain from expressing their ideas
further comprehensive guidance on techniques and because they are not in line with the prevailing view.
methods. This may lead to a “false” consensus.
The advantage of individual (or face-to-face) interviews

10.3.1 What type of interview?
is that they offer more room for clarification and expan-
One of the first issues that will need consideration is sion of the discussion with each individual. The safe
whether interviews will be conducted with individuals or and private environment is also more likely to create an
with groups of individuals. The key difference between atmosphere of trust and openness, which also makes it
134
easier to address issues that are taboo. Moreover, the approaches, as is often done in health worker research
researcher can link the results of the discussions with (see for example the illustrations in Table 10.2).
characteristics of the respondent (such as professional
experience or socioeconomic background). In general, A second choice that researchers need to make con-
individual interviews are easier to manage than group cerns the extent to which the interview or discussion
discussions. However, it can be difficult to keep indi- will be structured. Box 10.3 presents a short descrip-
vidual discussions focused, especially when they are tion of commonly distinguished types of interviews
unstructured, and there is a risk of generating a large according to their degree of structure (2, 13, 43, 44).
volume of data that will require significant effort to ana- In reality structured and unstructured interviews are on
lyse. Individual interviews also tend to be more time a continuum. The essential discriminating factor is the
consuming and costly compared to group interviews.4 room allowed to respondents to express and develop
their own points of view, and to the interviewer to ori-
Table 10.3 provides some guidelines on when to use ent the discussion. If the objective is to collect rich and
group versus individual interviews, also taking into deep data on topics on which little is known, the script
account logistical and financial considerations (43). should be less structured. This comes at a price of
A frequently used third option is to combine both yielding data that are less comparable across respond-
ents. Also, the less structured the interview, the more
skill and experience is required from the interviewer,
4 When the individual interviews are with experts, they are often and later from the data analyst.
referred to as “expert interviews”. For example, in order to
understand the policy and institutional context in which health
workers operate it can be useful to interview government A special case of group interview is the focus group
officials at national and district level, representatives of discussion, which is a semi-structured group discus-
health professional associations and nongovernmental
organizations active in the health sector, and others. sion around a specified set of topics. Here the interview
Table 10.3 Group discussions or individual interviews?
Factors to consider Use group interviews when… Use individual interviews when…
Group interaction Interaction of respondents may stimulate Group interaction is likely to be limited or
a richer response or new and valuable non-productive
insights
Group or peer pressure Group or peer pressure will be Group or peer pressure would inhibit
valuable in challenging the thinking of responses and cloud the meaning of
respondents and illuminating conflicting results
opinions
Sensitivity of subject matter Subject matter is not so sensitive that Subject matter is so sensitive that
respondents will temper responses or respondents would be unwilling to talk
withhold information openly in a group
Extent of issues to be covered There is a need to cover a small number There is need to cover a greater number
versus depth of individual responses of issues on a topic on which most of issues on a topic that requires a
respondents can say all that is relevant in greater depth of response per individual
less than 10 minutes
Requirement for interview guide Enough is known to establish a It may be necessary to develop the
meaningful topic guide interview guide by altering it after a
series of initial interviews
Logistics requirement An acceptable number of target Respondents are dispersed or not easily
respondents can be assembled in one assembled
location
Cost and training Quick turnaround is critical, and funds Quick turnaround is not critical, and
are limited budget will permit higher cost
Availability of qualified staff to Focus group facilitators are able to Interviewers are supportive and skilled
conduct the interview control and manage group discussions listeners
Source: Adapted from Frechtling, Sharp and Westat (43).
135
Box 10.3 Types of qualitative interviews according to their degree of structure
Structured or standardized interview. This type of interview uses a script that typically contains a number
of predetermined questions that are presented to the interviewee one after the other. Since all interviewees
are given identical cues, their answers can be reliably aggregated, which is often the main objective of the
technique. Data for quantitative analysis can also be collected, especially if the answers to the questions
are precoded.
Semi-structured interview. This term covers a wide range of cases but mostly refers to interviews where
the script contains a series of questions that, in contrast to structured interviews, can change sequence;
the interviewer also tends to have some latitude to ask further questions in response to what are seen
as significant replies. The answers are open ended and there is more emphasis on the interviewee
developing arguments.
Unstructured interview. The interviewer typically has only a list of topics or issues, often called an
interview guide or aide-memoire, that he or she expects to cover in the interview. The style of questioning
is usually informal and the phrasing and sequencing of questions will vary from interview to interview.
Sources: Adapted from Bryman (2), Denscombe (13), Frechtling, Sharp and Westat (43) and Atkinson (44).
script is mainly used as a memory support and the Another benefit of writing out the complete script is that
main technique is to prompt and probe the partici- it facilitates the reproduction of the interviews, which
pants. Prompt questions aspire to open the discussion may be of interest in order to contrast findings from dif-
on a topic in a neutral way, providing an input without ferent contexts. For example, using a similar script and
imposing preset ideas, while probing is used to trigger research design, health workers in Ethiopia were willing
deeper responses. to discuss corrupt practices such as absenteeism, illicit
charging, embezzlement of materials, pilfering of drugs
Writing out the complete script is good practice since it and holding a private practice within the public sector,
allows full transparency about the research design. Box while in contrast, health workers in Rwanda were more
10.4 shows an extract of a script for focus group dis- reluctant to acknowledge these malpractices. However,
cussions with health workers in Ethiopia.5 It starts with after further probing, health workers admitted that
a prompt stating the potential of performance problems there were corrupt practices, but that these were usu-
among health workers and then probes on the issues ally small scale, that they did not occur frequently and
of motivation, time use, labour market institutions and that they had become less common than in the past.
corruption. As the script served mostly as a check- Some illustrative quotes are reported in Box 10.5. The
list the topics were usually discussed when they came comparative analysis indicates that corrupt practices
up, but the script helped to guide the discussion and are less pervasive in Rwanda than in Ethiopia and the
to generate deeper data (the prompts were informed data also offer explanations why this is the case. The
by a brief preparatory literature review). Other classic health system in Rwanda provides better monitoring
probing and prompting techniques – such as remain- and accountability (by involving community workers in
ing silent, repeating the question, repeating the last few the management of health facilities, for example), is in
words from a discussant, offering an example as well the process of adopting performance pay (which also
as asking for an example, requesting a clarification or stimulates reciprocal monitoring) and provides better
more detail – were also used, though they are not visi- enforcement of sanctions.
ble in the script.
10.3.2 Participant selection

5 The full script, which is available upon request, contained the In contrast to quantitative research, where represent-
following sections (with aspired time allocation in brackets ativeness is the main concern for sample design, the
in minutes): Introduction and warm up (10’); Professional main objective of participant selection in qualitative
training, recruitment and career paths (30’); Contracts,
compensation, benefits and outside economic activities (20’); research is to include individuals that are well informed
Job satisfaction, motivation, and performance (30’); Impact about an issue, and have a broad range of views and
of HIV/AIDS (10’); Coaching other people (10’); and Thanks, experiences. This contributes to qualitative interviews
payment and signature (10’).
136
Box 10.4 Extract of semi-structured interview script of qualitative health

worker study in Ethiopia and Rwanda
JOB SATISFACTION, MOTIVATION AND PERFORMANCE
Time allocated: 30’ Time started: ……
Prompt:
In many countries, users complain about health services. For example, there are often complaints that
health workers are not very motivated, that they do not spend as much time as they should doing their
job, that they are not good at doing their job, and even sometimes that they are involved in illegal activities
such as stealing drugs and material and charging too much for services. How do you feel that the situation
is in Ethiopia?
Let’s start with the issue of motivation…
Issues to probe
Do you think most health workers are satisfied with their job?
Why do you think some health workers are unsatisfied in their job?
t payment, lack of equipment
Do you feel that there is a strong commitment to delivering good health care?
What do you think is the most important source of motivation for health workers?
t money and benefits
t the activities and responsibilities of the job
Do you think the commitment of health workers is different in different sectors (private, for-profit, not-for-
profit)? Why?
(…)
Let’s turn to corruption and inappropriate behaviour …
Issues to probe
To what extent do you feel there is a problem of corruption in the health sector?
t use of equipment for private purposes
t overcharging
t stealing/leakage of drugs and other material
Why do you think these problems arise?
What is done in your workplace to reduce these problems?
What sort of disciplinary measures are available and are they being used?
Source: From the authors.
being informative and generating rich data (2, 13). The choice of group size depends on the extent to
Selecting the participants for group discussions raises which one aspires to have an animated discussion ver-
additional issues related to group size, the number of sus an intimate exchange. Discussions in larger groups
groups, within-group dynamics and the desired variety are often more animated, while small groups may be
in participant profiles within a group. more conducive to discussions about sensitive topics.
137
However, this may not be realistic since the number

Box 10.5 Selected quotes from of focus groups is often fixed during the study design.
qualitative interviews in Ethiopia Another issue is within-group homogeneity. Status
and Rwanda homogeneity is for example a common concern, as
the interaction between participants of a very differ-
ent social status may be less productive. In the case
Corruption came and spread widely in the of health workers it seems advisable to have separate
last ten years. Now corruption is prevalent groups for different cadres. This point is of particular
among people at all levels of education and importance for research that addresses sensitive issues
in all areas. such as absenteeism, pilfering or other illicit behaviour.
—Medical officer in a provincial town in Ethiopia Workers that are in a hierarchical relationship are likely
to be uncomfortable about openly sharing their views
I have worked in a private pharmacy. Drugs (for example nurses with doctors).
come illegally and we know their source…
they are taken from public facilities. A final question is how within-group dynamics can be
assured and how much variety of opinion is wanted
—Health-care assistant in a provincial town
within a group. To ensure constructive group dynam-
in Ethiopia
ics, participants should, as a rule, not know each other.
However, the opposite has been argued when there
Low-paid staff might be involved in small
is a taboo regarding the topic under discussion, mak-
bribes… There is some humanity in this and
ing an open discussion between participants that do
assisting it is okay because people working in
not know each other uncomfortable. Kitzinger (45), for
other service institutions such as water supply
example, studying HIV, deliberately selected people
have the benefit of charging for free services.
who knew each other in order to have a discussion that
—Physician in Addis Ababa, Ethiopia
was “as natural as possible”. A potential problem with
this approach is that assumptions commonly shared
It happened that patients paid more than was
between participants are often not made explicit (18).
reported in the register but we identified most
of these cases. I also know that some have When the aim of the research is to explore an issue,
been sacked because there was fraud in the the richest data come from discussions where differ-
payment register. ent points of view are confronted. A good way to ensure
—Auxiliary health worker in Kigali, Rwanda variety of opinion among the participants is to pay spe-
cial attention to the selection of the participants. For a
I have seen a patient give 1000 francs to a study of health worker performance and career choice,
nurse. The nurse was frightened, which shows for instance, it may be useful to consider such potential
that it is not regular practice. influences as gender, age, family status (having chil-
—Health service client in Kigali, Rwanda dren or not), sector of work (public, private for-profit,
faith-based) and work activities (for example having
Health workers do no longer sell drugs taken two or more jobs). Box 10.6 reproduces the selection
from the health centre. This is something from sheet for nurses and midwives used in Ethiopia. Similar
the past. Perhaps in rural areas, but not here in sheets were compiled for doctors, assistant nurses and
the city. users of health services.
—Health service client in Kigali, Rwanda
The same approach can be used for individual inter-
Source: From the authors. views. Once the criteria and characteristics for
selection are determined, the participants are selected
using these criteria. In this case the richness of the data
comes from contrasting during the analysis the opin-
For reasons of comparability and transparency it is ions of the individuals with a different background.
advisable to use groups of similar size throughout the
study.
10.3.3 Study design and data collection:
The number of groups should be such that all partic- how to enhance validity and reliability
ipants that have well-informed and different views on The ways in which data are collected and analysed
the subject are included. Alternatively, groups can be affect the study findings. There is now a rich literature
added until responses can be fairly well predicted. on how best to deal with issues of validity and reliability
138
Box 10.6 Selection criteria for focus groups with eight participants on the
performance and career choice of nurses and midwives in Ethiopia
Selection criteria:
t Five nurses; all should have two years or more of nursing education after having completed
secondary schooling; at least two male, at least two female
t Three midwifes; at least one male, at least one female
t At least two of the females should be mothers
t At least two of the males should be fathers
t At least two participants who combine work in the private sector with work in the public health sector
t At least two participants should be working exclusively for the private for-profit sector
t At least one participant should be working for the private not-for-profit sector
t At least one participant should be working exclusively for the public sector (not engaged in private
health sector work at all)
t None of the participants should be very shy (to participate in a group discussion like this)
t The participants should not know each other and should not be working at the same facility
(46–49).6 Here the focus is on some hands-on issues to team, or the roles taken up by the researchers in the
consider when planning a qualitative study with health team, must be reconsidered if needed.
workers. t Permission. Health workforce studies usually rely
on interviewing both clinical and administrative
t Position of the researcher. The involvement of the health workers. Typically, health workers will want
researcher in the collection and analysis of data is to be assured that the employer has authorized the
a core feature of qualitative research. The sex, age study, especially when interviews take place during
and ethnic origins of the interviewer are known to working hours. A letter from the ministry of health will
influence the replies of the interviewees (13). This in most cases be a necessary element of authori-
may have far-reaching consequences for the valid- zation, although it may not be enough – especially
ity of the study. In applied health worker studies the when the employer is a private provider.
researchers themselves are often part of the health t Venue, timing and duration. The venue, timing
system they survey. Ministry officials and health and duration of the interview should be as conven-
workers are frequently in charge of the implemen- ient as possible for the interviewees. The place of
tation, if not the entire design and management, interview should encourage a private atmosphere,
of the research project. This may bias the results. where the interview is not disturbed. The workplace
A similar concern may arise if researchers are per- – often a health facility or the ministry of health –
ceived as representatives of donor agencies that are may be an appropriate place, except when sensitive
involved in the sector. Since the bias stemming from issues such as absenteeism or corruption are part
this “interviewer effect” is difficult to repair it must of the interview, or when health workers are asked
be anticipated and the composition of the research for an opinion on matters where their viewpoint may
diverge from those of their employers. The acoustic
6 Validity can be seen as the concern with the integrity of the qualities of the venue are important if the interview is
conclusion generated from research, while reliability refers to
recorded. For example, rain on iron rooftops or traf-
the degree to which a measure is stable (for more information
see Guba and Lincoln (46), LeCompte and Goetz (47), fic from a busy road can create loud background
Lincoln and Guba (48) and Mason (49)). Both validity and noise, making recordings inaudible.
reliability can be affected by a number of factors, not least
by the objectives of the study. More worldly issues, such as
t Compensation. When the costs related to partici-
the available budget and the motivation and control of the pating in the interview are not sufficiently covered,
funding body, also play a role.
139
Box 10.7 Extract from the introduction to focus group participants in Rwanda
Good afternoon. First of all we would like to thank you very much for your willingness to cooperate in this
group discussion. The aim of the discussion is to get a better idea of the human resources in the health
sector in Rwanda and to investigate what the problems and opportunities are for health workers. The
discussion today is part of a larger study that looks at a human resources policy for health workers in
Rwanda. The study is financed by the World Bank and approved by the Ministry of Health. If you wish, we
can read you the letter of approval by the Secretary-General.
The immediate objective of our meeting is to hear from health workers themselves how they look at
different aspects relating to their work. We would like to have an open and honest discussion with you
about different aspects of working in the health sector. We are here to listen to you and learn from you.
The discussion will go as follows. [Name] and I will lead the discussion. We prepared a number of topics
on which we will ask your opinion.
It is very important that you feel no restraints to speak your mind. Be as open, direct and sincere as you
can. The statements you make should be based on real-life experiences and observations. It is important
to us to know how the situation really is and not how the situation ought to be. Do say if you feel you don’t
agree with what someone else is saying.
Importantly, we would like to stress that your personal contributions and views will not be shared with
anyone outside this room. To guarantee the anonymity of participation in this discussion, we will not take
your name down. We will also have a number of other discussions, so your view will be balanced out with
what others say.
Also note that the discussion will be recorded. This is because we do not want to miss anything of what
you say. However, nothing of what you say will ever be made public with any of your names attached. The
results will only be written down as the report of a discussion with “a selected group of health workers”.
Lastly, the discussion is estimated to last approximately two hours and you will receive a compensation for
your travel, time and cooperation at the end of the discussion.
Are there any questions or remarks?
there is a risk that participants may not show up or have arrived. Whatever is decided, it is important to
may not be participative. Too much compensation, inform the participants about the modalities before
on the other hand, may induce strategic behav- the interview takes place.
iour, such as trying to please the interviewer with t Confidentiality. Anonymity and confidentiality are
the responses. As a minimum, the cost of trans- important because they may have an impact on the
port to get to the interview should be reimbursed. openness of the respondents. If participants sus-
For additional payments it is advisable to seek guid- pect that their answers can be traced back to their
ance from local practitioners. For example, against person in the final study results, they are less likely
a backdrop of perceived insufficient salaries, many to speak their mind. Therefore, warranting confiden-
health workers are accustomed to receiving finan- tiality and anonymity is usually key to the collection
cial incentives for taking part in seminars, training of quality data. As an illustration, Box 10.7 reprints
and research. Alternatively, financial incentives can an extract of a script containing the introduction
be substituted or complemented by non-financial for a focus group discussion with health workers in
rewards such as food, drinks or gifts. These may Rwanda. The reimbursement modalities for travel
also contribute towards relaxing the atmosphere and expenses and the financial reward for taking part in
offering a drink at the beginning of the interview can the study had been dealt with prior to the interview
help to bridge the waiting time until all participants
140
by a third person specifically dealing with partici- Once the data are video or audio recorded, the inter-
pant selection. views are transcribed literally8 and, if necessary,
t Interview skills. As the researcher takes active translated.9 This written version of the interview is
part in the production and analysis of the qualitative needed for in-depth analysis, and can be used for soft-
data, the more so the less structured the interview ware-supported analysis, which will be discussed in
script, the quality of a study is greatly determined more detail below. Whether the analysis is done man-
by the experience, skill and insight of the modera- ually or with the help of specialized software, the basic
tor. Apart from mastering discussion management approach is the same.
techniques, familiarity with the topic under study is
a necessary condition for generating high-quality In a first step the analyst identifies key themes in the
data (see for example Hurley (50)). transcripts, focusing on issues that are mentioned fre-
quently or consistently, that receive particular emphasis
or for which views expressed in the interviews diverge
10.3.4 Data recording and analysis
in a systematic way. This is typically done by writing
The most common options when it comes to record- codes (referring to themes) next to the transcribed text
ing data are note taking and audio recording. Notes and comparing them visually. The researcher compares
can be taken during or after the interview, with the lat- groups of coded text by leafing through the transcripts
ter being less intrusive. However, since the data then and then brings together themes in main themes and
depend entirely on the researcher’s memory, which is subthemes. The textual data are subsequently struc-
typically subject to bias and partial recall, audio record- tured in matrices, with a different column for each focus
ing (with the agreement of the participants) is the most group or individual, and rows for the different themes
commonly used technique.7 An alternative approach and subthemes. When assisted by a word processing
is to record the interview on video. Although this tech- programme the matrices can be created by moving
nique may provoke unease with the participants at the the text from the transcripts using cut and paste. The
beginning, the negative impact on openness and sin- advantage of the matrix approach is that it facilitates
cerity often disappears once the discussion is under comparison across as well as within groups or individ-
way. An advantage of video recording is that it allows uals, while avoiding excessive structure. It also allows
linking the responses to individual characteristics. Data emergence of issues and opinions as expressed by
embedded in images can also be creatively used, for the participants, without too much interpretation from
example to determine the degree of patient-centered- the researcher(s). Box 10.8 contains an extract from a
ness of a medical practitioner. Both video and audio matrix-based analysis of data collected through focus
recording allow that the raw data can be made avail- group discussions in Ethiopia. The matrix contains
able to other researchers. However, this raises the quotes from three different occupational groups relat-
question whether researchers can maintain their com- ing to their views on the public sector.
mitment to anonymity and confidential treatment of the
data. The increasing possibilities to copy and trans- If the data are analysed with the help of specialized
fer digital data files require that safeguards over and software, the first step is to provide the data from each
above conventional measures must be put in place, interview with an identifying composite label (for exam-
such as protecting files with passwords and requesting ple “Urban doctors Ethiopia” corresponding to the
colleague researchers to provide written guarantees heading of the last column in the matrix presented in
of participants’ anonymity. The base policy to safe- Box 10.8). The next step is to attribute at least one, but
guard confidentiality is to make the data anonymous typically more than one, code to each coherent part
by using unique numerical identifiers that refer to a sep- of text, or entire quote (these codes correspond to the
arate database containing personal information so that themes and subthemes in the matrix approach).
data analysts have no automatic access to participants’
identities.
8 Today’s digital recorders produce high-quality sound files that

