Supplementary data
Appendix 1: Search process ...................................................................................................... 3
Appendix 2: Survey inclusion flow chart ..................................................................................... 4
Appendix 3: Data availability ...................................................................................................... 5
Appendix 4: Country-specific sampling methods ........................................................................ 6
Appendix 5: Diabetes biomarker devices by country .................................................................28
Appendix 6: Blood pressure measurement devices by country .................................................30
Appendix 7: Cholesterol measurement devices by country .......................................................32
Appendix 8: National definitions of area residence ....................................................................33
Appendix 9: Summary of diabetes performance measures .......................................................37
Appendix 10: Unavailability of performance measures by country .............................................38
Appendix 11: Details on missing data by country ......................................................................40
Appendix 12: Map of included countries ....................................................................................42
Appendix 13: Sample characteristics ........................................................................................43
Appendix 14: Rural versus urban residence among study sample ............................................44
Appendix 15: Proportion of diabetes population living in rural or urban areas ...........................46
Appendix 16: Number of respondents with diabetes and diabetes prevalence by country .........48
Appendix 17: Age-adjusted proportion of individuals with diabetes achieving performance
measures ..................................................................................................................................50
Appendix 18: Population of individuals achieving and not achieving goal ..................................51
Appendix 19: Age-adjusted proportion of individuals with diabetes achieving performance
measures by sex .......................................................................................................................52
Appendix 20: Differences in achievement of ever tested among rural versus urban (reference
category) populations with diabetes by country .........................................................................53
Appendix 21: Differences in achievement of awareness of diagnosis among rural versus urban
(reference category) populations with diabetes by country ........................................................54
Appendix 22: Differences in achievement of glucose-lowering medication among rural versus
urban (reference category) populations with diabetes by country ..............................................55
Appendix 23: Differences in achievement of blood pressure-lowering medication among rural
versus urban (reference category) populations with diabetes by country...................................56
Appendix 24: Differences in achievement of glycemic control among rural versus urban
(reference category) populations with diabetes by country ........................................................57
Appendix 25: Differences in achievement of blood pressure control among rural versus urban
(reference category) populations with diabetes by country ........................................................58
Appendix 26: Differences in achievement of AB control among rural versus urban (reference
category) populations with diabetes by country .........................................................................59
1
Appendix 27: Sensitivity analyses 1 (unadjusted proportions) ...................................................60
Appendix 28: Sensitivity analyses 2 (less strict glycemic target of HbA1c <8.0% [FPG <9.2
mmol/L]) – Differences in achievement of diabetes performance measures among rural versus
urban (reference category) populations .....................................................................................61
Appendix 29: Sensitivity analyses 3 (population weights) – Differences in achievement of
diabetes performance measures among rural versus urban (reference category) populations ..62
Appendix 30: STROBE checklist ...............................................................................................63
Supplementary references ........................................................................................................66
2
Appendix 1: Search process
The following is our comprehensive, two-step methodology for identifying, accessing, and
pooling available national health surveys:
1. We identified all LMICs in which a World Health Organization (WHO) Stepwise Approach to
Surveillance (STEPS) survey had been conducted.1 We preferred STEPS surveys as they
use a standardized questionnaire template and represent the WHO’s official framework for
conducting surveillance for noncommunicable diseases (NCD) at the population level.2,3
Prior to 2019, we requested each STEPS survey from a list maintained on the WHO
website.4 The research team contacted the responsible party for each survey based on the
information provided on this website. If the contact information was outdated or unavailable,
the research team relied on publications utilizing STEPS data and electronic searches of the
survey or contact name. For the Caribbean region, country involvement was facilitated by
the Caribbean Public Health Agency (CARPHA). Beginning in 2019, we downloaded STEPS
surveys from the WHO NCD Microdata Repository.5 The final search date for STEPS
surveys was April 1, 2021.
2. For countries in which no eligible STEPS survey was available, we conducted a systematic
Google search in May 2020 to identify additional potentially eligible surveys. Our search
strategy is described below:
Search engine: Google
Search terms: “[country name]” AND (“population-based” OR household) AND (“blood
glucose” OR “plasma glucose” OR “blood sugar” OR hemoglobin OR haemoglobin OR A1c
OR HbA1c OR A1C OR Hb1c OR Hba1c OR HGBA1C OR “blood pressure” OR
hypertension OR hypertensive OR cholesterol OR LDL OR HDL OR lipoprotein OR
triglycerides OR triglyceride OR lipid OR lipids)
Number of hits reviewed: We reviewed the hits until we identified an eligible survey. If we
reviewed the first 50 hits (10 hits per page/5 pages reviewed) without identifying an eligible
survey, we stopped reviewing the hits and determined the country to not have any eligible
non-STEPS surveys.
Search date: April 8, 2020
3
Appendix 2: Survey inclusion flow chart
STEPS surveys
n=19 high-income
countries
n=31 conducted
before 2008
n=10 subnational
n=37 no data on rural
diabetes care
Non-STEPS surveys
132 STEPS
surveys identified
90 countries in
systematic search
35 eligible STEPS
surveys
49 eligible nonSTEPS surveys
33 STEPS surveys
included
9 non-STEPS
surveys included
n=2 no response to
clarifying emails
42 total surveys
included
4
n=40 no or unclear
data on rural
diabetes care
Appendix 3: Data availability
The generic versions of the World Health Organization STEPwise approach to
noncommunicable disease surveillance (WHO STEPS) instrument are available online
(accessed June 29, 2021) at the following links:
Version 2.1:
https://www.who.int/ncds/surveillance/steps/STEPS_Instrument_v2.1.pdf
Version 3.2:
https://www.who.int/ncds/surveillance/steps/instrument/STEPS_Instrument_V3.2.pdf
Data included in this study are publicly available for 37 of the 42 included country surveys.
Microdata can be downloaded at the following links:
Chile National Health Survey: https://www.minsal.cl/estudios_encuestas_salud/
China Health and Nutrition Survey: https://www.cpc.unc.edu/projects/china
El Salvador 2015 Encuesta Nacional de Enfermedades Crónicas No Transmisibles en
Población Adulta de El Salvador (ENECA-ELS):
https://data.amerigeoss.org/gl/dataset/encuesta-nacional-de-enfermedades-cronicas
Indonesia Family Life Survey (IFLS): https://www.rand.org/well-being/social-and-behavioralpolicy/data/FLS/IFLS.html
Mexico National Survey on Health and Nutrition (ENSANUT):
https://ensanut.insp.mx/encuestas/ensanut2018/descargas.php
Namibia Demographic and Health (DHS) Survey:
https://dhsprogram.com/methodology/survey/survey-display-363.cfm
STEPS Microdata repository: https://extranet.who.int/ncdsmicrodata/index.php/catalog/STEPS
For data that are not publicly accessible and for which we have arranged specific data-use
agreements (surveys in Burkina Faso, Fiji, Iran, Romania, and South Africa), data will be made
available from the authors upon reasonable request and with permission of the data owners.
5
Appendix 4: Country-specific sampling methods
Note: In order to ensure accuracy in reporting, these sampling methods are verbatim from the
methods sections of the specified sources.
Afghanistan: STEPS 2018
In the sampling methodology districts are used as primary sampling units (PSUs),
villages/blocks are the SSUs, and households within districts serves as TSUs. Based on the
guidelines of the WHO, the total number of the PSUs within a sampling frame should be greater
than 100 among which 50-100 PSUs should be randomly selected. The total number of districts
in 34 provinces of Afghanistan is 417. From 417 districts 55districts were selected based on the
available resources using Stepwise-Approach XLs form.
The total sample size was distributed proportionate to the size of the districts, then the sample
size of the districts was divided by 15 (maximum number of the household to interviewed within
an EA) and number of EAs within each district was calculated. Using the EPI sampling frame
EAs were selected within each district. Within each EA the total number of the households were
calculated and it was divided to calculate the sampling interval. The household with each
randomly selected, within each household interview with a randomly selected male or female
members was conducted.
Age range of participants included: 18-69 years
Source: Afghanistan STEPS 2018 Report. Available at:
https://extranet.who.int/ncdsmicrodata/index.php/catalog/782
Algeria: STEPS 2016-2017
A multi-stage cluster sample of households. One individual within the age range of the survey
was selected per household. Analysis weights were calculated by taking the inverse of the
probability of selection of each participant. These weights were adjusted for differences in the
age-sex composition of the sample population as compared to the target population.
Different weight variables are available per Step:
wStep1 - for interview data
wStep2 - for physical measures
wStep3 - for biochemical measures
This allows for differences in the weight calculation for each Step of the survey as the age-sex
composition of the respondents to each Step can differ slightly due to refusal or drop out.
Additionally, some countries perform subsampling for Step 2 and/or Step 3. When no
subsampling is done and response rates do not differ across Steps of the survey, the 3 weight
variables will be the same.
Age range of participants included: 18-69 years
Source: no report or fact sheet available. Sampling information obtained from:
https://extranet.who.int/ncdsmicrodata/index.php/catalog/91/study-description
Armenia: STEPS 2016
The STEPS survey of non-communicable disease (NCD) risk factors in Republic of Armenia
was carried out from September 2016to December 2016. The Republic of Armenia carried out
Step 1, Step 2, and Step3. Socio demographic and behavioral information was collected in Step
1. Physical measurements such as height, weight and blood pressure were collected in Step 2.
Biochemical measurements were collected to assess blood glucose and cholesterol levels and
urine analyze to assess salt intake levels in Step 3. The survey was a population-based survey
of adults aged 18-69A cluster sample design was used to produce representative data for that
6
age range in Armenia. A total of2349adults participated in the survey. The overall response rate
was42%.
Age range of participants included: 18-69 years
Source: Armenia STEPS Fact Sheet. Available at:
https://extranet.who.int/ncdsmicrodata/index.php/catalog/102
Azerbaijan: STEPS 2017
A multi-stage cluster sample of households. One individual within the age range of the survey
was selected per household. Analysis weights were calculated by taking the inverse of the
probability of selection of each participant. These weights were adjusted for differences in the
age-sex composition of the sample population as compared to the target population.
Different weight variables are available per Step:
wStep1 - for interview data
wStep2 - for physical measures
wStep3 - for biochemical measures
This allows for differences in the weight calculation for each Step of the survey as the age-sex
composition of the respondents to each Step can differ slightly due to refusal or drop out.
Additionally, some countries perform subsampling for Step 2 and/or Step 3. When no
subsampling is done and response rates do not differ across Steps of the survey, the 3 weight
variables will be the same.
Age range of participants included: 18-69 years
Source: no report or fact sheet available. Sampling information obtained from:
https://extranet.who.int/ncdsmicrodata/index.php/catalog/127/studydescription#page=overview&
tab=study-desc
Bangladesh: STEPS 2018
Sampling Procedure
A multistage complex sampling design was used to produce representative data for that age
range in Bangladesh.
Response Rate
The overall response rate was 83.8%.
Weighting
Analysis weights were calculated by taking the inverse of the probability of selection of each
participant. These weights were adjusted for differences in the age-sex composition of the
sample population as compared to the target population.
Different weight variables are available per Step:
wStep1 - for interview data
wStep2 - for physical measures
wStep3 - for biochemical measures
This allows for differences in the weight calculation for each Step of the survey as the age-sex
composition of the respondents to each Step can differ slightly due to refusal or drop out.
Additionally, some countries perform subsampling for Step 2 and/or Step 3. When no
subsampling is done and response rates do not differ across Steps of the survey, the 3 weight
variables will be the same.”Age range of participants included: 25 to 69 years
Source: https://extranet.who.int/ncdsmicrodata/index.php/catalog/770/studydescription#page=overview&tab=study-desc
Source: National Institute of Population Research and Training (NIPORT), Mitra and Associates,
and ICF International. 2013. Bangladesh Demographic and Health Survey 2011. Dhaka,
Bangladesh and Calverton, Maryland, USA: NIPORT, Mitra and Associates, and ICF
International.
7
Belarus: STEPS 2016-17
The sampling frame is a collection of data and materials from which are selected for the survey.
The optimal sampling frame should be complete, accurate and current. Best of all, the above
criteria are met by the results of the population census, which became the basis for constructing
the sample for the STEPS study. Census population represents a representative territorial
sampling frame in the form a hierarchical set of parcels grouped in a certain way. Plots
censuses are, on average, about the same size. For each site there is a schematic map that
provides a clear, non-overlapping demarcation of geographic districts, as well as information on
the population and the number of households.
The largest in size is the census area, which includes several instructor sites. The smallest unit
in the hierarchical structure of parcels by censuses - enumeration areas.A positive aspect of
using enumeration areas as primary sampling units (PSUs) is that they have a small and
approximately the same size (each includes about 100 HHs on average). Consequently this, the
PSU is a territory within which it is possible to effectively organize field work. To conduct a
population census, the territory of the Republic of Belarus was divided into almost 32 thousand
enumeration areas. Due to the fact that the last population census in the Republic of Belarus
was carried out in 2009, to update the sample, the current data of polyclinics were used,
medical outpatient clinics, FAPs and rural Soviet accounting in rural areas.