can be played on a computer, facilitating the transcription
process. It is still most reliable to transcribe the interviews
manually, as existing software packages that transfer speech
to text perform best when tuned to one voice, and do not
seem sufficiently accurate to transcribe group discussions
or even individual interviews. Software packages are only
available in a limited number of languages, and there also
7 Even in the case of audio recording the researcher often remain problems with recognizing specialized vocabulary.
takes notes during the interview, either to help guide the 9 Because there is a real risk that parts of the data get lost or
discussion or to capture additional information – for example, altered in the process of translation, accuracy is important.
embedded in the body language and facial expression of the This can for example be enhanced by back-translating the
interviewees. text in order to expose deviations from the original.
141
Box 10.8 Matrix-based analysis of data from focus group discussions in Ethiopia
Health-care Physicians and paramedical

assistants Nurses and midwives practitioners
Perceptions The main difference I work in the public. The present challenge for
about the between the public The payment does not the public sector is that
public and private is the compensate enough for the staff has no incentives,
sector payment and the ability the job we are working. In lack opportunity for career
of patients to pay for the private facilities, the development, etc.
the necessary medical payment is good and the
The payment in the public
services. workload is not much. The
does not compensate well for
problem in the private, if
The working condition the physicians’ work. I would
owners get an alternative
is not safe, medical not say that all are equally
they will fire you and bring
equipments and affected by workload. My
in the other.
supplies are lacking. experience is that workload
There are times In the public services, we was not that much a problem.
when staff is obliged provide service for the
I do not agree that in public
to sterilize gloves mass of the population.
facilities there is very high
for reuse while it is There are a lot of things
workload. In the private sector,
disposed off in the lacking. For workers, the
one can have a maximum of
private facilities. salary is not satisfactory. I
three surgeries per day. The
work in the private sector,
Medical supplies in the problem is that one has to
the payment is good.
government are very come in time and go out in
inadequate. The benefit package time; there is good payment.
such as pension, health
(…) In addition to that, the number
insurance and job
of private institutions to
security are good reasons
practice in is limited. Some
for working in the public
might not want to confront
facilities.
being jobless. There is no
(…) refined relationship between
the owner of facilities and the
rest of the profession due to
lack of experience.
(…)
The advantage of a software-based analysis is imme- analysis is more flexible in that the coding is made
diately clear, as it allows for more flexibility in the use easier, and once the coding is finalized, the data can
of codes. For example, it makes it easy to attach sev- be retrieved or combined in different ways using the
eral codes to the same quote, but also to revise any codes. It also allows for a more advanced degree of
hierarchical structure of the codes. The way codes are analysis by looking at patterns in the data. Because
conceived partly depends on whether the research a quote can receive multiple codes, one can look at
is exploratory, in which case open coding is used, or the association between codes, or to what extent some
whether it wants to test a theory, in which case the issues are raised together. A possible disadvantage is
researcher predominantly uses preconceived codes. the start-up cost related to learning the software. The
As in the manual approach, the codes are reorganized use of software for the analysis may therefore be less
throughout the analysis as insights in the data shape attractive for stand-alone small-scale projects. Box
up. The next phase in the research is then to identify 10.9 provides a practical example of a software-based
patterns between the coded data. Software-based analysis.
142
Box 10.9 Recording, transcription, coding and analysis: an illustration from

a health worker study in Rwanda
Ten focus group discussions were conducted in the health worker study in Rwanda. They were digitally
recorded, generating one sound file per discussion. The files were transcribed, resulting in written
accounts of between 10 and 15 pages each. If the discussion had been in Kinyarwanda, it was translated
into French. The analysis was carried out using the qualitative data analysis software, QSR NVivo 2.0. Its
main advantage lies in ease of data coding and its functions enabling the visualization of different data
cross-sections. It does not enable, however, the production of an overview of the data in a matrix structure.
First, all transcripts were coded: a code was attributed to each piece of data (quote). A total of 35 codes
were used, reflecting the broad interest of the study. Examples of codes included vocation, rural versus
urban, health sector exit, job satisfaction, remuneration, absenteeism, corruption. A quote could receive
different codes if referring to different topics at the same time, thus allowing examination of associations
between quotes. In total, 1203 quotes have been examined.
To address the issues of reliability and validity of the an important theme in the discussion. More quotes do
resulting analysis, the following actions can be taken: not necessarily convince the reader more. In the final
t As qualitative research is essentially concerned with report key quotes are either embedded in the text, or
the viewpoints of the participants, the validity of the put in boxes separate from the main text.
study results can be increased by presenting the
results to the study participants and integrating their Another important issue in reporting qualitative research
feedback into the research output. is to dedicate some space to the description of the
t Writing up the underlying study assumptions and techniques used. Being transparent and explicit about
contextual information helps readers and potential the study method by including scripts and participant
users to assess the extent to which the research selection sheets, flagging problems with implementa-
results may be comparable to another context. tion and describing how the analysis was carried out
t Internal reliability – or the degree to which other all increase the readers’ confidence in the results; they
researchers would come to the same conclusions also make a replication of the study in another context
– can be enhanced by including an audit that exam- possible.
ines the data collection and analysis procedures and
identifies potential sources of bias. A more intensive
approach is to have the same qualitative data coded 10.4 Summary and conclusions
by different researchers, especially in the case of
free coding, and compare the obtained results. This chapter has argued that qualitative techniques
can play an important role in improving our under-
standing of health workforce challenges and policies.
10.3.5 Reporting Qualitative techniques are particularly well suited to
Because qualitative research deals with data in the building an understanding of a complex environment
form of text, it can be a challenge to present the results in order to generate and fine-tune hypotheses (that
in a way that is easy to digest. It is tempting to provide can be tested by quantitative research), and to identify
too many quotes in the final report and the golden rule and address measurement issues. These strengths are
is therefore for the researcher to adopt only a small pro- reflected in a growing body of qualitative research on
portion of the quotes he or she would like to include. health workers. Many studies combine different quali-
What may help is to compile a “summary of quotes” tative methods: the most commonly used approaches
document after the analysis but before reporting. This are individual interviews and group discussions, while
can contain up to 10 quotes per theme, and may help a third approach, participant observation, is an option
to select the most appropriate quotes to be used in but tends to be used less frequently. Many also serve
the final report and presentation. Quotes are typically as a complement to a quantitative study, either to better
selected for their salience and because they reflect understand puzzles thrown up by quantitative research
or in preparation of a quantitative survey.
143
The implementation of qualitative research is fraught References

with a number of methodological and practical chal-
1. Becker HS. The epistemology of qualitative
lenges. An important aim of the chapter has therefore research. In: Jessor R, Colby A, Shweder RA, eds.
been to provide practical guidance on how to design Ethnography and human development: context and
and implement qualitative work. An important mes- meaning in social inquiry. Chicago, University of
sage in this regard is that obtaining high-quality data Chicago Press, 1996.
requires careful preparation and implementation of the 2. Bryman A. Social research methods. Oxford, Oxford
study. Even though qualitative methods have inherent University Press, 2001.
weaknesses – most importantly the weak objectivity 3. Bryman A, Burgess RG, eds. Analyzing qualitative
following from the involvement of the researcher in the data. New York, Routledge, 1994.
production and interpretation of the data – these can to 4. Denzin NK, Lincoln YS, eds. The SAGE handbook
a significant degree be contained in the design of the of qualitative research, 3rd ed. London, Sage
study. Reporting the applied method in a transparent Publications, 2005.
way (for example including scripts, participant selec- 5. Flick U, ed. The SAGE qualitative research kit.
tion and method of analysis) increases the credibility Thousand Oaks, CA, Sage Publications, 2007.
of study results, and is something that is still neglected 6. Greene JC. Qualitative program evaluation: practice
in many studies; it also allows replication of the stud- and promise. In: Denzin NK, Lincoln YS, eds.
ies in a different context. A second message is not to Handbook of qualitative research. Thousand Oaks,
consider qualitative research in isolation. Combining CA, Sage Publications, 1994.
different methodologies strengthens the validity of the 7. Seale C. Quality in qualitative research. Qualitative
study results, and the high degree of complementarity Inquiry, 1999, 5(4):465–478.
between qualitative and quantitative methods calls for 8. Silverman D. Doing qualitative research: a practical
combining the two approaches. handbook, 2nd ed. London, Sage Publications,
2005.
The value of future qualitative research on human 9. Silverman D. Interpreting qualitative data, 3rd ed.
resources for health will be judged by how much it London, Sage Publications, 2006.
helps to redirect policies. For example, human resource 10. Walker R. Applied qualitative research. Aldershot,
policies in the health sector have typically emphasized Gower Publishing Company, 1985.
a labour planning approach – relying on the assump- 11. Wolcott HF. Writing up qualitative research, 3rd ed.
tion that health workers are competent, motivated to Thousand Oaks, CA, Sage Publications, 2008.
serve the public, but passive in their career choices. 12. Holland J, Campbell J. Methods in development
Currently, health workers are increasingly considered, research: combining qualitative and quantitative
like other professionals, as purposely reacting to incen- approaches. Bourton Hall, IDTG Publishing, 2005.
tives from their environment, and trying to balance their 13. Denscombe M. The good research guide: for small-
own happiness and that of their family with a concern for scale social research projects. Buckingham, Open
patients. Research on health workers therefore increas- University Press, 1998.
ingly incorporates a labour market approach, using 14. Tesch R. Qualitative research: analysis types and
more concepts and ideas from labour and personnel software tools. New York, Falmer Press, 1990.
economics, as illustrated by recent studies on job sat- 15. Hammersley M. The relationship between qualitative
isfaction (51), women’s labour market participation (52) and quantitative research: paradigm loyalty versus
and organizational design and decision-making (53), methodological eclecticism. In: Richardson JTE, ed.
Handbook of research methods for psychology and
and the growing interest in the economics discipline
the social sciences. Leicester, BPS Books, 1996.
in use of qualitative methods (see for example Bewley
(54)). Given the limited knowledge on health worker 16. Bryman A. Quantitative and qualitative research
strategies in knowing the social world. In: May
labour markets, and given their complex institutional T, Williams M, eds. Knowing the social world.
setting, qualitative research techniques are bound Buckingham, Open University Press, 1998.
to be predominantly used in the early stages of this
17. Halfpenny P. The analysis of qualitative data.
new avenue in health workforce research. Qualitative Sociological Review, 1979, 27(4):799–825.
research that is sustained by careful design and robust
18. Morgan DL. Practical strategies for combining
methods will be a powerful tool to support the devel- qualitative and quantitative methods: applications for
opment of these and other strands of innovative health health research. Qualitative Health Research, 1998,
workforce research. 8(3):362–376.
19. Q-squared: combining qualitative and quantitative
approaches in poverty analysis. Centre for
International Studies (http://www.q-squared.ca/,
144
20. Jaffré Y, Olivier de Sardan JP. Une médecine 34. Robinson V, Carey M. Peopling skilled international
inhospitalière: les difficiles relations entre soignants migration: Indian doctors in the UK. International
et soignés dans cinq capitales d’Afrique de l’ouest. Migration, 2000, 38(1):89–108.
Paris, Karthala, 2003.
35. Hardill I, MacDonald S. Skilled international
21. McPake B et al. Informal economic activities of migration: the experience of nurses in the UK.
public health workers in Uganda: implications for Regional Studies, 2000, 34(7):681–692.
quality and accessibility of care. Social Science and
36. Troy PH, Wyness LA, McAuliffe E. Nurses’
Medicine, 1999, 49(7):849–865.
experiences of recruitment and migration from
22. Ferrinho P et al. Pilfering for survival: how health developing countries: a phenomenological
workers use access to drugs as a coping strategy. approach. Human Resources for Health, 2007,
Human Resources for Health, 2004, 2:4 (http:// 5:15 (http://www.human-resources-health.com/
www.human-resources-health.com/content/2/1/4, content/5/1/15, accessed 25 January 2009).
37. Jewkes R, Abrahams N, Mvo Z. Why do nurses
23. Kyaddondo D, Whyte SR. Working in a decentralized abuse patients? Reflections from South African
system: a threat to health workers’ respect and obstetric services. Social Sciences and Medicine,
survival in Uganda. International Journal of Health 1998, 47(11):1781–1795.
Planning and Management, 2003, 18(4):329–342.
38. Asenso-Okyere WK et al. The behaviour of health
24. Israr SM et al. Coping strategies of health personnel workers in an area of cost sharing: Ghana’s drug
during economic crisis: a case study from cash and carry system. Tropical Medicine and
Cameroon. Tropical Medicine and International International Health, 1999, 4(8):586–593.
Health, 2000, 5(4):288–292.
39. Manongi RN, Marchant TC, Bygbjerg IC. Improving
25. Gruen R et al. Dual job holding practitioners in motivation among primary health care workers in
Bangladesh: an exploration. Social Science and Tanzania: a health worker perspective. Human
Medicine, 2002, 54(2):267–279. Resources for Health, 2006, 4:6 (http://www.human-
resources-health.com/content/4/1/6, accessed 25
26. Ferrinho P et al. How and why public sector doctors
January 2009).
engage in private practice in Portuguese-speaking
African countries. Health Policy and Planning, 1998, 40. Lindelow M, Serneels P. The performance of
13(3):332–338 (http://heapol.oxfordjournals.org/cgi/ health workers in Ethiopia: results from qualitative
reprint/13/3/332, accessed 25 January 2009). research. Social Science and Medicine, 2006,
62(9):2225–2235.
27. Jumpa M, Jan S, Mills A. The role of regulation in
influencing income-generating activities among 41. Lievens T, Serneels P. Synthesis of focus group
public sector doctors in Peru. Human Resources discussions with health workers in Rwanda. World
for Health, 2007, 5:5 (http://www.human-resources- Bank Research Report. Washington, DC, World
health.com/content/5/1/5, accessed 25 January Bank and Rwanda Ministry of Health, 2006.
2009).
42. Serneels P, Lindelow M, Lievens T. Qualitative
28. Belli P, Gotsadze G, Shahriari H. Out-of-pocket and research to prepare quantitative analysis:
informal payments in health sector: evidence from absenteeism among health workers in two African
Georgia. Health Policy, 2004, 70(1):109–123. countries. In: Amin S, Das J, Goldstein M, eds.
Are you being served? New tools for measuring
29. Franco LM et al. Determinants and consequences
service delivery. Washington, DC, World Bank, 2008
of health worker motivation in hospitals in Jordan
(http://go.worldbank.org/F6KIIC0700, accessed 25
and Georgia. Social Science and Medicine, 2004,
January 2009).
58(2):343–355.
43. Frechtling J, Sharp Westat L, eds. User-friendly
30. Dieleman M et al. Identifying factors for job
handbook for mixed method evaluations. Arlington,
motivation of rural health workers in north Viet Nam.
VA, National Science Foundation, 1997 (http://www.
Human Resources for Health, 2003, 1:10 (http://
nsf.gov/pubs/1997/nsf97153/start.htm, accessed 25
January 2009).
44. Atkinson R. The life story interview. Qualitative
31. Mathauer I, Imhoff I. Health worker motivation in
Research Methods Series, volume 44. London, Sage
Africa: the role of non-financial incentives and
Publications, 1998.
human resource management tools. Human
Resources for Health, 2006, 4:24 (http://www. 45. Kitzinger J. Understanding AIDS: researching
human-resources-health.com/content/4/1/24, audience perceptions of acquired immune
accessed 25 January 2009). deficiency syndrome. In: Eldridge J, ed. Getting
the message: news, truth and power. London,
32. Brewer CS et al. Nursing recruitment and retention
Routledge, 1993.
in New York State: a qualitative workforce needs
assessment. Policy, Politics and Nursing Practice, 46. Guba EG, Lincoln YS. Competing paradigms in
2006, 7(1):54–63. qualitative research. In: Denzin NK, Lincoln YS, eds.
Handbook of qualitative research. Thousand Oaks,
33. Muula AS, Maseko FC. How are health professionals
CA, Sage Publications, 1994.
earning their living in Malawi? BMC Health Services
Research, 2006, 6:97 (http://www.biomedcentral.
com/1472–6963/6/97, accessed 25 January 2009).
145
47. LeCompte MD, Goetz JP. Problems of reliability