Age range of participants included: 18-69 years
Source: Translated directly from the Belarus STEPS 2016 report. Available at:
https://extranet.who.int/ncdsmicrodata/index.php/catalog/100/related_materials
Benin: STEPS 2015
“The STEPS survey on risk factors for non-communicable diseases in Benin was conducted
from October to December 2015. It was a population-based survey of adults aged 18 to 69
years. A 3-stage sampling frame was used to produce representative data for this age group in
Benin. The information required for the investigation was collected electronically using a manual
device. The survey was implemented by the National Program for the Fight against NonCommunicable Diseases (PNLMNT) of the Ministry of Health of Benin. A total of 5,126 adults
participated in the STEPS survey conducted in Benin. The overall response rate was 98.6%.
The 1st survey took place in 2008. A third survey is planned for 2020 if the financial situation
allows it.”
Age range of participants included: 18-69 years
Source: Translated directly from the Benin STEPS 2015 report. Available at:
https://extranet.who.int/ncdsmicrodata/index.php/catalog/107/download/1044
Bhutan: STEPS 2014
Sampling procedure
To achieve a nationally representative sample, a multistage sampling method was used to
select enumeration areas, households and eligible participants at each of the selected
households in three stages. The 2005 National Census was chosen as the basis for the
sampling frame, with “Geogs” (blocks) in rural areas and towns in urban areas forming the
primary sampling units (PSUs). Since the population distribution for urbanicity is 70:30
(rural:urban), 63 PSUs in rural and 14 PSUs in urban areas were chosen. PSUs were selected
through the probability proportionate to size (PPS) sampling using the number of households in
each PSU. Two secondary sampling units (SSUs) for every rural PSU and 4 SSUs for every
urban PSU were selected. This led to the selection of 126 SSUs from rural and 56 SSUs from
urban areas. This was also carried out by PPS sampling, using the number of households in
each SSU. A total of 16 households from each SSU (both rural and urban) were selected using
systematic random sampling. The sampling frame for this was the list of households with a
8
unique identification number (ID) developed by the enumerators for the survey. At the
household level, the Kish sampling method was used to randomly select one eligible member
(aged 18–69 years) of the household for the survey. The Kish method ranks eligible household
members in order of decreasing age, starting with males and then females, and randomly
selects a respondent using the automated program for Kish selection in the handheld personal
digital assistant (PDA).
Age range of participants included: 18-69 years
Source: Bhutan STEPS report. Available at:
https://www.who.int/ncds/surveillance/steps/bhutan/en/
Burkina Faso: STEPS 2013
“Sampling methodology: The study was conducted on a sample obtained from a three-stage
cluster stratified as recommended by the WHO for STEPS screening surveys. risk factors for
noncommunicable diseases. The sampling frame used was that derived from the general
census of the population and habitat 2006 (RGPH 2006) and updated in 2010 during the survey
Demographic and Health Survey of Burkina Faso (EDS-BF, 2010). This update concerned the
enumeration areas (EAs) that correspond to the cluster as part of this study.
Selection of clusters: The choice of clusters was made according to a systematic random
selection proportional to their size (in number of households) within strata (regions). To do this
clusters were organized by stratum and place of residence (urban / rural). A total of 240 clusters
of which 185 were in rural areas and 55 in urban areas were selected for the investigation.
Selection of households: Households were randomly drawn after an enumeration exhaustive list
of all households in the cluster. A draw tool designed on Excel by the team. The technique was
used in the field for selecting households to investigate. In total, 20 households in clusters were
selected to participate in the study.
Selection of individuals: The choice of individuals was made randomly using Kish's method. In
total, an individual aged 25 to 64 living in a selected household was fired for participate in the
survey.”
Age range of participants included: 25-64 years
Source, translated from: Rapport de l’enquete national sur la prevalence des principaux facteurs
de risques communs aux maladies non transmissibles au Burkina Faso Enquete STEPS 2013.
Available at: http://www.who.int/chp/steps/burkina_faso/en/.
Cambodia: STEPS 2010
“The initial planned sample size was designed to involve 5,760 persons in accordance with the
NCD multi-stage cluster survey method (1.5 design effect, 95% confidence interval, 5% margin
or error, and 50% baseline levels of the indicators) in order to provide an equivalent distribution
of the participants in regards to age groups and gender after taking into consideration that the
estimated potential rate for non-response in each group and refusals in the nest stages would
equal to 20%. Estimates were obtained for each of the following eight age/sex groups: men
ahed 25-34 years, 35-44 years, 45-54 years, and 55-64 years; and women aged 25-34 years,
35-44 years, 45-54 years, and 55-64 years.
The survey was designed to cover all geographical areas of Cambodia and a 3-stage sampling
process as part of the multi-stage cluster sampling was carried out to randomly select the target
population: random selection of communes (Khum in rural areas and its equivalent Sangkat in
urban area) as primary sampling unit (PSU), followed by villages (Phum) for the second
sampling unit (SSU), and by households for the elementary units (EU). Finally, all members of
the randomly chose households aged 25-64 years were invited to participate in this survey. The
selection process was performed identically for urban and rural areas in order to get a selfweighted estimate for the whole population of the country. A total of 180 clusters with 34
clusters from the urban area and 146 clusters from the rural area were randomly selected.”
9
Age range of participants included: 25-64 years
Source: Cambodia STEPS 2010 survey report. Available at:
https://www.who.int/ncds/surveillance/steps/cambodia/en/
Chile: NHS 2009-10
“The sampling frame was constituted from the Population and Housing Census 2002. The
design of the study was transversal, with a random sample of complex type households
(stratified and multi-stage by clusters) with national, regional and area representation rural /
urban. The target population was adults older than or equal to 15 years. The survey had a
response rate in the eligible population of 85%. The refusal rate was of 12%. 5,434 people were
interviewed. A nurse performed clinical and examinations to 5,043 participants and 4,956
accepted laboratory tests (blood and urine). The total sample loss of the oversized sample was
28% (this including rejection, non-contact and other causes of random loss). The raw sample
was designed with overrepresentation of some population groups (older adults, regions other
than the Metropolitan Region and rural areas) to increase sample efficiency and homogenize
the accuracy of the estimators. The expansion of the sample data is because it grants each
participant the weight that corresponds to it according to the design sample and at the same
time corrects the distortion of the raw sample, making it coincide with the census population
projection for January 2010 for Chilean adults over 15 years of age.“
Age range of participants included: 15 years or older
Source, translated from: Resumen Ejecutivo: Encuesta Nacional de Salud ENS Chile 2009-10.
Available at: http://epi.minsal.cl/encuesta-ens-anteriores/.
China: CHNS 2009
“The China Health and Nutrition Survey is a longitudinal study across 228 communities within
nine provinces of China. Surveys began in 1989, with subsequent surveys every 2–4 years, for
a total of nine rounds between 1989 and 2011. The China Health and Nutrition Survey was
designed to provide representation of rural, urban and suburban areas varying substantially in
geography, economic development, public resources and health indicators,13 and it is the only
large-scale, longitudinal study of its kind in China. The original survey in 1989 used a
multistage, random cluster design in eight provinces (Liaoning, Jiangsu, Shandong, Henan,
Hubei, Hunan, Guangxi and Guizhou) to select a stratified probability sample; a ninth province,
Heilongjiang, was added in 1997 using a similar sampling strategy. Essentially, two cities (one
large and one small city—usually the provincial capital and a lower income city) and four
counties (stratified by income: one high, one low and two middle income counties) were
selected in each province. Within cities, two urban and two suburban communities were
selected; within counties, one community in the capital city and three rural villages were chosen.
Twenty households per community were then selected for participation. The study met the
standards for the ethical treatment of participants and was approved by the Institutional Review
Boards of the University of North Carolina at Chapel Hill and the Institute of Nutrition and Food
Safety, Chinese Center for Disease Control and Prevention.”
Age range of participants included: all ages
Source: Attard, Samantha M.; Herring, Amy H.; Wang, Huiling; Howard, Annie Green;
Thompson, Amanda L.; Adair, Linda S.; Mayer-Davis, Elizabeth J.; & Gordon-Larsen, Penny.
(2015). Implications of Iron Deficiency/Anemia on the Classification of Diabetes Using HbA1c.
Nutrition & Diabetes, 5, e166.
El Salvador: ENECA-ELS 2015
The sample selection was carried out in a two-stage and probabilistic manner; the sample
framework was the population census conducted in El Salvador in 2007. A cartographic update
of the census segments conducted by Digestyc in 2015 was carried out and these were divided
10
into clusters, which were composed of 12 to 25 dwellings and finally to all persons in the
dwellings that met the inclusion criteria.
The data collection process was carried out in two stages: in the first stage, each of the selected
houses was visited, where all the members of the household who met the inclusion criteria were
listed in a family file. The objective of the study was explained to the eligible persons and they
were given the consent form to read it; the document was read to those who had difficulty
reading and it was explained to them that they could withdraw from the study at any time if they
chose to do so. Once the reading was finished, they were invited to participate in the study;
those who accepted signed the informed consent form or placed their fingerprint, and then
proceeded to conduct the survey.
If a person was ill at the time of the survey or had been diagnosed during the application of the
survey, he/she was referred to a health facility. The actual fieldwork was conducted from
October 2014 to March 2015. The second measurement was performed with a minimum interval
of three months after the first one, in order to confirm the CKD. Thus in January 2015, the
remeasurement was carried out, ending in March 2015.Out of a total of 1032 persons to be
remeasured, 725 underwent such remeasurement. After the study, 4817 questionnaires that
met all the required methodological conditions were completed. These were used to form the
database for the analysis of the results. Estimates were made according to sex, 3 age groups
(20 to 40, 41 to 60 and 60 and over), urban and rural area of residence and Minsal health
regions.
Age range of participants included: ≥20 years
Source: Ministerio de Salud, 2015. Encuesta Nacional de Enfermedades Crónicas no
transimibles en Población Adulta de El Salvador. San Salvador.[Translated]
Ethiopia: STEPS 2015:
According to the WHO step-wise approach to the surveillance of NCD risk factors, a communitybased cross sectional study was carried out.
The target population for this survey included all men and women age15-69 years old who have
been living at their place of residence for at least six months. This target population included all
people who consider Ethiopia to be their primary place of residence. This definition included
those individuals residing in Ethiopia regardless of their citizenship status. . People with the
following characteristics were not included: those who were not a permanent resident of
Ethiopia, and those who were institutionalized including people residing in hospitals, prisons,
nursing homes, and other similar institutions or residents whose primary residences are military
camps or dormitories. Furthermore, critically ill, mentally disabled and those with some type of
physical disability that is not suitable for physical measurement were excluded from this study.
In general, the target population of the study included individuals 15-69 years old and residing in
all geographic areas of the country.
Age range of participants included: 15 to 69 years
Source: Ethiopia STEPS 2015 Report. Available at:
https://extranet.who.int/ncdsmicrodata/index.php/catalog/794
Fiji: EHS 2009
“The sample frame (188 800 people aged ≥40 years; 50.3% female; 49.4% Melanesian Fijian,
44.9% Indo-Fijian, and 5.7% of other ethnicity; 43.2% rural dwellers) included all 8 provinces of
Viti Levu, Fiji’s main island, where 79.1% of the total population resides. Using an anticipated
prevalence of vision impairment of 11.0% in the target population (actual was 11.4%; 95%
confidence interval [CI] = 9.9% to 13.2%), absolute precision of ±2.2% (20% relative difference),
with 95% confidence, a design effect of 1.4 and a response rate of 80%, the sample size was
11
determined to be 1354 persons. From the sample frame, 34 clusters of 40 people were
required. Across Viti Levu, the clusters were selected through probability proportionate to size
sampling, using national census data.”
Age range of participants included: 40 to 90 years
Source: pasted verbatim from email exchange with study team.
Additional reference: Brian G, Ramke J, Maher L, Page A, Szetu J. The prevalence of diabetes
among adults aged 40 years and over in Fiji. N Z Med J. 2010; 123(1327):68–75. PMID:
21358785
Georgia: STEPS 2016
“The STEPS survey of noncommunicable disease (NCD) risk factors in Georgia was carried out
from June 2016 to September 2016. Georgia carried out Step 1, Step 2 and Step 3. Socio
demographic and behavioural information was collected in Step 1. Physical measurements such
as height, weight and blood pressure were collected in Step 2. Biochemical measurements were
collected to assess blood glucose and cholesterol levels in Step 3. The survey was a
population-based survey of adults aged 18-69. A Multi-stage cluster sampling design was used
to produce representative data for that age range in Georgia. A total of 5554 adults participated
in the survey. The overall response rate was 75.7%.”
Age range of participants included: 18 to 69 years
Source: Georgia STEPS Survey 2016 Fact Sheet.
Available at: http://www.who.int/chp/steps/georgia/en/.