and validity in ethnographic research. Review of
Educational Research, 1982, 52(1):31–60.
48. Lincoln YS, Guba E. Naturalistic inquiry. Beverly
Hills, CA, Sage Publications, 1985.
49. Mason J. Qualitative researching, 2nd ed. London,
Sage Publications, 2002.
50. Hurley RE. Qualitative research and the profound
grasp of the obvious. Health Services Research,
1999, 34(5 part 2):1119–1136 (http://www.
pubmedcentral.nih.gov/picrender.fcgi?artid=108905
6&blobtype=pdf, accessed 25 January 2009).
51. Bitsch V, Hogberg M. Exploring horticultural
employees’ attitudes toward their jobs: a qualitative
analysis based on Herzberg’s theory of job
satisfaction. Journal of Agricultural and Applied
Economics, December 2005.
52. Kwagala B. Integrating women’s reproductive roles
with productive activities in commerce: the case
of businesswomen in Kampala, Uganda. Urban
Studies, 1999, 36(9):1535–1550.
53. Friebel G, Panova E, Shaw K. Insider privatization
and careers: a study of a Russian firm in transition.
In: Bender S et al., eds. The analysis of firms and
employees: quantitative and qualitative approaches.
Chicago, University of Chicago Press, 2008.
54. Bewley TF. Why wages don’t fall during a recession.
Cambridge, MA, Harvard University Press, 1999.
146
Analysis and synthesis of information
11 on human resources for health

from multiple sources: selected
case studies
GUSTAVO NIGENDA, JOSÉ ARTURO RUIZ, CHRISTOPHER H HERBST,
AARUSHI BHATNAGAR, KRISHNA D RAO, INDRANI SARAN, SHOMIKHO RAHA
11.1 Introduction and programme decision-making when there are multi-

ple data sources (as is often the case in HRH analysis).
Having timely and reliable measurements of the health Whereas single research studies seek to maximize
workforce is becoming increasingly important to deci- scientific rigour through internal validity, triangulation
sion-makers, programme managers, development seeks to make the best public health decisions based
partners and other stakeholders, as a cornerstone to on the available information. The optimal use of pre-
monitoring and evaluation of health systems perform- existing data sources by means of triangulation allows
ance. Yet monitoring human resources for health (HRH) for a rapid understanding of the situation and facilitates
can be challenging and complex. Variances in the esti- timely, evidence-based decisions.
mated stock, skills mix and other characteristics of
HRH may occur, depending on the data source and the In this chapter, three case studies are presented on the
adjustment method used to correct for data deficien- uses of triangulation for HRH analysis with illustrative
cies (if any). Ideally, the data should be derived from applications in selected low- and middle-income coun-
a comprehensive, harmonized health information sys- tries. Both the data and methodological dimensions of
tem of all persons trained or working in the promotion, approaches to triangulation are discussed. Case stud-
protection or improvement of population health, along ies from Mexico, India and Zambia are used to review
with nature of skills obtained or used. However, the the potential for extracting relevant data for measur-
data sources used to populate such systems tend to ing a specific indicator of HRH metrics from different
be fragmented and incomplete, which can lead to bias information sources, examine how triangulation can be
in any measurements derived from them. Unfortunately, used to gauge the reliability of the available information
the methodologies for measuring health workforce and investigate how new information can be teased out
dynamics lag far behind the demand for information by triangulating data across different types of sources.
and evidence.
As reviewed in Chapter 3 of this Handbook, a number

of sources can potentially produce data relevant for
11.2 Identification of potential
HRH analysis, including population and establish- information sources and their
ment censuses and surveys, administrative records use to estimate indicators
and qualitative studies. Using a combination of differ- of health labour market
ent sources can provide better-quality measurements
of health workforce characteristics and give a much
participation in Mexico
more rounded picture (1). It is good practice to recon- The health workforce is a heterogeneous group, with
cile data from different sources rather than to rely on the activities of health workers ranging widely, depend-
only one source of information. Dependence on single ing on how they are incorporated into the labour market.
sources increases the risk of making decisions based Measuring health labour dynamics can be complex,
on statistics that are incomplete or biased (2). but there are also practical opportunities for generat-
ing precise measurements from standard statistical
A strategy of triangulation – or cross-examination and sources. In contrast with the general population, the
synthesis of the available data on a central theme across health workforce is generally a highly qualified group
different sources and using different methodologies – of workers for whom the level and field of education
can be used to assess and reconcile potential variances tends to be closely correlated with the nature of their
in coverage, classification and reporting of information. jobs. The development of evidence-based HRH pol-
This approach can be effective in supporting policy icies requires ongoing assessment of any potential
147
imbalances between formal education attained by for obtaining regular updates on the working status of
health workers and their current occupational status health personnel (5). The Mexican employment survey
and work activities (3). has been carried out quarterly since 1988, collecting
information on, for example, sociodemographic char-
In Mexico, different sources can be used to assess the acteristics of the population (including level and field of
number of health professionals and their working situa- education), labour force activity (participation, employ-
tion. They include the Population and Housing Census ment status, occupation, industrial sector) and working
(Censo General de Población y Vivienda) and the conditions (income, benefits, hours worked).
National Survey on Employment (Encuesta Nacional de
Empleo). Censuses are conducted every 10 years by The availability of questions in the census and employ-
the Central Statistical Office (4). The 2000 census cov- ment survey on education and labour activity enables
ered a number of topics related to education and labour an estimate to be made from both sources of the
market activities (Table 11.1). In addition, a midterm number of physicians, nurses and other health profes-
count gathers information from a representative sam- sional groups across the country, and their employment
ple of the population between censuses. However, the status. Among those currently employed, it is possible
latest (2005) count did not include questions on field to assess the type of activity they perform and whether
of education or labour activity, and so did not produce it is related to the academic field in which they were
the required information for health workforce analy- educated. The available data also offer estimates on
sis. While the previous (1995) count collected some the number of people with a health-related education
relevant information, certain questionnaire wording who are not active in the formal labour force (for exam-
differences compared to the census hindered comple domestic labour, retired, poor health).
parability in measurement of some indicators across
these two sources. An important constraint was the ina- Such information is valuable for measuring labour
bility to track field of education over time. wastage of HRH, which refers to qualified human
resources who, though of eligible age for labour force
As in many countries, labour force surveys are carried participation, are in inadequate employment situations
out regularly in Mexico, representing a good option because they are not working or because they carry
Table 11.1 Questions on education and labour activity included in the national population census and
intercensal counts, Mexico, 1995–2005
Source Population and Housing XII Population and Housing Population and Housing
Indicator Count, 1995 Census, 2000 Count, 2005
Educational Highest year/grade of formal Highest year/grade of formal Highest year/grade of formal
attainment education completed by the education completed by the education completed by the
respondent. Choice of nine respondent. Choice of eight respondent. Choice of 10
response options. response options. response options.
Field of education n.a. Name of course of study (for n.a.
respondents with higher levels
of educational attainment).
Open question.
Labour activity Labour force participation Labour force participation n.a.
during the previous week. during the previous week.
Choice of nine response Choice of eight response
options. options.
Occupation Occupational or job title. Open Occupational or job title. Open n.a.
question. question.
Nature of work activities or Nature of work activities or n.a.
tasks. Open question. tasks. Open question.
Branch of economic Place of work by type. Choice Place of work by type. Open n.a.
activity of eight response options. question.
Main economic activity at place Main economic activity at place n.a.
of work. Open question. of work. Open question.
n.a. Not available (no information collected).
Source: Instituto Nacional de Estadística, Geografía e Informática (4).
148
Analysis and synthesis of information on human resources for health from multiple sources: selected case studies
out activities that do not correspond to their training (6). t Unemployed: individuals that actively place pres-
From a health systems strengthening perspective, fail- sure on the labour market in searching for a job or
ure of qualified persons to put into practice the specific waiting for the outcome of a job application (during
skills stemming from their education implies wastage of a referenced period of time, such as the week prior
a social, public or private investment that does not yield to data collection). The definition may also extend to
a benefit for the population, or for the workers them- those looking for work outside the formal economic
selves. Those trained in health services provision who sector.
remain outside it due to various reasons (inadequate t Inactive-eligible: individuals not currently seeking
market absorption capacity, personal motivation) fall employment but who would be eligible for work by
into various categories: virtue of their skills, age and ability. This encom-
t Underemployed: individuals that have completed passes those who have chosen to withdraw from the
their formal education and are currently employed labour market as a personal option (in the short or
but perform activities not related to their training. long term), including those dedicated to domestic
This includes trained health professionals working labour.
outside the health services, signifying a mismatch t Inactive-ineligible: individuals unavailable or una-
of occupational skills. The concept is consistent with ble to work, for example due to retirement, studies,
the International Labour Organization’s definition work-limiting disability or other reason.
of underemployment as encompassing individu-
als whose “employment is inadequate, in relation to Table 11.2 shows selected findings from two differ-
specified norms or alternative employment, account ent sources on indicators of health labour wastage in
being taken of [their] occupational skill” (7).1 Mexico (8). Microdata, or data collected from individual
respondents, are used from the last population census
1 An alternative to the skill-related approach to assessing and from the national employment survey round corre-
underemployment is in relation to a time criterion, such as in
the case of health professionals employed in health services sponding to the same period (third quarter of the year
having worked less than a threshold related to working time 2000). The indicators relate to the physician and nurs-
(for example less than what is considered as full-time work in ing workforces, calculated using the same definitions
a country).
Table 11.2 Stock and distribution of the physician and nursing workforce by
labour force status, based on census and survey data, Mexico, 2000
Data/Source Census Employment survey