Guyana: STEPS 2016
“A response rate of 66.68% will be selected based on the experience and response rates of
other surveys over the years such as the recent Demographic Health Survey 2009. [...] STEPS
3 involve taking blood samples from a proportion of the sample, in this case 50% of the sample,
in order to measure raised blood glucose levels and abnormal blood lipids. [...] The STEPS
sample will be prepared by the Bureau of Statistics Guyana following the recommended STEPS
sample methodology. A multi-stage cluster sampling design will be used. Guyana is divided into
10 administrative regions and within the administrative regions there are seven towns and each
region is further divided into enumeration districts. For the STEPS survey 288 enumeration
districts will be selected using the population probability sampling method and from each
enumeration district 12 households will be selected giving a total sample size of 3456. Further at
the household level each participant will be randomly selected by the electronic tablet. For
STEP 3 50% of the sample will be randomly selected to participate. A re-listing of some
households may also be necessary, such as those interior region locations, in which case in
addition to household listings, enumeration districts maps will also be provided so that a relisting can be done where required.”
Age range of participants included: 18 to 69 years
Source: STEPwise Approach to Chronic Disease risk factor surveillance (STEPS): Guyana’s
Implementation Plan. June 20, 2016. Ministry of Public Health, Guyana.
India: NFHS 2015-16
“The NFHS-4 sample was designed to provide estimates of all key indicators at the national and
state levels, as well as estimates for most key indicators at the district level (for all 640 districts
in India, as of the 2011 Census). The total sample size of approximately 572,000 households for
India was based on the size needed to produce reliable indicator estimates for each district and
for urban and rural areas in districts in which the urban population accounted for 30-70 percent
of the total district population. The rural sample was selected through a two-stage sample
design with villages as the Primary Sampling Units (PSUs) at the first stage (selected with
probability proportional to size), followed by a random selection of 22 households in each PSU
12
at the second stage. In urban areas, there was also a two-stage sample design with Census
Enumeration Blocks (CEB) selected at the first stage and a random selection of 22 households
in each CEB at the second stage. At the second stage in both urban and rural areas,
households were selected after conducting a complete mapping and household listing operation
in the selected first-stage units.”
Age range of participants included: women 15-49 years, men 15-54 years
Source: Ministry of Health and Family Welfare (MoHFW) - Government of India. India - National
Family Health Survey 2015-2016. Report generated on: February 7, 2018.
Indonesia: IFLS 2014-15
“Because it is a longitudinal survey, IFLS5 drew its sample from IFLS1, IFLS2, IFLS2+, IFLS3
and IFLS4. The IFLS1 sampling scheme stratified on provinces and urban/rural location, then
randomly sampled within these strata (see Frankenberg and Karoly, 1995, for a detailed
description). Provinces were selected to maximize representation of the population, capture the
cultural and socioeconomic diversity of Indonesia, and be cost effective to survey given the size
and terrain of the country. For mainly cost-effectiveness reasons, 14 of the then existing 27
provinces were excluded.3 The resulting sample included 13 of Indonesia’s 27 provinces
containing 83% of the population: four provinces on Sumatra (North Sumatra, West Sumatra,
South Sumatra, and Lampung), all five of the Javanese provinces (DKI Jakarta, West Java,
Central Java, DI Yogyakarta, and East Java), and four provinces covering the remaining major
island groups (Bali, West Nusa Tenggara, South Kalimantan, and South Sulawesi).
Within each of the 13 provinces, enumeration areas (EAs) were randomly chosen from a
nationally representative sample frame used in the 1993 SUSENAS, a socioeconomic survey of
about 60,000 households. The IFLS randomly selected 321 enumeration areas in the 13
provinces, over-sampling urban EAs and EAs in smaller provinces to facilitate urban-rural and
Javanese–non-Javanese comparisons.
Within a selected EA, households were randomly selected based upon 1993 SUSENAS listings
obtained from regional BPS office. A household was defined as a group of people whose
members reside in the same dwelling and share food from the same cooking pot (the standard
BPS definition). Twenty households were selected from each urban EA, and 30 households
were selected from each rural EA. This strategy minimized expensive travel between rural EAs
while balancing the costs of correlations among households. For IFLS1 a total of 7,730
households were sampled to obtain a final sample size goal of 7,000 completed households.
This strategy was based on BPS experience of about 90% completion rates. In fact, IFLS1
exceeded that target and interviews were conducted with 7,224 households in late 1993 and
early 1994. In IFLS1 it was determined to be too costly to interview all household members, so a
sampling scheme was used to randomly select several members within a household to provide
detailed individual information.”
Age range of participants included: all ages
Source: Strauss, J., F. Witoelar, and B. Sikoki. “The Fifth Wave of the Indonesia Family Life
Survey (IFLS5): Overview and Field Report”. March 2016. WR-1143/1-NIA/NICHD.
Iran: STEPS 2016
“The sampling part, which includes determining the sample size and the cluster head, belongs
to the pre-study phase and was planned in the form of a specific protocol for sample size and
statistical sampling. All experts in the quality control team supervised the finding of samples and
cluster heads.
In order to estimate the prevalence rate of the risk factors for non-communicable diseases in the
country in 1395, a sampling method proportionate to the population was used, which is a
common approach in survey studies. Therefore, the selected sample size was proportionated to
the population of that province. On the other hand, for estimating the prevalence of the risk
13
factors in the province, in order to be on the safe side, the smallest sample size for achieving
the predicted rates was calculated at 95%. This rate was equal to 384 samples, which was
selected as the smallest sample size in the least populated province, Ilam. The required sample
size for other provinces was therefore calculated according to the population of that province
proportionate to the population of the reference province, Ilam. Besides, to control the nonresponse error, 10% was added to the calculated sample size in each province. In order to
decrease costs and increase efficiency, for provinces with 800 samples or more, weights were
given to their samples. Weight-giving is an effective method used in surveys in order to
decrease the sample size. This was achieved in the selected provinces by considering the
calculated sample size as half and the sampling weight as double. The total sample size was
calculated to be 30150 and to achieve this sample size, sampling from 3015 clusters was
required.”
Age range of participants included: 18 and older
Source: Iran STEPS 2015 report.
Available at: https://www.who.int/ncds/surveillance/steps/STEPS_2016_Atlas_EN.pdf?ua=1
Iraq: STEPS 2015
“The sample frame consisted of the population of Iraq of (18+) years for both sexes residing in
the urban and rural area. It was based on the results of listing and numbering operation for the
year 2009 that covered all governorates. Due to the unstable conditions at the time of the
survey three governorates (Naynawa, Salahaddin and Al-Anbar) were excluded. A major
challenge confronted was the late demographic change due to population movement,
displacement and migration. All permanent residents of (18+) years of age, who were resident
in Iraq within one month at the time of implementation of the survey were considered eligible.
A cross‐sectional community based survey covering 15 governorates in Iraq. A Multi-stage
cluster sampling technique was depended to select the minimum representative sample size to
estimate the prevalence of the risk factors of noncommunicable disease through direct
interview, physical examination and laboratory examination of blood samples of study
participants. A total of 412 clusters were randomly selected each contain ten households. One
subject from each household was randomly selected using KISH table to participate in the
survey with a total sample size of 4120. The Sample was designed to provide estimates on a
number of indicators on the situation of Noncommunicable diseases risk factors in Iraq at the
national level. A national based rather than a governorate based sample is selected. A multi
stage cluster sampling was used with stratification to urban and rural areas. Primary sampling
units (PSUs) were the blocks, which consisted of 70 households or more before selection.”
Age range of participants included: 18 years and older
Source: Iraq STEPS 2015 report.
Available at: https://www.who.int/ncds/surveillance/steps/Iraq_2015_STEPS_Report.pdf
Jordan: STEPS 2019
A national cross-sectional survey was conducted adopting a two-stage stratified-cluster
sampling design. The margin error was (5%) and the confidence level was set at 95%. The
Jordan Population and Housing Census 2015 was used as a sampling frame for Jordanians. A
sample of 3000 households was randomly drawn to represent the Jordanian population. It was
designed in a probability proportional to size (PPS) way to provide valid and reliable survey
estimates across the entire Kingdom of Jordan - rural and urban areas, the twelve governorates
and the smaller communities within. The sample also ensured reliable estimates in terms of
geographical distribution, where Jordan was divided into three regions; north, centre, and south,
also at governorate level. The north of Jordan covered Ajloun, Irbid, Jerash, and Mafraq, the
centre region covered Amman, Balqa, Madaba, and Zarqa, and the south region covered
Aqaba, Karak, Ma’an, and Tafieleh. Furthermore, each governorate was subdivided into area
14
units called census blocks, which were the Primary Sampling Units (PSU-Blocks) for this survey
(on average a PSU comprises 50-70 households). The PSU-Blocks were then regrouped to
form clusters. From each PSU, eight households were randomly drawn with an equal probability
systematic selection. A household was defined as a group of people living in the same dwelling
space who eat meals together, acknowledging the authority of a man or a woman as the head
of the household. After the household selection and obtaining the permission of household
residents to participate in the survey, all the eligible household members were entered into the
STEPS program, which ran a random selection to choose one member household.
Age range of participants included: 18 to 69 years
Source: Jordan STEPS 2019 Report. Available at:
https://extranet.who.int/ncdsmicrodata/index.php/catalog/853
Kenya: STEPS 2015
“The 2015 Kenya STEPs survey was a national cross-sectional household survey designed to
provide estimates for indicators on risk factors for non-communicable diseases for persons age
18 – 69 years. The sample was designed with a sample size of 6,000 individuals to allow
national estimates by sex (male and female) and residence (urban and rural areas). The survey
used the fifth National Sample Surveys and Evaluation Programme (NASSEP V) master sample
frame that was developed and maintained by KNBS. The frame was developed using the
Enumeration Areas (EAs) generated from the 2009 Kenya Population and Housing Census to
form 5,360 clusters split into four equal sub-samples. A three-stage cluster sample design was
adopted for the survey involving selection of clusters, households and eligible individuals. In the
first stage, 200 clusters (100 urban and 100 rural) were selected from one sub-sample of
NASSEP V frame. A uniform sample of 30 households from the listed households in each
cluster was selected in the second stage of sampling. The last stage of sampling was done
using Personal Digital Assistants (PDAs) at the time of survey, where one individual was
randomly selected from all eligible listed household members using a programmed KISH
method of sampling.”
Age range of participants included: 18 to 69 years
Source: WHO: Kenya STEPwise Survey for Non Communicable Diseases Risk Factors 2015
Report. Available at: http://www.who.int/chp/steps/Kenya_2015_STEPS_Report.pdf?ua=1.
Kyrgyzstan: STEPS 2013
A multi-stage cluster sample of households. One individual within the age range of the survey
was selected per household.
Analysis weights were calculated by taking the inverse of the probability of selection of each
participant. These weights were adjusted for differences in the age-sex composition of the
sample population as compared to the target population.
Different weight variables are available per Step:
wStep1 - for interview data
wStep2 - for physical measures
wStep3 - for biochemical measures
This allows for differences in the weight calculation for each Step of the survey as the age-sex
composition of the respondents to each Step can differ slightly due to refusal or drop out.
Age range of participants included: 25 to 64 years
Source: no report or fact sheet available. Sampling information obtained from:
https://extranet.who.int/ncdsmicrodata/index.php/catalog/271/studydescription#page=overview&tab=study-desc
Lao People’s Democratic Republic: STEPS 2013
15
A multi-stage cluster sample of households. One individual within the age range of the survey
was selected per household. Analysis weights were calculated by taking the inverse of the
probability of selection of each participant. These weights were adjusted for differences in the
age-sex composition of the sample population as compared to the target population.
Different weight variables are available per Step:
wStep1 - for interview data
wStep2 - for physical measures
wStep3 - for biochemical measures
This allows for differences in the weight calculation for each Step of the survey as the age-sex
composition of the respondents to each Step can differ slightly due to refusal or drop out.
Additionally, some countries perform subsampling for Step 2 and/or Step 3. When no
subsampling is done and response rates do not differ across Steps of the survey, the 3 weight
variables will be the same.
Age range of participants included: 18 to 64 years
Source: no report or fact sheet available. Sampling information obtained from:
https://extranet.who.int/ncdsmicrodata/index.php/catalog/588/studydescription#page=sampling&tab=study-desc
Malawi: STEPS 2009
This was a national community based cross-sectional survey, using WHO STEPwise approach
for assessing risk factors for chronic non-communicable diseases. The approach includes the
use of a questionnaire for gathering demographic and behavioural information (Step 1), then
moving to physical measurements (Step 2) and then biochemistry tests (Step 3). In addition,
there are three modules of risk factor assessment, namely core, expanded and optional. The
STEPS Survey instrument was adapted and tested by the core team and data collectors.