Indicator
Number % Number %
Physicians
Employed 142 923 70 189 930 74
Underemployed 26 733 13 28 457 11
Unemployed 10 892 5 7 036 3
Domestic labour 7 895 4 14 556 6
Not available for work 16 335 8 17 509 7
Total 204 778 100 257 488 100
Nursing professionals
Employed 57 834 55 62 406 64
Underemployed 16 128 15 7 666 8
Unemployed 7 143 7 1 254 1
Domestic labour 16 971 16 19 530 20
Not available for work 6 659 6 7 124 7
Total 104 735 100 97 980 100
Note: Underemployed includes those who have completed university-level training in medicine or nursing
but perform work activities not related to their education. Not available for work includes those who are
retired, studying or have a work-limiting disability. Percentages may not sum to 100% due to rounding.
Source: Instituto Nacional de Estadística, Geografía e Informática (8).
149
from both data sources. Despite the similar methodol- calculating relevant indicators from the latter. Countries
ogy used, important differences are found in the total are encouraged to compile and publish the available
number of physicians (results from the survey being information from all sources even if at first glance the
some 20% higher than those from the census), but the data may seem of questionable quality. This could help
labour force indicators (in percentage terms) are gen- prospects, which might otherwise be lost, for develop-
erally consistent across both sources. ing techniques for evaluating and adjusting such data
and learning from experience.
Regarding nursing professionals, in contrast to the
case of physicians, estimates of the total stock are
lower from the survey source than from the census 11.3 Using multiple sources
(7% lower), and the indicators of labour force par-
ticipation show a greater asymmetry. For example,
of information to produce
the difference in the employment rate among nurses best estimates of India’s
is 9 percentage points across sources, and for the health workforce
underemployment and unemployment rates some 6
In most low- and middle-income countries, HRH esti-
percentage points each.
mates reported by ministries of health and allied
agencies need strengthening in terms of comprehen-
It is possible that differences in sampling, training of field
siveness, reliability and timeliness. In India, routine
enumerators and supervisors, coding of self-reports of
information on the national health workforce suffers
respondents’ occupation descriptions, data entry and
from several limitations. Reports from state professional
processing operations, or other methodological con-
regulatory bodies – which form the basis of official
siderations may lie behind observed discrepancies in
health workforce statistics – tend to be inaccurate as
the estimated indicators of labour force activity among
a result of non-adjustment for health workers leaving
health professionals from the two different data sources
the workforce due to death, migration or retirement,
in Mexico.
or of double-counting of workers registered in more
It is also clear from this case study that standard sta- than one state. Further, not all states follow the same
tistical sources can be a useful tool for HRH analysis in registering procedure, which raises issues of compara-
many countries. In the Mexican context, the available bility; for example, the Delhi Medical Council requires
data allow quantification of the essentially qualita- practising physicians to re-register every five years, a
tive concept of health labour wastage, which in turn process that is not followed in other states. Finally, cer-
assists in raising awareness among decision-makers tain categories of health workers (such as biomedical
of the magnitude of HRH challenges and guiding pol- researchers, physiotherapists and laboratory technol-
icy recommendations (6). Despite perceived barriers ogists) are not recorded in official statistics, thereby
to access to a skilled health worker among some seg- making it difficult to estimate the overall size and com-
ments of the population, the data demonstrate that there position of the health workforce (9, 10).
simultaneously exists within the country a large number
As previously discussed, a variety of data sources
of trained providers who are not working in health serv-
available in most countries can be used to provide
ices delivery. While the present analysis was limited to
useful information on HRH metrics. In this analysis,
physicians and nurses, it could certainly be extended
estimates of the numbers of health workers from both
to other health professional and technician groups, and
official sources and (population-based) standard sta-
eventually to other countries where the questionnaire
tistical sources are cross-examined:
wordings from such sources are similar.
t Official statistics on the numbers of registered phy-
Data evaluation, and possibly adjustment to compen- sicians and nurses were obtained from the medical
sate for data deficiencies, is crucial to validate HRH and nursing councils of India (based on compiled
information sources to support policy and planning. The reports from their state counterparts, covering both
cross-examination checks presented here are informa- the public and private sectors); other statistics on
tive, but do not provide a basis for formal evaluation HRH were drawn from various Ministry of Health
or adjustment of the reported numbers. Repeating this reports (11–13).
exercise across successive censuses or over multiple t Tallies of the health workforce for various cadres
countries is likely to lead to improvements in estima- were drawn from the 2001 Census of India, which
tion and evaluation procedures. In Mexico, options collected information on the self-reported occupa-
for time-trend analysis were limited due to differences tion of all individuals in the country (14). The present
in the data collection tools between the census and analysis uses a census microdata sample of 20%
intercensal counts, which limited the possibility of of rural and 50% of urban enumeration blocks in
all districts, representing a sample size of about
150
Figure 11.1 Density of the health workforce by cadre, according to data source, India 2005
19.5
TOTAL 19.7
1.4
Others 3.1
Census
0.5
Other traditional Employment survey
0.7
medicine practitioners
Ministry of Health
1.8
AYUSH practitioners 2.6 official statistics
6.5
2.2
Pharmacists 1.7
0.2
Dentists 0.2
0.5
7.4
Nursing & midwifery
7.1
personnel 12.8
6.1
Physicians & surgeons 4.3 Sources: Indian Nursing Council (11),
5.9 Medical Council of India (12), Central
Bureau of Health Intelligence (13), Cen-
0 5 10 15 20
sus of India 2001 (14), National Sample
Density of health workers (per 10 000 population) Survey Organization 2004–2005 (15).
300 million, and weighted to represent the total number of cases where occupational information was
population. missing, some employed individuals were recognized
t Estimates of the health workforce for various cad- as health workers based on their branch of economic
res were also obtained from the 2004–2005 activity or their educational qualifications.
round of a quinquennial survey on employment
and unemployment, implemented by the National The final set of health worker categories for which
Sample Survey Organisation (15). The nationally estimates were produced comprised physicians and
representative household survey recorded the prin- surgeons (allopathic medical practitioners), nurs-
cipal self-reported economic activity of employed ing and midwifery personnel, dentists, pharmacists,
individuals. AYUSH practitioners (ayurvedic, yoga, unani, sidha,
homeopathy), other traditional medicine practitioners,
Certain adjustments were made to make the employ- and others (dieticians, opticians, dental assistants,
ment survey and census estimates comparable. While physiotherapists, medial assistants and technicians,
the census estimates were recorded in March 2001, other hospital staff).
the survey was conducted between July 2004 and
June 2005. On the assumption that growth in the health As seen in Figure 11.1, official Ministry of Health esti-
workforce follows that of the general population, the mates are generally higher than those of the census
census estimates were inflated by 8% to reflect demo- and survey sources and, significantly, there is better
graphic growth between 2001 and 2005. agreement between the latter two. Measures were most
similar for the physician and dentist workforces across
Health workers were identified in the census and all three sources. In total, the census and employment
employment survey samples using the National survey estimates suggest that there are some 2.1 mil-
Classification of Occupations (16). The census used lion health workers in India, which translates into a
the latest 2004 version of the classification (also known density of approximately 20 health workers per 10 000
as NCO-04), whereas the survey used the earlier 1968 population. Comparative statistics were not available
codes (NCO-68). In the estimates presented here, on the total number of health workers from administra-
NCO-68 codes were converted to NCO-04 with little tive sources. Also excluded from these estimates are
loss of information. To further improve comparability the roughly 1.5 million community health workers, for
between the two, certain health worker categories were whom a distinct occupational code is not included in
either split or merged together. For example, because the current classification (although it is possible that
the function of nurses and midwives is often similar some of these workers are assimilated under nursing
they were merged into a single category. In a limited and midwifery personnel).
151
Figure 11.2 Density of the health workforce (per 10 000 population) based on census versus survey data,
by state, India 2005
60
50
Mizoram
Delhi Goa
Sikkim
Density based on census data
40 Kerala
Nagaland
Arunachal Pradesh
30 Himachal
Jammu & Kashmir Pradesh Maharashtra
Punjab
West Bengal Tamil Nadu
Manipur
20 Orissa Uttaranchal
Tripura Gujarat
Assam Andhra Pradesh
Jharkhand Karnataka
10 Bihar Meghalaya Haryana
Uttar Pradesh Rajasthan

Madhya
Pradesh Chhattisgarh
0
0 10 20 30 40 50 60 Sources: Census of India 2001 (14),
National Sample Survey Organization
Density based on employment survey data 2004–2005 (15).
Further analysis of the census and survey data reveals example, the census estimates have been shown to
considerable variation in the density of the health have good correspondence with those obtained from
workforce (all cadres combined) across the states in the survey source, at the aggregate level, indicating
India (Figure 11.2). For a majority of states, the cen- good reliability.
sus estimates are higher than those obtained from the
employment survey. The latter’s small sample size pre-
vents robust disaggregated estimates at the state level. 11.4 Triangulation of data from two
Estimates of the heath workforce drawing on census
different sources for monitoring
and employment survey data have several advantages. health worker absenteeism and
They are based on population counts and hence avoid ghost workers in Zambia
the problem of double-counting, cover a wide range
Health worker absenteeism fuels inefficiencies and
of health occupations, are available for all areas of the
inequities in many health labour markets in the devel-
country and have fewer comparability issues because
oping world. As with other dimensions of the health
they are based on standard occupational codes.
workforce crisis, relevant data and information are
However, these information sources also have certain
needed for governments and partners to be able to
limitations. Notably, they cannot provide health work-
address and monitor the problem. Often underused,
force estimates on an ongoing basis, as the Indian
health facility assessments offer an ideal mechanism
employment survey is repeated only once every five
for obtaining and producing information and evidence
years and the census every 10 years.
on various forms of worker absenteeism. Measuring
this indicator of labour dynamics requires the collec-
Overall, the census appears to be the best available
tion of only limited additional data, and the subsequent
source for health workforce estimates. The large size of
triangulation of the new facility-based data with infor-
the microdata sample used here – covering every dis-
mation from routine administrative registries (17).
trict in the country and, within each district, both urban
and rural areas – allows for robust estimates across
Health worker absenteeism comes in several forms. A
health worker categories and by state. In the Indian
distinction should be made between: (i) absenteeism
152
of health workers registered and generally present Figure 11.3 Percentage of facility-based health
at the facility but absent at a particular point in time workers on the duty roster but not accounted for on
(henceforth referred to as absenteeism); and (ii) health the day of the assessment, by cadre, Zambia 2006
workers absent from the workforce altogether, not
found on facility staffing lists, but nevertheless listed on TOTAL 23
official payroll records (henceforth referred to as ghost Other clinical personnel <1
workers). The former may include scheduled absence
Other technicians 21
for official duty or personal reasons (i.e. not currently
present due, for example, to part-time work status, Pharmacy technicians 8
attendance at training or meetings, secondment, vaca- Laboratory technicians 11

tion, maternity or parental leave), unscheduled absence
Radiography technicians 15
(for example sickness or other emergency circum-
Orthopaedist technicians 36
stance), dual employment (i.e. current work practice in
another location) or unauthorized or unexplained leave. Pharmacists 17
The latter includes exit from the facility-based work- Environmental
13
health workers
force (but not from the payroll) due to death, long-term
Hygienists <1
illness or even fraud. In some contexts, ghost workers
reflect a strategy among health personnel to overcome Physiotherapists 30
unsatisfactory remuneration or working conditions. Nutritionists 21
Medical assistants 17
The fundamental requirement for obtaining reliable
information on absenteeism on ghost workers is the Nurses 23
collection of primary data by means of facility-based Physicians 45

assessments, specifically on health workers registered
0 10 20 30 40 50
at the facility level (i.e. on staff inventory lists), those
meant to be working at the time of data collection (on Source: Herbst and Gijsbrechts (17).
duty rosters) and those actually present at the time of
data collection (those headcounted). Timely data from
administrative sources are also required on health
Figure 11.4 Percentage of health workers
workers listed on (public or private) payroll records. The
recorded on the Ministry of Health payroll but not
respective indicators are calculated as follows: registered at the facility level, Zambia 2006
t absenteeism = facility-based health workers reported
on duty rosters minus those actually headcounted;
t ghost workers = number of health workers listed on TOTAL (average) 28
central payroll records minus those registered at the
facility level.
Health management &
28
For basic snapshots on absenteeism and ghost work- support workers
ers, the minimal data requirements on facility-based
workers for all three components (inventory lists, duty
rosters and headcounts) are name, cadre, unique iden- Dentistry cadres 34
tification number and salary source. Without knowledge

of name and cadre, absentees cannot be identified.
Medical, nursing & other
Without the identification number, double-counting of 19
health cadres
workers having jobs at more than one facility cannot
be prevented. Salary source is crucial for triangulating
0 10 20 30 40
facility data with administrative payroll data to estimate
the extent of ghost workers in the health system. At the % on payroll not captured by the survey
level of the facility, information is also needed on own-
Source: Herbst and Gijsbrechts (17).
ership, or operating authority of the establishment.
An enumeration of health facilities in Zambia obtained

data on health workers who were listed to be on duty at (Figure 11.3) (see also Chapter 7 for an illustrative
the time of the facility visit (i.e. on the duty roster) and example from Kenya). Findings on the estimated pro-
those actually present (i.e. headcounted or accounted portion of ghost workers in health facilities in Zambia
for), producing a snapshot in time on absenteeism can be seen in Figure 11.4.
153
Despite efforts to adjust the facility-level data against 11.5 Summary and conclusions
payroll data for enhanced comparability, non-negligi-
ble information gaps on salary source of health workers This chapter has presented selected case studies on
may have compromised measurement accuracy (17). approaches to triangulation, or cross-examination of
Computation of indicators on ghost workers requires multiple information sources for in-depth HRH analy-
triangulation of data on health workers by salary source sis. Triangulation represents a cost-effective strategy
– as opposed to sector of employment – with payroll for using diverse datasets to explore a single issue.
records. Although in Zambia the vast majority of health In monitoring and evaluation of health workforce pol-
workers found in public sector facilities are indeed on icies and programmes, using one source or method
the public sector payroll (an assumption made during is rarely sufficient. Frameworks have been formulated
analysis), this is not necessarily typical. that can support efforts to coherently combine statis-
tics from different sources and for different units (18).
The production of useful information on health worker Different kinds of data, methods and approaches will
absenteeism (at a given moment in time) and ghost often yield somewhat different results. Exploring these
workers is only possible if the underlying data from inconsistencies can be an important means of better
which this is derived are not only adequately collected understanding the multifaceted and complex nature of
and triangulated (using well-designed tools and ana- health labour market dynamics. Areas where the data
lytical techniques), but also accurately reflect the triangulation produces similar or converging results will
day-to-day scenario on the ground. To minimize poten- increase confidence in preliminary findings.
tial biases, it is important to take into account certain
considerations, notably the timing of data collection at The technique can also help articulate recommenda-
the facility level: snapshots of absenteeism may vary tions for strengthening future evaluation efforts. This
depending on whether data collection occurs at the chapter has highlighted the strengths and limitations
beginning or end of a workday, as levels of absentee- of using some combination of population censuses,
ism are often higher in the afternoons. Announcing employment surveys, health facility assessments
data collection in advance to those with managerial and routine administrative records for HRH assess-
decision-making authority may also skew the picture ment. Ensuring a strategic agenda for data collection,
on absenteeism in some scenarios. Facility managers processing, analysis and use – including common (and
may either want to ensure staff presence since high detailed) occupational classification, unique identifiers
levels of absenteeism could be seen to reflect badly for practising health workers, ongoing update of admin-
on managerial performance – or, conversely, they may istrative registries to account for worker absences and
want to expose absentees as evidence to support attrition, consistency over time in the design of data
future human resources planning and negotiations. collection instruments, and systematic dissemina-
tion of findings – opens up new possibilities for HRH
It is not certain which combination of reasons for analysis to support timely recommendations for evi-
observed worker absences at a moment in time dence-based decision-making.
explains the results from the data triangulation exer-
cise for Zambia. The method entailed comparisons of
staff present, listed or on the payroll on the day of the
survey. However, the facility assessment did not collect
detailed information that would help improve under-
standing of the dynamics of absenteeism, such as
duration of absence or its underlying basis (for exam-
ple, expected versus non-expected). Improving routine
HRH management information systems, including daily
updates of staffing schedules at the facility level, would
allow better elucidation of staff behaviours and work-
force efficiencies on an ongoing basis.
154
References 14. Census of India 2001. Office of the Registrar

General and Census Commissioner of India (http://
1. Galin P. Guía para optimizar la utilización de www.censusindia.net, accessed 27 January 2009).
la información disponible sobre empleo en el
sector salud. In: Dal Poz M et al., eds. Relaciones 15. Government of India, Ministry of Statistics and
laborales en el sector salud. Quito, Organización Programme Implementation, National Sample Survey
Panamericana de la Salud, 2000. Organisation (http://mospi.nic.in/mospi_about_nsso.
htm, accessed 27 January 2009).
2. AbouZahr C, Adjei S, Kanchanachitra C. From data
to policy: good practices and cautionary tales. 16. National Classification of Occupations 2004.
Lancet, 2007, 369:1039–1046. Government of India, Ministry of Labour, 2004 (http://
dget.nic.in/nco, accessed 27 January 2009).
3. Nigenda G et al. University-trained nurses in
Mexico: an assessment of educational attrition 17. Herbst CH, Gijsbrechts D. Information on stock,
and labor wastage. Salud Pública de México, profiles, and distribution of health workers in Zambia:
2006, 48(1):22–29 (http://www.scielosp.org/ analysis of the Health Facility Census data. World
scielo.php?script=sci_arttext&pid=S0036– Bank Human Resources for Health Program Paper.
36342006000100005&lng=en&nrm=iso, accessed Washington, DC, World Bank, 2007.
25 January 2009). 18. Hoffmann E. Developing labour account estimates:
4. Instituto Nacional de Estadística, Geografía e issues and approaches. In: Household accounting:
Informática (INEGI) (http://www.inegi.gob.mx, experiences in concepts and compilation. Studies in
accessed 25 January 2009). Methods, Series F, No. 75. New York, United Nations
Statistics Division, 2000.
5. Gupta N et al. Assessing human resources for
health: what can be learned from labour force
surveys? Human Resources for Health, 2003,
1:5 (http://www.human-resources-health.com/
6. Nigenda G, Ruiz JA, Bejarano R. Educational and
labor wastage of doctors in Mexico: towards the
construction of a common methodology. Human
Resources for Health, 2005, 3:3 (http://www.human-
resources-health.com/content/3/1/3, accessed 26
January 2009).
7. Hussmanns R, Mehran F, Verma V. Surveys of
economically active population, employment,
unemployment and underemployment: an ILO
manual on concepts and methods. Geneva,
International Labour Office, 1990.
8. XII Population and Household Census and National
Survey on Employment. Mexico, Instituto Nacional
de Estadística, Geografía e Informática, 2000.
9. Financing and delivery of health care services
in India. Background paper of the National
Commission on Macroeconomics and Health.
New Delhi, Government of India, Ministry of Health
and Family Welfare, 2005 (http://www.who.int/
macrohealth/action/national_reports/en/index.html,
10. Not enough here ... too many there: health workforce
in India. New Delhi, World Health Organization
Country Office for India, 2007 (http://www.whoindia.
org/EN/Section2/Section404_1264.htm, accessed
27 January 2009).
11. Indian Nursing Council (http://www.
indiannursingcouncil.org, accessed 27 January
2009).
12. Medical Council of India (http://www.mciindia.org,
13. Medical health and nursing manpower statistics
2005. New Delhi, Government of India, Ministry
of Health and Family Welfare, Central Bureau of
Health Intelligence (http://cbhidghs.nic.in/index3.
asp?sslid=929&subsublinkid=828, accessed 27
January 2009).
155
156
Part IV:
DATA DISSEMINATION
AND USE
157
Getting information and evidence into
12 policy-making and practice: strategies