Age range of participants included: 18 to 69 years
Source: Malawi Steps 2009 Report. Available at:
https://extranet.who.int/ncdsmicrodata/index.php/catalog/629
Mexico: ENSANUT 2018
The ENSANUT 2018-19 is a national, urban and rural probabilistic survey. The units of analysis
defined for the survey are the following: - Household is the set of people related by some
kinship or not who usually sleep in a dwelling under the same roof, benefiting from a common
income contributed by one or more of the household members. - Population aged 0 to 4 years
(preschoolers)- Population aged 5 to 9 years (schoolchildren)- Population aged 10 to 19 years
(adolescents)- Population aged 20 years and older (adults)- Utilizers
Once the PSUs and strata were constructed, the PSUs for the 2018-19 ENSANUT were
selected in two stages: first, INEGI selected a master sample of PSUs with probability
proportional to their number of dwellings in the year 2012, then, for the 2018-19 ENSANUT, a
subsample of PSUs with equal probability was selected within each stratum. Finally, in each
PSU, dwellings were selected with equal probability; on average, five dwellings were selected in
each PSU of the high urban stratum and 20 dwellings were selected in the PSUs of the rural
and urban complement strata.
Whenever possible, one adult, one adolescent, one schoolchild and one preschooler were
selected from each household with equal probability. Also, whenever possible, up to two users
of medical services during the last 15 days were selected in 40% of the dwellings, and in the
remaining 60% of the dwellings, up to one user was selected.
Age range of participants included: All ages
16
Source: ENSANUT Report. Available at:
https://ensanut.insp.mx/encuestas/ensanut2018/informes.php [Translated]
Moldova: STEPS 2013
“A total of 4807 randomly selected respondents participated in the survey. They were all aged
18–69 years, and the group comprised both sexes, as well as residents of all districts and the
territorial administrative unit “Gagauz-Yeri”, along with Chişinãu and Balti municipalities. The
survey did not cover the districts from the left bank of the Nistru River and the municipality of
Bender. A two-stage cluster sampling procedure was carried out to select randomly participants
from among the target population. Cluster sectors from the 2004 Moldova Population Census
were used as a basic unit. Given the differences in lifestyle and disease status between
populations in urban and rural areas, the target population was stratified into urban and rural
areas of residence for the STEPS survey. At the first stage, within each stratum, primary
sampling units (PSUs) (enumeration areas (EAs)) were selected systematically with probability
proportional to the 2004 Population Census EAs (measure of size equal to the number of
population in the EAs, provided by the census). Before selection, the census sectors were
sorted geographically from north to south within each stratum, in order to ensure additional
implicit stratification according to geographical criteria. A total of 400 clusters representing 400
EAs were selected from the 10 991 census EAs. These probabilistically selected clusters were
used also in Moldova’s DHS conducted in 2005, and the Multiple Indicator Cluster Surveys
(MICS) conducted in 2012. Cartographic materials from the Population Census conducted in
Moldova in 2004 were not available, thus it was not possible to use them for the STEPS survey.
Therefore, for the first stage the probabilistic samples from the abovementioned surveys were
used.
Out of the 400 selected clusters, 167 were rural and 233 were urban. The distribution of the
sample of 400 PSUs (EAs) for the DHS/MICS surveys was inversely proportional to the number
of population within each stratum, taking into account that the response rate is lower in urban
areas than rural owing to the smaller average size of the households in urban areas compared
with rural areas. Thus, disproportional allocation with oversampling for urban areas was applied
in the STEPS survey. A final weighting adjustment procedure was carried out to enable
estimates at national and urban/rural levels.
At the second stage, 15 households (secondary sampling units (SSUs)) were selected within
each of the 400 PSUs. From the updated list of households used for the MICS 2012 survey, 15
households were selected randomly per cluster, using the Microsoft Excel® random sample
tool. A total of 6000 individuals were selected from among the 400 clusters. The Kish method
(17) was applied for the random selection of one individual aged 18–69 years from each
household.
Age of participants included: 18-69 years
Source: Republic of Moldova STEPS 2013 report. Available at:
https://www.who.int/ncds/surveillance/steps/Moldova_2013_STEPS_Report.pdf
Mongolia: STEPS 2019
A multistage stratified sampling design was used to produce representative data for that age
range in Mongolia. A total of 6654 adults participated in the survey. Analysis weights were
calculated by taking the inverse of the probability of selection of each participant. These weights
were adjusted for differences in the age-sex composition of the sample population as compared
to the target population.
Different weight variables are available per Step:
wStep1 - for interview data
17
wStep2 - for physical measures
wStep3 - for biochemical measures
This allows for differences in the weight calculation for each Step of the survey as the age-sex
composition of the respondents to each Step can differ slightly due to refusal or drop out.
Additionally, some countries perform subsampling for Step 2 and/or Step 3. When no
subsampling is done and response rates do not differ across Steps of the survey, the 3 weight
variables will be the same.
Source: No report available. Sampling information obtained from
https://extranet.who.int/ncdsmicrodata/index.php/catalog/836/studydescription#page=sampling&tab=study-desc
Morocco: STEPS 2017
One of the essential elements for establishing a probability sampling plan is the constitution
an adequate sampling frame. For the purpose of the STEPS survey, the sampling frame used to
meet the sampling need was the 2014 master sample, developed by the HCP based on data
from the 2014 population and housing census. It has the advantage extrapolate the sample
results to the target population and estimate the accuracy desired. The stratification of
observation units belonging to any sampling frame makes it possible to design sampling plans
ensuring optimal sample size; a significant reduction in costs and a substantial improvement in
the accuracy of expected estimators. However, the choice of criteria allowing the population to
be divided into homogeneous groups (strata) and having recent and reliable data on these
criteria is a task that requires generally considerable efforts (updating the sampling frame) both
in terms of methodological than that of data collection.
In Morocco, the particularity of cities containing several social categories for which, synthesizing
the vector of heterogeneous demographic and socioeconomic behavior into a representative
characteristic makes stratification a difficult task. The stratification adopted was geographical for
the two environments according to the weight in terms of households, each of which has a
specific stratification: For urban units, the criteria used were the administrative division into
regions, provinces / prefectures and the dominant habitat type. As for the rural environment, the
primary units were stratified according to the geographical criterion, and the type of relief
dominant at the municipal level.
Age range of participants included: 18 years and older
Source: Morocco STEPS report [translated online]:
https://extranet.who.int/ncdsmicrodata/index.php/catalog/544/study-description
Namibia: DHS 2013
“The sample for the 2013 NDHS was a stratified sample selected in two stages. In the first
stage, 554 EAs were selected with a stratified probability proportional to size within the sampling
frame. The EA size is the number of households residing in the EA and recorded in the 2011
NPHC. Stratification was achieved by separating each region into urban and rural areas.