and mechanisms
GÜLIN GEDIK, ULYSSES PANISSET, MARIO R DAL POZ, FELIX RIGOLI
12.1 Introduction implement, monitor and evaluate strategic action. In

addition, the policy imperatives of global and regional
In a rapidly changing world, the need to address huge initiatives and other strategies to strengthen health sys-
deficits in the human resources for health (HRH) devel- tems in countries (such as those described in Chapter
opment agenda has sparked growing interest and 1) call for robust monitoring and evaluation mech-
concern at global, regional and country level. In order anisms to assess the extent to which countries are
to act rapidly, effectively and efficiently to address these making progress or lagging behind. To make well-
deficits, policy- and decision-makers require updated informed decisions about HRH, decision-makers and
and readily available information and evidence based other stakeholders, including the private sector and
on solid data. civil society, need the best available evidence about
what works or looks promising.
A major difficulty experienced by decision-makers
in addressing HRH challenges is the complex diver- The significant challenge of scaling up and improv-
sity and simultaneity of pressing key issues such as ing performance of HRH and health services delivery
imbalances in global and national distribution and in a relatively short time, and integrating and coordi-
production of health workers, progressive increase nating decision-making in health systems with other
in international migration, sociopolitical restrictions sectors of government and society, requires new ways
on public sector operations (including freeze on pub- of thinking and organizing policy-making. There have
lic sector recruitment due to structural adjustment been several studies in the past decade examining the
programmes), the need to scale up priority health inter- processes, which have attempted to address the gap
ventions, impacts of HIV on the health workforce and between what is known about effective health systems
addressing worker motivation and retention. Many of interventions and the evidence used in decision-mak-
these challenges are long-term processes with cumu- ing. The issue of how to access evidence synthesis and
lative effects on workforce shortages and imbalances. adapt it for application to the local context is a prob-
Resulting human resources problems can manifest in lem common to all countries at different levels of social
different ways, such as labour disruptions (for example and economic development, and needs addressing
strikes) because of low morale among health workers, through international cooperation (1).
or even hospital closure due to inability to meet mini-
mum staffing norms. The possible solutions can only Senior policy-makers in low- and middle-income
be identified through examination of the roots of these countries have stressed that access to high-quality
problems, though these may not be apparent at first information and evidence that is timely and relevant
glance. is critical to their ability to make evidence-informed
decisions. Although necessary, evidence alone is cer-
Countries have made efforts to develop HRH policies tainly not sufficient, given that health interventions are
and implementation plans within the context of health affected by a wide range of contextual, political, cultural
sector reforms with varied levels of success. In gen- and sociological factors that shape the decision-mak-
eral, policies may be elaborated and implemented ing process (2). There is a significant knowledge gap in
with several inputs and factors, such as political and our understanding of the mechanisms, structures and
financial constraints, tacit knowledge and professional factors that policy-makers experience. To overcome
experience. In the absence of reliable and validated this gap, identification of these barriers and facilitat-
data, it has been difficult to establish evidence-based ing factors in each context is required, through better
policies, that is, policies informed through the use of research on how evidence-informed policy-making
health indicators and research results to formulate,
159
can integrate context, values and politics in pursuit of identified as one of the main obstacles for effective
improved health of populations (3). knowledge transfer and the use of research in deci-
sion-making processes. One of the fundamental factors
The debate over evidence in public health has often identified in approaches that recognize the complexity
focused on “the linear use of research evidence in a of health systems policy-making and action is the need
programmatic rather than policy context” – and has not to improve interaction between different stakeholders
taken account of the capacity to implement the pol- in the decision-making process (9). This is essential
icy (4). The capacity to take into consideration political to produce feasible policies that can be implemented
factors and act on evidence is fundamental to the appli- according to a specific context. Such a stakeholder-
cation of an evidence-informed policy, but is lacking oriented model has been called an “interactive model”
in most initiatives to date. Active participation of pol- (4) – the ongoing process to “adopt, adapt, and act” on
icy-makers and other stakeholders in the identification knowledge, using research to inform policy while simul-
of problems and systematization of evidence is vital, taneously weighting the politics of different interests,
but rare (5). A more integrated policy process – where motivations and values.
agenda setting, policy formulation, implementation and
feedback are closely related to each other – is required A systematic review of studies evaluating policy-mak-
to make sure that the use of the best scientific evidence ers’ perceptions of their use of evidence identified the
is effectively implemented in different contexts (6). following facilitating factors: (i) frequent two-way per-
sonal contact between researchers and policy-makers;
Earlier chapters of this Handbook highlighted what (ii) an estimation that the research produced is timely,
needs to be monitored, what types of information are relevant and of sound quality; and (iii) the presentation
required, what are the possible sources of data and of results including summaries with clear policy rec-
information, and how they can be collected and proc- ommendations. The most commonly identified barriers
essed. This chapter looks at how this information and were absence of interaction, timeliness, relevance and
research results can be collated, disseminated, shared credibility; lack of translation to the user; and mutual
and used to facilitate policy dialogue and implemen- distrust between researchers and policy-makers (10).
tation. The following sections take stock of various
initiatives in place to build effective health workforce Utilization of research evidence requires active promo-
information and evidence generation and dissemination of these facilitating factors and overcoming barriers
tion mechanisms at subnational, national, regional and in innovative ways to take account of the complexity of
international levels, and how to improve the use of sci- the policy-making process. The creation of an enabling,
entific evidence in policy formulation and monitoring interactive, learning environment with different stake-
and evaluation of its implementation. holders is critical (1, 11, 12). Three main strategies have
been proposed to promote the application of informa-
tion and evidence to policy- and decision-making:
12.2 Strategies to get evidence t strategies to promote researcher “push”, concen-
trating on diffusion to a broad audience – including
into policy and practice dissemination of findings through concise policy
The term “knowledge translation” (or wording alter- briefs and syntheses of existing evidence;
natives that basically encompass the same concept, t strategies to promote policy-maker “pull”, focused
such as “capacity building”, “knowledge management” on the needs of users, thereby creating an appetite
and “linkage and exchange”) is increasingly used to for research results – including capacity develop-
describe a series of activities undertaken to generate ment for policy-makers in commissioning research
knowledge targeting user needs, to disseminate this and development of rapid response mechanisms;
knowledge, to build decision-makers’ capacity for its t strategies to promote linkage and exchange, or
uptake and to adapt and track its application in specific building and maintaining relationships in order to
contexts (7). In the field of health, it has been described exchange knowledge and ideas – including joint
as a dynamic and iterative process that includes syn- mechanisms to set priorities for health and health
thesis, dissemination, exchange and ethically sound workforce policies, and forums and workshops for
application of knowledge to provide more effective exchange between researchers, policy-makers and
health services and products, strengthen health sys- civil society representatives regarding the evidence
tems and improve population health (8). base and its application to specific policies (7, 13).
The traditional separation between researchers, pol- Policy- and decision-making require research results
icy-makers, research funders, programme managers, focused on problem solving and addressing demand
health practitioners and health service users has been for specific policy issues. Thus, the definition of the
160
Getting information and evidence into policy-making and practice: strategies and mechanisms
research question based on a problem identified by of the actors involved in this process, they must also
the decision-maker (pull) facilitates the research-to- develop a stakeholder-oriented work environment (21).
policy link. Sponsors and financers of health research Interactive approaches, ones that create and strengthen
are increasingly realizing that their calls for proposals to a learning environment with constant exchange or link-
study the intricacies of health systems must rely on the ages of ideas and experiences among a varied set of
demand of policy- and decision-makers, as opposed stakeholders, are essential for promoting the appropri-
to the more traditional researcher-oriented definition ate use of research evidence (22). The process may be
of research projects. Ensuring sustained interactions driven at the beginning by decision-makers and other
between researchers and policy-makers appears to users (for example health services practitioners), with
make a difference in supporting evidence-informed researchers acting as resources (23). Simultaneously,
decision-making processes (14). it may facilitate the implementation of the policy, as
inputs from different users will improve the capacity to
Besides the need for direct interaction with policy- identify factors enabling implementation (Figure 12.1)
makers, HRH research also entails an interdisciplinary (24).
approach to tackle the multiple aspects related to
human resources in health systems. This poses a chal- Thus, it is critical to identify, promote, establish and
lenge on developing innovative methodologies and strengthen effective mechanisms for supporting the
networking, integrating mechanisms that bring to HRH development, implementation and monitoring and
research the expertise in financing, legislation and evaluation of HRH policies. The mechanisms should
regulation, health systems management, behavioural ensure not only the collection of data or generation of
sciences and other key areas for the better under- evidence, but also dissemination and utilization for pol-
standing of HRH dynamics. icy and managerial decisions. Viable HRH policies can
be developed, committed and implemented if they are
An additional aspect of HRH research is that most of based on information and evidence, principles and val-
the available studies focus on developed countries ues, and respond to the various stakeholders’ interests.
(15). It has been estimated that only 5% of published Any policy dialogue, therefore, needs to use mecha-
articles on health policy and systems research focus on nisms that ensure stakeholders’ participation is taken
low- and middle-income countries (16). Researchers into account, but modulated with the inflow of infor-
looking into HRH issues for low- and middle-income mation. Health workforce policies formulated only by
countries frequently run into difficulties with availability stakeholders’ consensus risk being interest driven (for
of quality (quantitative and qualitative) data and defini- example, doctors creating the policy for the medical
tion of appropriate performance indicators required as workforce); conversely, policies made behind the doors
raw material on which research can be shaped (17). of a government cabinet based purely on principles or
values risk being unrealistic or dogmatic (for example,
The issue of context specificity between and within forcing health professionals to work in deprived areas
countries also constitutes a major challenge. People without thinking about incentives for their performance
working towards strengthening of the health system
in a given jurisdiction do so in very specific political,
socioeconomic and cultural environments. Particularly
in HRH policy issues – in which financial, political, Figure 12.1 Basis for HRH policies
legal and cultural factors are of great significance
– it is extremely difficult to draw generalized conclu-
sions from a study conducted in one country, within
Principles and values
a specific context and at a specific point in time.
Furthermore, subnational or local-level characteristics
can differ markedly within a given country, especially
in situations of decentralized health systems. In other
words, one size definitively does not fit all, demanding
the development of mechanisms for contextualization HRH policies
of research evidence.
As health systems interventions are social experiments Stakeholders’ Information and

with an impact on large parts of the population, peo- interests evidence
ple affected by policies are inevitably key stakeholders
(18–20). Researchers and policy-makers must not only
know and understand the constraints and opportunities Source: Adapted from Muir Gray (24).
161
and retention). Information flows can balance these 12.3.1 Development and contributions
otherwise partial points of view. of health sector observatories
Since the 1970s, various types of cooperative mech-
To this end, several different but complementary
anisms – with the label “observatory” or other – have
approaches could be followed:
been established in countries at different levels of social
t strengthening health workforce information systems
and economic development, focusing on different pub-
to generate the data needed for monitoring and
lic health themes. These have included observatories
evaluation of HRH policies and programmes;
charged with aiding local health and social care pol-
t building research capacities to produce and dissem-
icy decision-making in France through the production
inate information and evidence for the formulation of
of useful information, and the public health observa-
policy options among decision-makers and other
tories in the United Kingdom, which were considered
stakeholders;
to have broken new ground in the provision of health
t facilitating knowledge translation platforms for
intelligence by not only providing information but also
information and evidence sharing among all key
context and perspective (29).
stakeholders;
t establishing and strengthening cooperative mech-
At the regional level, the implementation of wide-
anisms and processes, such as health workforce
spread health-care reforms and increasing recognition
observatories, with the aim of integrating HRH
of the need for better monitoring in the late 1990s wit-
information and evidence (including analyses
nessed the emergence of observatories on health
of best practices) with policy development and
systems and human resources, notably the European
implementation.
Observatory on Health Systems and Policies (30) and
the Observatory of Human Resources in Health Sector
Other chapters in this Handbook tended to focus on
Reforms (among countries of Latin America and the
the first two approaches; the rest of this chapter looks
Caribbean) (31).
into examples of the latter two. In particular, knowl-
edge translation platforms (for example discussion
When the European Observatory on Health Systems and
forums or virtual networks) are valuable for facilitating
Policies was initiated, it was described as a new part-
ongoing interactions between researchers and policy-
nership aiming to bridge the gap between academia
makers and helping transform scientific evidence into
and policy-makers. The rationale was explained by one
policy and practice (25). An example of a successful
of the observatory’s directors as follows:
knowledge translation platform for health systems and
policies focusing on low- and middle-income countries There is some information about what reforms
is given in Box 12.1 (26–28). work and where they work. But it is not read-
ily accessible or easy to interpret. Information
Health workforce observatories offer opportunities for is mostly in academic journals and much of it is
facilitating planning, monitoring and evaluation of HRH unpublished. The expertise of the observatory is
policy implementation by means of fostering mechanisms to collate information, analyse it, structure it, and
for timely access to and use of relevant data and infor- present findings clearly to those responsible for
mation. The next section concentrates on this approach, making and implementing health policy.
with examples from regional and country level.
Source: Figueras (32).
In the region of the Americas, stimulated by efforts of

12.3 A mechanism to harness the Pan American Health Organization (PAHO) to mon-
the HRH agenda: health itor ongoing health sector reforms and their impacts on
workforce observatories health labour, an initiative was launched in 1999 bring-
ing together 23 countries under an umbrella network of
Health workforce observatories are being increas- HRH observatories. This resulted in the development
ingly recognized as a potentially valuable mechanism of a community of professionals engaged in health
to improve the information and knowledge base on workforce policies. The regional observatory has been
the HRH situation and to facilitate policy develop- defined as “a cooperative initiative among the coun-
ment and monitoring. Although they have evolved in tries of the region aimed at producing information and
different ways in different contexts, they share a funda- knowledge in order to improve human resource policy
mental feature: a cooperative network initiative among decisions as well as contributing to human resources
countries and partners to produce and share informa- development within the health sector on the basis of
tion and knowledge necessary for improving human sharing experiences among countries” (33). Thus the
resources policy decisions. observatories are intended to be not just information
162
Box 12.1 Evidence-Informed Policy Network: EVIPNet
In response to a call “to establish mechanisms to transfer knowledge in support of evidence-based public
health and health-care delivery systems, and evidence-based health-related policies”, in 2005 the World
Health Organization launched the collaborative Evidence-Informed Policy Network (EVIPNet) (26).
EVIPNet addresses the integration of two fundamental enabling factors in the process of evidence-
informed policy-making: best practices to promote interaction among stakeholders that generate and use
evidence; and capacity to implement research in a local context (27). Focusing on low- and middle-income
countries, EVIPNet promotes partnerships at the country level between policy-makers, researchers and
civil society in order to facilitate both policy development and policy implementation through the use of the
best scientific evidence available.
EVIPNet comprises a series of linked (but distinct) networks that bring together national teams, which are
coordinated at regional and global levels. Africa, Asia and the Americas each host regional networks,
which together work in some 25 countries. At the regional level, EVIPNet is supported by WHO regional
offices and by small secretariats responsible for promoting regional coordination. A global steering group
facilitates exchange between regions and supports international-level activities. The work is guided by a
group of international experts in the evidence-to-policy field, which provides strategic guidance, organizes
technical review of EVIPNet proposals, develops innovative methodologies and provides direct technical
support to country teams where necessary. A project priority is to identify best practices in developing
capacity of both policy-makers and researchers in working with scientific evidence-to-policy links, while
engaging citizens and advocacy groups to sustain demand for evidence-informed health policy.
Since its inception, EVIPNet has worked collaboratively with the Regional East African Community
Health Policy Initiative, an institutional mechanism whose mission is to access, synthesize, package
and communicate evidence required for health policy and practice and for influencing policy-relevant
research agendas for improved health equity (28). Similar initiatives are emerging in other countries,
focusing on supporting governments to use research evidence for health policy decisions – although the
specific scope of activities, nature of collaborative efforts and research evaluation methods are unique
across contexts. For example, country teams may organize workshops to build technical capacity; several
promote the inclusion of journalists as observers and advocates. Thus EVIPNet has contributed to creating
a supportive political environment for delivery of evidence-informed policies.
Recently, country teams have began to produce policy briefs based on systematic reviews of research,
offering policy options that include governance, delivery and financial arrangements adapted to specific
country or district contexts and the reality of existing resources. For instance, in February 2008, EVIPNet
Africa country teams produced policy briefs to address current policy challenges related to malaria
treatment services in their respective countries.
disseminators, but also proactive actors in policy-mak- levels: subnational, national, intercountry and regional”
ing. For example, regular (biennial) regional meetings (35) (Box 12.2, page 166).
provide a forum for interactive discussion, sharing and
strategic planning. Similarly, with support from WHO, the Eastern
Mediterranean Region Observatory on Human
More recently, with increasing global attention and Resources for Health grew out of an existing regional
resource mobilization to address the health workforce observatory on health systems with the purpose “to
crisis in Africa, the Africa Health Workforce Observatory assist Member States in using a proactive approach
(34) has evolved as part of the action agenda. With its and sharing the best and most innovative options to
secretariat housed by the World Health Organization tackle HRH-related challenges” (36). Its outputs in the
(WHO), its mission is to “support actions that address public domain include updated profiles on the health
HRH challenges by promoting, developing and sustain- workforce situation within and across countries, analy-
ing a solid knowledge base for HRH information at all ses of correlations between significant health workforce
163
determinants, and operational and policy linkages to unions and associations; major private and nongovern-
national and regional HRH benchmarks and targets. mental organizations providing health services; civil
society (for example health services consumer and cli-
Although the functions of and triggering force for ent groups); and development partners.
the emergence of these various observatories differ
slightly, depending on specific contexts and needs, National health workforce observatories can serve as
each works to bridge the gap between evidence and mechanisms to promote collaborative engagement
policies. The common objectives can be summarized among stakeholders and enhance their roles and
as: to produce information and knowledge necessary contributions to HRH development. However, just as
to improve human resources policy decisions, planning regional observatories evolved in different ways, so
and implementation; to share country experiences in too may national observatories have different priori-
human resources development; and to facilitate policy ties, structures, memberships and ways of working,
dialogue. Recent significant contributions attributable as driven by country contexts and interests. Table
to regional HRH observatories have included: 12.1 provides an illustrative example of potential stake-
t the creation and dissemination of databases of holders and their roles in the context of the Sudanese
cross-nationally comparable statistics on various national observatory (38).
indicators of health workforce metrics – for example,
the public interface for the African atlas of the health At the same time, some common principles and fea-
workforce is maintained on the regional observatory tures among operating national HRH observatories can
web site (37); be identified:
t the generation of regional and country profiles of the t They build on existing structures and mechanisms,
HRH situation and trends for overview analysis and such as national technical multisectoral and partner
benchmarking; working groups, avoiding duplication of efforts and
t the undertaking of intercountry and national studies parallel structures.
on specific policy-relevant health workforce topics t Organizational flexibility accommodates diverse
– for example, joint activities across regional observ- sociopolitical contexts (although the actual mod-
atories for capacity assessment of HRH units in els of organization vary from country to country,
national ministries of health (in the WHO regions of according to existing institutional arrangements and
Africa, the Americas and the Eastern Mediterranean) leadership).
and for mapping the health management workforce t Coordination functions are assumed through either
(in the African and Eastern Mediterranean regions). a small secretariat or focal point.
t Active involvement of the ministry of health, espe-
cially in a leading role, is essential.
12.3.2 National health
t They can contribute to strengthening capacities
workforce observatories and empowerment of the HRH units of ministries of
While regional observatories may play an essential role health and other stakeholders.
in terms of organization and coordination, national HRH
observatories lie at the core of achieving the desired In line with these principles, Boxes 12.3, 12.4 and 12.5
objectives discussed above. They operate to synergize offer examples across different operational frame-
efforts and consolidate resources at the country level, works for HRH observatories (39–42). The stimulation
contributing to a commonly agreed workplan among all of joint work and products among participating stake-
stakeholders to respond to country needs. holders can contribute to continuation of advocacy
and commitment for observatories, and therefore to
National observatories involve networks of all part- their sustainability. An outstanding question remains
ners in health workforce development in the country. regarding whether the national observatory should
Routine responsibilities of network members often be integrated as part of the health ministry’s HRH unit
entail monitoring and documenting implementation of (such as in the case of Brazil), or as an autonomous
HRH policies and strategies. While members may each interinstitutional group with guidance from the national
operate under their own identities, tasks are divided authority. There are advantages to both options, but
in a coordinated manner and the results of the work also risks – in the former case, close proximity to the
are shared in a systematic way. Observatory networks political directives may inhibit widespread stakeholder
typically involve representatives at various levels: gov- participation; in the latter, the information generated
ernment (ministry of health, ministry of education, may not be relevant for policy and planning.
ministry of finance, public service commission, national
statistics office); academic and training institutions; Overall, experiences from different contexts have shown
health professional regulatory bodies; health workers’ that HRH observatories can be an effective mechanism
164
Table 12.1 Main stakeholders and their role in support of the national HRH observatory in Sudan
Stakeholder Current rolea Potential role

Federal Ministry of Health Annual health statistics report Improving scope and quality of official statistical
Administrative records of the ministry’s reporting
HRH department Periodical reports from HRH department
Mappings and surveys on the health Maintaining a centralized database on health workers
workforce
Ministry of Higher Annual report on medical schools (staff Inclusion of other health training institutions
Education and students) Improving scope and quality of the report
Health workforce production and education policies
Ministry of Labour No obvious role Records and reports on health-related jobs
Records of scholarships for health personnel
Labour market dynamics
Sudan Medical Council Registry of doctors, dentists and Periodically updated registry (relicensing)
pharmacists Capacity building
Council for Allied Health Registry of nurses and paramedical Periodically updated registry (relicensing)
Professions staff Capacity building
Sudan Medical Records of registered doctors enrolled Annual report on intake and graduation
Specialization Board for specialized training
Sudan Health and Social No role Records of membership (regularly updated)
Professions Trade Union
Sudan Doctors Union No role Records of membership
Records of doctors abroad
Capacity building
Negotiation
Army Medical Corps No obvious role Records and reports on health workers affiliated to
the armed forces
Police Health Services No obvious role Records and reports on health workers affiliated to
Department police services
Secretariat for Sudanese Records of some categories of migrant Records of all categories of migrant health workers
Working Abroad Sudanese health personnel Annual analytical report on health workers abroad
Health Insurance Fund No obvious role Records of health workers affiliated to the National
Health Insurance Fund
Annual report on characteristics of health workers
Quality of practice
National Centre for No role Facilitation of access to data and information across
Information different government agencies
Support for analysis and other technical aspects of
data collection and use
Ministry of Health/ Records of health workers in the Comprehensive records on all health workers
Government of Southern country’s southern region Annual report on characteristics of health workers
Sudan
Health workforce development policies
Facilitation of dialogue
Coordination
WHO Country Office for Technical support to the Federal Support for health workforce assessments
Sudan Ministry of Health Support and technical assistance in the area of
information and communication technology
Regional networking and exchange of experience
a. Current role as assessed in early 2007.
Source: Badr (38).
165
to improve information and knowledge flows in support Table 12.2 reviews a selection of studies undertaken
of evidence-based decision-making to address health in the context of the Brazilian observatory network and
workforce challenges in countries. As illustration, their role in influencing national HRH policies.
Box 12.2 Africa Health Workforce Observatory
Among a series of recent initiatives to monitor and address the HRH crisis in Africa, home to critical
shortages and imbalances of health personnel, was the establishment the Africa Health Workforce
Observatory (34).
The observatory was started with a small secretariat based in the WHO Regional Office for Africa. A
range of activities aimed to improve HRH data, information and evidence across the region. The main
tasks and functions include information dissemination, facilitating networking for sharing of experiences,
intercountry studies to contribute to national policy-making processes and support for additional national-
level activities.
The work of the regional secretariat is guided by a steering group comprising representatives of the
observatory’s key partners, including the regional economic communities of the African Union, the New
Partnership for Africa’s Development, academic and training institutions, international organizations (WHO,
World Bank, European Commission) and bilateral agencies (for example the United States Agency for
International Development).
A core priority of the regional observatory is to support the establishment and strengthening of national
HRH observatories. The Ethiopia Health Workforce Observatory was the first launched at the national level,
followed by those in Ghana and the United Republic of Tanzania. Others are also in the process of being
establishing in a number of countries, often with additional support from subregional groupings such as
the East, Central and Southern African Health Community or the West African Health Organization. The
national health workforce observatories work to bring together stakeholders for policy dialogue on HRH
issues at the country level.
Box 12.3 Brazil Human Resources for Health Observatory
In Brazil, the national Human Resources for Health Observatory (Observatório de Recursos Humanos en
Saúde do Brasil) comprises a network of some two dozen workstations that has been gradually developed
across universities and state-level health departments. The work is led by a secretariat based in the
Federal Ministry of Health in partnership with the WHO Country Office for Brazil, which provided financial
and technical support.
The workstations focus on the following:
t HRH research oriented to specific topics and requests proposed by the Ministry of Health;
t sharing and disseminating the results of such research as a public good through different channels, but
most notably via the observatory’s web portal (39);
t joint activities and products, including regular meetings and newsletters;
t facilitating the entry of new members into the network;
t cooperating with other centres in Brazil and in other countries of the Latin America and Caribbean
region in the promotion of the HRH research agenda.
Over the years, the different workstations have produced a number of thematic research reports, some
of which have been compiled into a series of books. The network has also created several tools and
instruments available in the public domain for HRH practitioners, such as methodologies for contracting
public health workers, software for analysing health services labour markets and databases of health
professions training institutions.
166
Box 12.4 Ghana Health Workforce Observatory
The Ghana Health Workforce Observatory started as an HRH technical team in 2005. With support
from key stakeholders involved in health workforce planning, monitoring, training and management, the
technical team developed a five-year plan to guide HRH development policies and strategies in the
country (40). The team was eventually expanded and transformed into an observatory. Bringing on board
many other national and international partners, it was established as part of the larger regional Africa
Health Workforce Observatory (see Box 12.2). The Ghana observatory and its official web site were
launched by the Deputy Minister of Health in December 2007 (41).
The Ghana observatory is composed of an advisory body, a technical committee and a secretariat with
representation from the Ministry of Health and other government ministries and agencies, universities and
other stakeholder institutions. Its objectives are to:
t strengthen the stewardship and regulation capacity of the Ministry of Health and its agencies;
t support and promote evidence-informed HRH policy-making;
t strengthen the knowledge base and use of data and information in HRH for policy and
decision-making;
t increase capacity in evaluating and monitoring the human resource situation and trends in the
health sector;
t provide a forum and a network for sharing experience among HRH data producers and users.
Since its inauguration the observatory has convened several events, including an HRH round-table
conference in June 2008 that sought to instigate action on the five-year strategic plan and also identify
gaps in the existing document. The role of the observatory would be to serve as machinery for the
implementation of decisions agreed upon during this first (and any subsequent) stakeholder conference.
Source: Contributed by James Antwi, Deputy Director, Human Resources for Health Development, Ministry of Health, Ghana.
Table 12.2 Influences of selected achievements of the Brazilian Human Resources for Health
Observatory on policy processes
Resulting policy decisions Health workforce study or analysis

Incentives for health workers in rural areas Profile of HRH stock and distribution: snapshot and cohort
studies
Creation of a joint high-level working group with the Ministry Trends in education and training for health professions
of Health and Ministry of Education (expansion, mix)
Long-distance training programmes Assessment of managerial skills of district (municipal)
health teams
Policy dialogue and governmental proposal for regulatory Contractual arrangements of the national health system in
norms (decree/law) the public sector (federal and state levels)
Expansion of education and training programmes for Assessment of workforce skills mix, with attention to
selected health professions (including nursing and certain selected health professions (e.g. team composition of dental
medical specialization programmes) workforce, including dentists, auxiliary dentists and dental
hygiene technicians)
Permanent negotiation round table Professional practices and interests (e.g. conflicts in scopes
of practice)
Regulation of new health professions
167
Box 12.5 Andean Human Resources for Health Observatory Network
The Andean Human Resources for Health Observatory Network (Rede ObservaRH Edmundo Granda) a
was developed under the auspices of the Observatories of Human Resources for Health of the Americas
and with the support of a technical cooperation agreement between PAHO and the Brazilian Government.
The network gathers partners from national HRH observatories from six countries: Bolivia, Chile, Colombia,
Ecuador, Peru and Venezuela. Its function is to support local actions among participating national teams,
with specific activities depending on the internal capacity of each country (42).
What is the Andean Human Resources for Health Observatory Network?
The observatory network is a cooperative mechanism encompassing the countries of the Andean
subregion. Its purpose is the exchange and joint generation of updated information, experiences and
knowledge to make decisions related to the development of human resources in health as a strategic
factor of the national health systems.
What are its objectives?
The objectives of the observatory network include:
t dissemination of strategic information that reflects the environment, advances and processes in HRH
issues for the Andean countries;
t generation of a space for communication among stakeholders in HRH issues in member countries;
t presentation of different perspectives in the analysis of HRH challenges by the various interested
parties, including universities, research units, governmental institutions, representative organizations
(schools, unions), health services providers and health services users;
t Provision of timely and relevant information for HRH policy-makers;
t Follow-up on commitments for the fulfilment of regional goals for HRH development.
What does it cover?
t research and systematized experiences in matters of HRH;
t official data and statistics on human resources in the health sector;
t connections among participating workstations in the Andean observatory network, the Brazilian
observatory network, and the regional network for the Americas based at PAHO headquarters.
How does it work?
The Andean Human Resources for Health Observatory Network:
t has a multicentric character, with each country assuming responsibility for the national contents;
t is anchored by a dynamic subregional node that is in continuous communication with the participating
countries;
t has information management mechanisms in place at various levels (regional, country and institutional)
for the intake, processing, updating and validation of HRH information;
t offers a common web-based platform for country administrators to post their national information.
a. The naming of the observatory network after Edmundo Granda (1946–2008) was agreed upon in memory of the distin-
guished Ecuadorian professor for his invaluable contributions to the field of public health, particularly HRH issues.
Source: Contributed by Mónica Padilla, Advisor, Human Resources for Health, Andean Subregion, PAHO.
168
12.4 Opportunities and directions long-term, intentional and coordinated efforts for HRH
development at the international, national, regional and
The development and sustainability of long-term, com- subregional levels (43).
prehensive HRH policies and plans is a common
challenge for all countries to ensure the health work- A number of critical success factors in moving forward
force is prepared to meet current and future health with health workforce observatories and enhancing
system objectives and population health needs with their benefits have been identified (44). They include:
equitable and adequate coverage. In many coun- t championing HRH issues in the country (including
tries, this entails strengthening institutional capacity for high-level political commitment and leadership);
defining appropriate policies and revising them period- t engaging all key stakeholders in joint planning and
ically, which itself is dependent upon close cooperation nurturing joint work;
among a wide range of stakeholders in the policy dia- t ensuring effective coordination;
logue from the inception stages. This includes not only t building awareness and capacity in HRH issues;
the ministry of health but also other sectors: finance t developing approaches from different experiences;
department, public service commission, educational t harmonizing standards, definitions and indicators
bodies, health professional regulatory bodies and for HRH profiling and analysis;
associations, programme managers (in the public, t supporting networks of HRH researchers;
parapublic and private sectors), development partners t institutionalizing coordination mechanisms (man-
and health services users’ groups. In order to achieve date and legitimacy);
balance across the different perspectives, and being t creating a communication mechanism for tracing
mindful that each country has specific challenges and the available information for public use (for exam-
contexts for workforce development, a common frame- ple a web site);
work needs to be identified where collaborative efforts t mobilizing resources (technical and financial).
can be focused. To this end, this chapter has exam-
ined various strategies and mechanisms to ground In summary, health workforce observatories and other
HRH policies and strategies in scientific information knowledge transfer platforms present dynamic and
and evidence. evolving networks, which can only stand to bene-
fit from increasing numbers of participating countries
The sustainable provision of timely, reliable and relevant and institutions. Securing the initial commitment is
data, information and evidence to improve HRH pol- often a challenge in establishing such mechanisms,
icies requires mechanisms to facilitate dissemination, but maintaining interest and commitment (often while
access and use in policy-making processes. Different coping with political changes and shifting donor prior-
knowledge transfer platforms (such as EVIPNet) and ities) is another critical concern. This requires regular
cooperative mechanisms (health workforce observato- dissemination of products that have proven to be use-
ries) can offer the opportunity for global and national ful, active sharing and exchange of information and
health institutions to take advantage of the latest inno- knowledge, and continuous advocacy activities. The
vative and sound tools to support decision-making. last-mentioned should including a strong, actionable
Their value lies in the forums they extend to facilitate communications strategy – aimed at policy-makers,
sharing of information and experiences and promotion managers, researchers and other relevant stakeholders
of collaborations at the national, regional and interna- inclusively – to advocate recognition of health workers
tional levels. as the foundation of health systems and implemen-
tation of effective strategies to directly address HRH
In particular, health workforce observatories can be a bottlenecks, barriers and funding gaps at all levels.
good mechanism to facilitate the steering and nego-
tiating processes of cooperative partnerships. They
can contribute to the strengthening of working rela-
tions and development of joint agendas among
stakeholders. Their informational products help bring
new evidence on the health workforce situation to a
broader audience, often in a standardized way to fos-
ter better understanding and dialogue for comparisons
and benchmarking. Perhaps most crucially, they can
effectively raise the priority of HRH issues in the health
development agenda; for example, through its working
groups and mobilization efforts, the Observatories of
Human Resources for Health of the Americas network
has been credited as central in shaping the agenda for
169
References 15. Siddiqi K, Newell J, Robinson M. Getting evidence

into practice: what works in developing countries?
1. Clancy CM, Cronin K. Evidence-based decision International Journal for Quality in Health Care, 2005,
making: global evidence, local decisions. Health 17(5):447–453.
Affairs, 2005, 24(1):151–162 (http://content.
healthaffairs.org/cgi/content/full/24/1/151, accessed 16. Strengthening health systems: the role and promise
29 January 2009). of policy and systems research. Geneva, Alliance for
Health Policy and Systems Research, 2004 (http://
2. Wallace A et al. Meeting the challenge: developing www.who.int/alliance-hpsr/resources/Strengthening_
systematic reviewing in social policy. Policy Politics, complet.pdf, accessed 29 January 2009).
2004, 32:445–470.
17. Stepping stones to improving the monitoring of vital
3. Sheldon TA. Making evidence synthesis more events. Geneva, Health Metrics Network, 2007
useful for management and policy-making. Journal (http://www.who.int/healthmetrics/tools/logbook/en/
of Health Services Research and Policy, 2005, move/web/index.html, accessed 29 January 2009).
10(Suppl. 1):1–5 (http://jhsrp.rsmjournals.com/cgi/
reprint/10/suppl_1/1, accessed 29 January 2009). 18. Trostle J, Bronfman M, Langer A. How do
researchers influence decision-makers? Case
4. Bowen S, Zwi AB. Pathways to “evidence-informed” studies of Mexican policies. Health Policy and
policy and practice: a framework for action. Planning, 1999, 14(2):103–114 (http://heapol.
PLoS Medicine, 2005, 2(7):1–14 (http://medicine. oxfordjournals.org/cgi/reprint/14/2/103?ck=nck,
plosjournals.org/archive/1549–1676/2/7/pdf/10.1371_ accessed 29 January 2009).
journal.pmed.0020166-L.pdf, accessed 29 January
2009). 19. De Savigny D et al. Fixing health systems: linking
research, development, systems, and partnerships.
5. Lavis J et al. Towards systematic reviews that inform Ottawa, International Development Research Centre
health care management and policy-making. Journal and Tanzania Ministry of Health, 2004.
of Health Services Research and Policy, 2005,
10(Suppl. 1):35–48 (http://jhsrp.rsmjournals.com/cgi/ 20. Regional East African Community Health-Policy
reprint/10/suppl_1/35, accessed 29 January 2009). Initiative. REACH-Policy Initiative: prospectus.
Arusha, East African Community, 2005 (http://
6. Tomson G et al. Decision-makers and the usefulness www.idrc.ca/uploads/user-S/11551301781REACH_
of research evidence in policy implementation: Prospectus.pdf, accessed 29 January 2009).
a case study from Lao PDR. Social Science and
Medicine, 2005, 61(6):1291–1299. 21. Sauerborn R, Nitayarumphong S, Gerhardus A.
Strategies to enhance the use of health systems
7. Tetroe JM et al. Health research funding agencies’ research for health sector reform. Tropical Medicine
support and promotion of knowledge translation: and International Health, 1999, 4(12):827–835.
an international study. Milbank Quarterly, 2008,
86(1):125–155. 22. Lomas J. Connecting research and policy. ISUMA,
2000, 1(1):140–144 (http://portals.wi.wur.nl/files/docs/
8. About knowledge translation. Canadian Institutes of ppme/lomas_e.pdf, accessed 29 January 2009).
Health Research (http://www.cihr-irsc.gc.ca/e/29418.
html, accessed 29 January 2009). 23. Pawson R et al. Realist review: a new method of
systematic review designed for complex policy
9. Walt G, Gilson L. Reforming the health sector in interventions. Journal of Health Services Research
developing countries: the central role of policy and Policy, 2005, 10(Suppl. 1):21–34.
analysis. Health Policy and Planning, 1994,
9(4):353–370. 24. Muir Gray JA. Evidence-based healthcare: how to
make health policy and management decisions, 2nd
10. Innvaer S et al. Health policy-makers’ perceptions of ed. Edinburgh, Churchill Livingstone, 2001.
their use of evidence: a systematic review. Journal
of Health Services Research and Policy, 2002, 25. Lavis JN et al. Assessing country-level efforts to
7(4):239–244. link research to action. Bulletin of the World Health
Organization, 2006, 84(8):620–628 (http://www.who.
11. Lavis JN et al. How can research organizations more int/bulletin/volumes/84/8/06–030312.pdf, accessed
effectively transfer research knowledge to decision 29 January 2009).
makers? Milbank Quarterly, 2003, 81(2):221–248.
26. Evidence-Informed Policy Network (EVIPNet). World
12. Lavis JN et al. Use of research to inform public Health Organization (http://www.who.int/rpc/evipnet,
policymaking. Lancet, 2004, 364(9445):1615–1621. accessed 29 January 2009).
13. What is health policy and systems research and why 27. Hamid M et al. EVIPNet: translating the spirit of
does it matter? Briefing Note No. 1. Geneva, Alliance Mexico. Lancet, 2005, 366:1758–1760.
for Health Policy and Systems Research and World
Health Organization, 2007 (http://www.who.int/ 28. van Kammen J, de Savigny D, Sewankambo N.
alliance-hpsr/resources/Alliance%20HPSR%20-%20 Using knowledge brokering to promote evidence-
Briefing%20Note%201.pdf, accessed 29 January based policy-making: the need for support
2009). structures. Bulletin of the World Health Organization,
2006, 84(8):608–612 (http://www.who.int/bulletin/
14. Lavis J et al. Examining the role of health services volumes/84/8/05–028308.pdf, accessed 29 January
research in public policymaking. Milbank Quarterly, 2009).
2002, 80(1):125–154.
170
29. Hemmings J, Wilkinson J. What is a public 43. Toronto Call to Action 2006–2015: towards a decade
health observatory? Journal of Epidemiology and of human resources in health for the Americas.
Community Health, 2003, 57:324–326 (http://jech. Report of the regional meeting of the Observatory of
bmj.com/cgi/content/full/57/5/324, accessed 29 Human Resources in Health, 4–7 October 2005. Pan
January 2009). American Health Organization et al., 2005 (http://
www.observatoriorh.org/Toronto/CallAction_eng1.
30. European Observatory on Health Systems and
Policies (http://www.euro.who.int/observatory,
accessed 29 January 2009). 44. Report of the meeting of the Africa Health Workforce
Observatory, 26–29 September 2006, Arusha,
31. El Observatorio de Recursos Humanos en las
Tanzania. Africa Health Workforce Observatory,
Reformas Sectoriales en Salud (http://www.
2006 (http://www.afro.who.int/hrh-observatory/
observatoriorh.org, accessed 29 January 2009).
documentcentre/meetingreports/arusha_meeting_
32. Figueras J (as interviewed by Richards T). European report_09_07.pdf, accessed 29 January 2009).
observatory will promote better health policy.
British Medical Journal, 1999, 318(7180):352
(http://www.pubmedcentral.nih.gov/articlerender.
fcgi?artid=1114839, accessed 29 January 2009).
33. Rigoli F, Arteaga O. The experience of the Latin
America and Caribbean Observatory of Human
Resources in Health. Report prepared for the Joint
Learning Initiative, 2004 (http://www.observatoriorh.
org/eng/pdfs/arteagaObervatoryHR.pdf, accessed
29 January 2009).
34. Africa Health Workforce Observatory (http://www.
afro.who.int/hrh-observatory/, accessed 29 January
2009).
35. Africa Health Workforce Observatory: concept and
implementation strategy. Geneva, World Health
Organization et al., 2006 (http://www.afro.who.
int/hrh-observatory/documentcentre/concept_
implementation_strategy.pdf, accessed 29 January
2009).
36. Eastern Mediterranean Region Observatory on
Human Resources for Health (http://www.emro.who.
int/hrh%2Dobs/, accessed 29 January 2009).
37. African atlas of the health workforce. World
Health Organization and Africa Health Workforce
Observatory (http://www.afro.who.int/hrh-
observatory/hwinformation/index.html, accessed 29
January 2009).
38. Badr EE. Establishing an observatory on human
resources for health in Sudan. Report prepared for
the World Health Organization and Federal Ministry
of Health of the Republic of Sudan, 2007 (http://
www.hrhobservatory.sd/Documents/observatory-
Badr.pdf, accessed 29 January 2009).
39. Rede ObservaRH: Observatório de Recursos
Humanos en Saúde do Brasil (http://www.
ObservaRH.org.br, accessed 29 January 2009).
40. Ministry of Health of the Republic of Ghana. Human
resource policies and strategies for the health
sector, 2007–2011. Accra, Ministry of Health of
the Republic of Ghana, Quality Health Partners
and World Health Organization, 2007 (http://www.
ghanahrhobservatory.org/downloads/hrh_policy_
and_plan.pdf, accessed 29 January 2009).
41. Ghana Health Workforce Observatory (http://www.
ghanahrhobservatory.org, accessed 29 January
2009).
42. Observatorio Andino de Recursos Humanos en
Salud (http://bvsde.per.paho.org/oarhs2/index.php,
171
172
Index
Index
Note: Page numbers in bold type refer to main entries
A health workers 14–22, 85–86

occupations 16, 18–19
Absenteeism 29, 56, 152–153 COFOG (Classification of the Functions of Government) 65
Accounting systems 65–67, 71–72, 76 Computer-based records, see Information technology
Administrative records 27, 30, 66–67, 113–126 Congo, workforce mortality 55
Africa (see also under the names of specific countries) Coping strategies 133
health worker shortfalls and imbalances 79 COPP (Classification of the Outlays of Producers According to
out-migration 53, 54–55, 58 Purpose) 65
Africa Health Workforce Observatory 163, 166 Corrupt practices 133, 136, 137, 138, 152–154
Age distribution of health workers 27, 28, 86, 87, 110 Costs, see Expenditure
Algeria, out-migration 54 Côte d’Ivoire, workforce mortality 55
Americas (see also under the names of specific countries) CPC (Central Product Classification) 65
education and training 42 Cuba, out-migration 54
Andean Human Resources for Health Observatory Network 168 Cultural competency 41
Angola, out-migration 55 Cultural factors 129, 161
Antigua and Barbuda, out-migration 55
Argentina
census data 105 D
educational attainment 109 Data collection
gender analysis 106, 107, 109 to complement existing sources 27
Assessment of Human Resources for Health (WHO) 84 guides and recommendations 21
Australia, out-migration 54 health facility assessments 81–82
human resources information systems 117
B methods 81–82, 84
qualitative research 138–141
Brazil
tools 21, 66, 72, 83–84
census data 105
workforce entry 43
educational attainment 109
Data management 33, 43, 117
gender analysis 106, 107, 109
Data presentation 31
HRH information and management system 122–125
Data protection 43, 117–118
HRH observatory 166, 167
Data quality 8, 117, 126
Data security, see Data protection
C Data sources (see also Administrative records; Labour force
surveys; Population census data) 14, 25, 27, 30–31
Cambodia
education and training 44
census data 105
entry to the workforce 44–45
expenditure 66–67
synthesis 147–154
Canada
workforce transitions and exits 51–52
out-migration 54
Data verification 35, 44–45, 76
retirement 56
Definitions
working hours estimation 69
data 4
Caribbean countries, out-migration 54–55
health facilities 86
Census data, see Population census data
health sector 66
Central African Republic, workforce mortality 55
health workers 13–14, 85
Central Product Classification (CPC) 65
indicators 4
Certification and licensing, see Professional certification and
monitoring and evaluation 4
licensing
remuneration 70–71
Chile
Demographic factors 27
census data 105
Denmark, exits from the workforce 56
Density of health workers 28, 32, 88, 105, 151–152
gender analysis 107, 109
Distribution of health workers (see also Geographical distribu-
China, out-migration 54
tion of health workers; Skills mix) 92–94
Civil service payroll registries 30
Dual practice 28, 85, 133
Classification of the Functions of Government (COFOG) 65
Classification of the Outlays of Producers According to Purpose
(COPP) 65 E
Classification systems
Earnings, see Remuneration
economic sectors 20
Eastern Mediterranean Region Observatory on Human
education and training programmes 14–16, 17
Resources for Health 163–164
expenditure 65–66, 71–72, 74
Economic sectors, classification 18–19
health facilities 86
173
Ecuador G
census data 105
Gender distribution of health workers 28, 103–111
Geographical distribution of health workers
attrition rates 97
Education and training 37–40
data presentation 31
accreditation of institutions 39–40, 44
data synthesis 152
capacity and output of institutions 6, 40, 42, 44, 45
education and training issues 37–38
classification of programmes 14–16, 17
facility-based assessments 92–94
eligibility requirements 39, 44
indicator 28
expenditure 68–69
transitions within a country 52
facility-based assessments 94, 95
Germany, out-migration 54
gender analysis 108–109
Ghana
indicators 44
HRH observatory 167
recruitment and selection 39, 44
out-migration 54
Egypt, out-migration 54
Ghost workers 30, 152–154
Employment relationship 71
Government expenditure and financing
Employment surveys, see Labour force surveys
for education and training 46
Entry to the workforce (see also Education and training) 37–47,
government accounts 73, 74–76
86–87
indicators 64
certification and licensing 41
Grenada, out-migration 55
education and training 39–40
Guyana, out-migration 55
international in-migration 40
measurement issues 42–46
recruiting back individuals 57 H
recruitment into labour market 41–42
Haiti, out-migration 54
Equity in the workforce (see also Gender distribution of health
Health accounts 65, 71–72, 76
workers) 65
Health facility assessments (HFAs) 27, 30, 32, 79–100
Ethiopia
active health workforce 88–91
absenteeism in the workforce 132
data collection methods and tools 81–85
qualitative research 132, 136, 137, 138, 139, 142
definitions and classification 85–86
workforce mortality 55
entry to the workforce 86–87
European Observatory on Health Systems and Policies 162
exits from the workforce 97–98
European System of Accounts 65
indicators for 80
Eurostat, expenditure data collection 66, 72
workforce distribution 92–94
Evidence-based policy-making 6, 159–169
workforce education, motivation and performance 94–97
Evidence-Informed Policy Network (EVIPNet) 163
Health Facility Census (HFC) 83
Exits from the workforce 49, 50, 53–56
Health Metrics Network 6
facility-based assessments 97–98
Health occupations 16–19
gender analysis 109–110
Health workforce
indicators 51
activity indicators 28–29
intersectoral transitions 56
classification 14–22, 85–86, 104
migration, see Migration of health workers
composition (see also Gender distribution of health workers;
mortality 55
Skills mix) 65
other reasons 56
definition 13, 66, 85
retirement 56
distribution (see also Geographical distribution of health
Expenditure (see also Remuneration) 63–77
workers) 92
classification 65–66, 71–72, 74
dynamics (see also Transitions within workforce) 25
components of 66, 72
indicators 28–29
for education and training 46
monitoring 25–34
by funding agent 32, 73
motivation and performance 8–9, 39, 80, 95–96, 133–134
government accounts 74
observatories 126, 162–169
health accounts 71–72
productivity indicators 29
indicators 64–65
renewal and loss indicators 29
measuring 65–71
working lifespan 8–9, 38, 49, 57, 76, 80, 98, 103, 106, 118
national accounts 72–74
HFC (Health Facility Census) 83
sub-accounts 76
Household surveys (see also Labour force surveys) 27, 30
Human resources information systems (HRIS) 113–126
F data management and use 117–118
development steps 116–117
Facility Audit of Service Quality 84
implementation issues 114–116
Facility-based assessments, see Health facility assessments
licensure information system 119–120
(HFAs)
Hungary
Fiji, out-migration 55
census data 105
Flow-based estimates 51, 52
Focus groups 135–136, 137–138, 139, 140, 142
gender analysis 107, 109
Former Soviet Union 54
Funding for HRH systems, see Resources for HRH systems
174
Index
I K
iHRIS Qualify 119 Kenya
Imbalances and inequalities (see also Gender distribution of active health workforce 88–91
health workers) 31 distribution of health workers 92–93
Implementation issues 33–35, 114 education and training 94, 95
India facility-based assessments 86, 94–96
geographical distribution of health workers 152 international out-migration 53
out-migration 54 nursing workforce 32–33
workforce analysis 150–152 out-migration 54
Indicators 8, 26–29 workforce entry 86, 87
definition 4 workforce mortality 55
disaggregation 26–27 working hours 96–97
education and training 44–45 Key indicators, see Indicators
expenditure 64–65
for health facility assessments 80
for monitoring of HRH system implementation 34, 35, 123–124 L
remuneration 29, 65–66 Labour costs, see Expenditure
workforce entry 42, 44–46 Labour force surveys 30, 32–33, 67–68, 152
workforce transitions and exits 51–53 Language and cultural competency 41
Inequalities, see Gender distribution of health workers; Lesotho, occupational transition 53
Geographical distribution of health workers Liberia
Inflows of health workers, see Entry to the workforce out-migration 55
Information and monitoring systems (see also Human resources workforce mortality 55
information systems (HRIS)) Licensing of health workers, see Professional certification and
assessments for selected countries 6 licensing
comparability of data 8
critical requirements 34
implementation issues 33–35, 114
M
indicators for monitoring implementation and use 35 Macro International, Service Provision Assessment (SPA) 83
management issues 6, 33–34, 114–115 Madagascar, workforce mortality 55
strengthening 6–8, 35, 116 Management and support workers, see Non-clinical workers
Information quality, see Data quality Management issues 6, 33–34, 114–115
Information sources, see Data sources MEASURE Evaluation 84
Information technology Metadata 31, 66
access to 7, 33, 43, 115–116 Mexico
open source software 119, 126 geographical distribution of health workers 31
In-migration, see Migration of health workers health-care expenditure 73
In-service training 68–69, 94, 95 identifying data sources 147–150
Integrated Public Use Microdata Series (IPUMS) 21, 104 out-migration 54
International initiatives 3–7 workforce analysis 149–150
International Labour Organization 16, 21, 65, 149 Migration of health workers
International migration, see Migration of health workers certification and licensing issues 41
International Standard Classification of Education (ISCED) codes of practice 57
14, 15–16, 21, 26, 39, 43, 46, 65, 85 indicators 45
International Standard Classification of Occupations (ISCO) international in-migration 40, 41, 45
9, 14, 16, 18–19, 21, 26, 65, 85, 104 international out-migration 40, 52, 53–55
International Standard Industrial Classification of All Economic language and cultural competency 41
Activities (ISIC) 14, 20, 21, 26, 65, 66, 86 policy issues 42
Intersectoral mobility, see Exits from the workforce qualitative research 134
Interviewing methods 82, 134–138 transitions within a country 52
Intrasectoral mobility 50, 56 Millennium Development Goals 3, 4, 5, 79
IPUMS (Integrated Public Use Microdata Series) 21, 104 Monitoring systems, see Information and monitoring systems
Iran, out-migration 54 Morocco, out-migration 54
Ireland, out-migration 54 Mortality among health workers 52, 55, 57
ISCED (International Standard Classification of Education) Motivation and performance 8–9, 39, 80, 95–96, 133–134
14, 15–16, 21, 26, 39, 43, 46, 65, 85 Movements within workforce, see Transitions within workforce
ISCO (International Standard Classification of Occupations) Mozambique, out-migration 55
9, 14, 16, 18–19, 21, 26, 65, 85, 104 Multiple data sources, analysis and synthesis 31, 147–154
ISIC (International Standard Industrial Classification of All
Economic Activities) 14, 20, 21, 26, 65, 66, 86
N
Namibia, labour force surveys 32
J National accounts 65, 72–74
Jamaica, out-migration 54 National databases 33
Japanese International Cooperation Agency (JICA) 83 National health workforce observatories 164–168
JICA (Japanese International Cooperation Agency) 83 Nepal
international out-migration 53
out-migration 54
175
Netherlands Productivity indicators 29

exits from the workforce 56 Professional certification and licensing 27, 41, 116
workforce data sources 68 case study 119–120
New Zealand, out-migration 54 indicators 45
Nigeria nurses seeking employment abroad 53, 54
entry to the workforce 99 registries 30
exits from the workforce 97–98, 99 Public sector expenditures, see Government expenditure and
out-migration 54 financing
workforce entry 86, 87
Non-clinical workers 8, 13, 14, 16, 20, 108
Non-health activities of health service personnel 108 Q
Non-monetary benefits 95–96 Qualitative methods 129–144
Nursing data recording and analysis 141–143
gender analysis 109–110 design and data collection 138–141
international migration 53, 54 illustrations 133–134
qualification and registration 54, 120, 121 interviews 134–138
recruiting back individuals 57 value and potential 130–132
workforce analysis in Mexico 149, 150 Quantitative Service Delivery Survey (World Bank) 84
workforce in Kenya 32–33
R
O Recruitment into labour market (see also Entry to the workforce)
Observatories 126, 162–169 41–42
Occupational groups indicators 45
classification 16–19, 85–86 recruiting back individuals 57
gender analysis 106–108 Registries, professional (see also Professional certification and
occupational titles 18 licensing) 30
skills mix 26, 28, 32, 92, 151 Reimbersements 71
transitions between 50, 52, 53, 56 Remuneration
OECD (Organisation for Economic Co-operation and components of 72–73, 74
Development) 53, 54–55, 66, 67, 72 defining 70–71
Open source software 119, 126 estimation process 67–68
Organisation for Economic Co-operation and Development indicators 29, 65–66
(OECD) 53, 54–55, 66, 67, 72 non-health activities of health service personnel 68
Outflows of health workers, see Exits from the workforce non-observed payments 68–69
special groups 68
Renewal and loss indicators 29
P Resource flows, health system 71–72
Pacific Islands, out-migration 54–55 Resources for HRH systems 3–4, 5, 7, 33, 114
PAHO, see Pan American Health Organization (PAHO) Retention of health workers (see also Exits from the workforce)
Pakistan 39, 95–96
international out-migration 53 Retirement of health workers 52–53, 56, 57
out-migration 54 Romania
Pan American Health Organization (PAHO) census data 105
health spending measurement 72 educational attainment 109
HRH observatories 162–163 gender analysis 106, 107, 109
Partners for Health Reformplus (PHRplus) 83–84 Rural/urban distribution, see Geographical distribution of health
Performance indicators, see Indicators workers
Performance of workforce 8–9, 39, 80, 95–96, 133–134 Russian Federation, labour data estimation 69
Peru, health-care expenditure 73 Rwanda
Philippines absenteeism in the workforce 132
census data 105 census data 105
educational attainment 109 educational attainment 109
gender analysis 106, 107, 109 gender analysis 107, 109
international out-migration 40 health facility assessment 32
occupational transition 53 health workers salaries 32
out-migration 54 qualitative research 132, 136, 137, 138, 140, 143
PHRplus (Partners for Health Reformplus) 83–84 workforce entry 43
Poland, out-migration 54 workforce mortality 55
Policy-making 159–169
Political factors 46, 161
Population census data 21, 27 S
distribution of health workers 31 SAM (Service Availability Mapping) 83
for gender analysis 104–111 Sampling methods 81, 84, 85, 99
for workforce analysis 30, 152 Self-employed income 70–71
Portugal, health accounts 74 Service Availability Mapping (SAM) 83
Pre-service training, see Education and training Service Provision Assessment (SPA) 83
Private sector Sexual diversity, see Gender distribution of health workers
facility-based assessments 81, 85 Shortages of health workers 3, 32, 90–91
workforce data 26, 46 Sierra Leone
176
Index
out-migration 55 United States Agency for International Development (USAID) 65

workforce mortality 55 United States President’s Emergency Plan for AIDS Relief
Situation Assessment of Human Resources in the Public Sector 4, 84
83–84 Urban/rural distribution, see Geographical distribution of health
Skills mix 26, 28, 32, 92, 151 workers
SNA (System of National Accounts) 63, 65, 67, 74 USAID (United States Agency for International Development)
Socioeconomic factors 27, 161 3, 65, 83
South Africa
census data 105
educational attainment 109 V
gender analysis 106, 107, 109 Viet Nam
international migration 53 census data 105
national accounts 75 educational attainment 109
out-migration 54 gender analysis 106, 107, 109
SPA (Service Provision Assessment) 83 geographical distribution of health workers 31
Specialisms, classification 16 out-migration 54
Sri Lanka, sectoral transitions 53 Vocational training, see Education and training
Standardization issues (see also Classification systems;
Indicators) 8, 21, 26
Stock-based estimates 51, 52
W
Sub-accounts 76 WHO, see World Health Organization (WHO)
Sub-Saharan Africa 3, 4, 57 Women (see also Gender distribution of health workers)
out-migration 54–55 Workforce, see Health workforce
Sudan Working hours
health workforce monitoring system 120–122 estimation 68, 69
HRH observatory 165 facility-based assessments 96–97
Support workers, see Non-clinical workers Working lifespan 8–9, 38, 49, 57, 76, 80, 98, 103, 106, 118
Survey data 21, 27 World Bank 3, 4, 105, 132, 140, 166
household 27, 30 expenditure data collection 65
labour force 30, 32–33, 67–68, 152 Quantitative Service Delivery Survey 84
System of Health Accounts 65–66, 72 World Health Organization (WHO) 3, 7, 11, 16, 21, 38, 49, 55,
System of National Accounts (SNA) 63, 65, 67, 74 121, 125, 166
Assessment of Human Resources for Health 84
Evidence-Informed Policy Network (EVIPNet) 163
T expenditure data collection 65, 66, 72, 76
Tanzania health workforce definition 13
out-migration 55 monitoring toolkit 4
workforce mortality 55 primary health-care services coverage 79
Thailand Service Availability Mapping (SAM) 83
census data 104, 105 The world health report 4, 8, 49
exits from the workforce 110
gender analysis 106, 107, 109, 110
Y
Training, see Education and training Yugoslavia, out-migration 54
Transitions within workforce (see also Exits from the workforce;
Migration of health workers) 49–53
geographical transitions within a country 52
Z
indicators 51–52 Zambia
occupational transition 50, 52, 53, 56 absenteeism and ghost workers 30, 153
between sectors 50, 56 active health workforce 88
Trinidad and Tobago, out-migration 54, 55 analysis of multiple data sources 152–154
Turnover of workforce, see Transitions within workforce distribution of health workers 93–94
out-migration 54
U Zimbabwe
Uganda international out-migration 53
census data 104, 105 out-migration 54
international out-migration 53
licensure information system 119–120
student nurses 120, 121
Unemployment of health workers 28, 52
United Kingdom
in-migration 54
international in-migration 53, 57–58
United States of America 39, 53, 68, 104, 133
recruiting back nurses 57
177
About the publishing agencies
The United States Agency for International Development (USAID) provides support in
developing countries to ensure that health systems are effective, efficient and equitable. An
important component of USAID’s strategy is to strengthen the ability of health workers to
provide quality health care, notably by supporting initiatives for addressing workforce devel-
opment and quality improvement. USAID works through cooperative agreements, contracts
and partnerships with international organizations and other donor agencies to implement and
maximize the effectiveness of its global health activities.
t For more information on USAID’s human resources for health (HRH) programme, visit:
www.usaid.gov/our_work/global_health/hs/techareas/workers.html
The World Bank is committed to helping client countries address and achieve the health-
related Millennium Development Goals. The Bank works with client countries’ governments to
develop sustainable projects based on sound policies and strategies. With the country’s own
vision for its development as the starting-point, the Bank is engaged in building sustainable
health systems, of which human resources are a main focus. Securing a qualified health work-
force is part of the Bank’s poverty alleviation strategy. Increasing the management and analysis
skills of health leaders also constitutes an important goal for strengthening health systems.
t For more information on the World Bank’s HRH programme, visit:
go.worldbank.org/XR4K48D5M0
The World Health Organization (WHO) provides leadership on global health matters, includ-
ing shaping the health research agenda and articulating evidence-based policy options. WHO
works with partners and countries to plan workforce strategies for health systems strength-
ening and priority health interventions; strengthen the information and knowledge base to
support decision-making for policies and programmes; and develop tools and guidelines for
building capacity in addressing workforce issues among countries and stakeholders.
t For more information on WHO’s HRH programme, visit: www.who.int/hrh
t For information on WHO regional office activities on HRH, visit:

www.who.int/hrh/activities/regional/en/index.html
178
Handbook on
Monitoring and
Evaluation of
Human Resources
for Health
A
skilled health worker can make the difference between life and death. It is
our job to assure our citizens that health workers will be deployed when
and where they are needed to save lives and that they possess skills
that are adequate wherever they work, in public, private or not-for-profit
establishments. This new Handbook is welcomed, as it gives us the tools we need to
actively monitor and better manage the workforce. The core and common methods
described here will help us all to enhance public confidence in the health system and
enable the health workforce to be at the right place at the right time to make a difference,
both for each of us as individuals and for our communities.
Sally K. Stansfield
Executive Secretary
Health Metrics Network
,6%1

Eng-1

Uploaded by

Copyright:

Available Formats

Eng-1

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Eng-1

Uploaded by

Copyright:

Available Formats

Handbook on

with special applications for

Edited by Mario R Dal Poz, Neeru Gupta,

ISBN 978 92 4 154770 3 (NLM classiﬁcation: W 76)

© World Health Organization 2009

List of contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii

Acronyms and abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii

1.5 Road map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

1.6 Further information and comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2 Boundaries of the health workforce: definition and classifications of health workers. . . . . . . . . 13

2.2 Who are health workers? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.4 Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Part II: MONITORING THE STAGES OF THE WORKING LIFESPAN. . . . . . . . . . . . . . . . . 23

3.2 Core indicators for HRH analysis: what needs to be monitored? . . . . . . . . . . . . . . . . . . 26

3.3 Overview of potential data sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

3.4 Illustrative analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

3.6 Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

4.2 Framework for monitoring entry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

4.4 Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

5 Monitoring health workforce transitions and exits . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

5.3 Indicators and measurement strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

5.4 Illustrative analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

5.5 Concluding remarks: implications for policy and planning . . . . . . . . . . . . . . . . . . . . . 57

Part III: MEASUREMENT STRATEGIES AND CASE STUDIES . . . . . . . . . . . . . . . . . . . . 61

6.2 What should be measured . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

6.3 Approaches to measuring HRH expenditure . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

6.4 Measurement frameworks and applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

6.5 Summary, conclusions and further developments . . . . . . . . . . . . . . . . . . . . . . . . . 76

7 Use of facility-based assessments in health workforce analysis . . . . . . . . . . . . . . . . . . . . 79

7.3 Overview of key HFA methodologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

7.4 Some limitations of HFA methodologies for HRH data . . . . . . . . . . . . . . . . . . . . . . . 85

7.5 Empirical examples based on HFA data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

7.6 Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

8.2 Importance of gender considerations in health workforce analysis . . . . . . . . . . . . . . . 103

8.3 Using census data for health workforce analysis . . . . . . . . . . . . . . . . . . . . . . . . . 104

8.4 Empirical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

8.5 Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .110

9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .113

9.3 Country case studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .119

9.4 Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

10.2 Qualitative methods: their value and potential. . . . . . . . . . . . . . . . . . . . . . . . . . 130

10.3 Issues in the design and implementation of a qualitative study . . . . . . . . . . . . . . . . . 132

10.4 Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143

11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .147

11.5 Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

Part IV: DATA DISSEMINATION AND USE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

12.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159

12.4 Opportunities and directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

About the publishing agencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .176

Table 2.1 Framework for deﬁning the health workforce. . . . . . . . . . . . . . . . . . . . . . . . . . . . .14

Box 1.1 Deﬁning monitoring and evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Carissa F. Etienne Yaw Ansu Gloria D. Steele

We would like to acknowledge a number of people and organizations whose contributions

Ghanim Mustafa Alsheikh, World Health Organization Gustavo Nigenda,

Danny de Vries, Capacity Project/ Célia Regina Pierantoni,

David B Evans, World Health Organization Shomikho Raha, World Bank

Nancy Fronczak, Felix Rigoli, Pan American Health Organization/