Therefore, the 13 regions were stratified into 26 sampling strata: 13 rural strata, and 13 urban
strata. Samples were selected independently in each stratum, with a predetermined number of
EAs selected as shown in Table A.3. Implicit stratification with proportional allocation was
achieved at each of the lower administrative unit levels by sorting the sampling frame before the
sample selection. Sorting was done according to the constituency and the EA code within a
sampling stratum, and by using a probability proportional-to-size selection procedure.
After the selection of EAs and before the main survey, a household listing operation was carried
out in all selected EAs, and the resulting lists of households served as a sampling frame for the
selection of households in the second stage. Some of the selected EAs may large. To limit the
18
amount of work done to list each household, selected EAs with more than 200 households were
segmented by the listing team in the field before the household listing. Only one segment was
selected for the survey, with probability proportional to the segment size. Household listing was
conducted only in the selected segment (see detailed instructions for segmentation in the DHS
Manual for Household Listing). So a 2013 NDHS cluster is either an EA or a segment of an EA.
In the second-stage selection, a fixed number of 20 households was selected in every urban
cluster and rural cluster, by equal probability systematic sampling. A spreadsheet indicating the
selected household numbers for each cluster was prepared. The survey interviewers
interviewed only the pre-selected households. To prevent bias, no replacements and no
changes of the pre-selected households were allowed in the implementing stages. In half of the
selected households where there was no male survey, all women age 15-49 were interviewed;
in the other half of the selected households where there was a male survey, all males and
females age 15-64 were interviewed.”
Age range of participants included: women 15 to 64 years
Source: The Nambia Ministry of Health and Social Services (MoHSS) and ICF International.
2014. The Namibia Demographic and Health Survey 2013. Windhoek, Namibia, and Rockville,
Maryland, USA: MoHSS and ICF International.
Nepal: STEPS 2019
STEPS-2019 is national cross-sectional population-based household survey that used multistage cluster sampling design to sample households and eligible adult men and women (15-69
years of age) for questionnaire interview and physical examination (anthropometry, blood
pressure measurement, blood glucose and cholesterol and urine sample for salt).
Survey population included men and women aged 15-69 years who have been the usual
residents of the household for at least six months and have stayed in the household the night
before the survey. People with the follow characteristics were not included: Those whose
primary place of residence was in military base or group quarters, Those residing in hospitals,
prisons, nursing homes and other institutions, Those too frail and mentally unfit to participate in
the study, Those with any physical disability, Those unable or unwilling to give informed
consent.
Sampling of Primary units (clusters):
This national representative sample was selected through multistage cluster sampling.
Sampling frame consisting of the distribution of oldwards as in census 2011 was obtained from
Central Bureau of Statistics (CBS). Then, in each of the province, the oldwards were compared
with current classification of metropolitan, sub metropolitan, municipality, and rural municipalities
and recorded as per new classification which has been recently updated by the government of
Nepal. The location of the new classifications were matched with the oldwards and, finally, used
as the sampling frame for selecting Primary Sampling Units (PSUs) for 2019 STEPS survey.
As a trade-off between survey costs and reducing the standard error, it was decided to sample
25 survey participants from each cluster, requiring sampling of 36.12 ~37 clusters in each of 7
provinces i.e. 259 clusters at national level.
Within each Province, the numbers of clusters were assigned to the three sub-strata in
metropolitan, sub-metropolitan, municipality and rural municipality in proportion to the share of
population in each of these 3 substrata in the total Province population.
Sampling of households and individuals from clusters:
19
A total of 25 households were sampled from each of the cluster. A sampling frame of the all
households in the sampled PSUs was obtained through a complete household listing and
mapping carried out in the sampled PSUs in September 6 to December 6 2018.
Sampling frame for selection of households from each PSU was prepared by conducting
household listing and mapping. The team of enumerators visited the sampling PSUs and carried
out a complete mapping of all the households in the PSU. If the sampled cluster were large, (if
the population exceeds 300), cluster was segmented. In that case, field team started from
northeast corner of each PSU and prepared an enumeration area of 300 household’s with at
least one person aged 15 years or more. Household listing questionnaire was used to list all of
the household’s members in selected PSUs. The listing was carried out electronically using
Android ODK software. Mapping was done along with household listing. Drawing a location map
of the cluster as well a detailed sketch map of all structures residing in the cluster was done
These materials guided the interviewers to return to the pre-selected households for interview.
This lists of the households so prepared from all sampled PSUs served as the sampling frame
for the selection of households in the next stage. From the prepare list, 25 households per PSU
were sampled using equal systematic random sampling after determining the sampling interval
by dividing the number of listed household by 25 and by randomly selecting the starting number
between 0 and the sampling interval. From each of the selected, one adult member was
sampled randomly for participation in the survey using the android tablet.
Age range of participants included: 15 to 69 years
Source: Nepal STEPS 2019 Report. Available at:
https://extranet.who.int/ncdsmicrodata/index.php/catalog/771
Romania: SEPHAR III 2015-2016
Like previous SEPHAR surveys, following a multi- stratified sampling procedure, a
representative sample of 2000 Romanian adults aged between 18 and 80 years has been
randomly selected from the database of the Romanian population general direction of data
records following the principle of equality of chances of being enrolled in the study, regardless of
the size of the place of residency.
The stratification criteria were: territorial regions (based on the recommendations of the National
Institute of Statis- tics), type of residence (rural and urban), gender (men and women), and age
groups (18–24, 25–34, 35–44, 45–54, 55– 64, and 65 – 80 years) using the data from the last
Census available [8]. For an adult Romanian population of 16 269 839 adult citizens [8], of
which 40.41% are estimated to be hypertensive patients based on SEPHAR II results [1], with a
maximum error of 2.18% at a confidence level of 95%, the minimum required sample size was
1379 study participants.
Identification of the selected study participants respected the law for the protection of personal
data of individuals, in the manner that we did not reach a person with a precise identity but only
a person with certain demographic characteristics (a person of a certain sex, of an age within a
certain age category from a certain locality). About 1 month previous to the study conduction in
each locality, the selected study participants were informed about the survey conduction and
their selection because of their demographic characteristics and were invited to send a
response letter to the study organizers regarding their availability to participate in the study.
During the two study visits, scheduled at 4-day interval, all enrolled individuals were evaluated
by: 71-item survey questionnaire, anthropometric, and BP measurements together with
20
investigations for target organ damage, blood, and urine sample collection after proper fasting
time (8 – 14-h prior).
The SEPHAR III survey was conducted in two stages: the first between 16 and 23 November
2015 in the Bucharest– Ilfov region and the second between 15 February and 25 April 2016 in
the remaining of the 82 survey sites (41 cities and 41 communes).
Response rate was calculated as the ratio between the total number of included study participants with
eligible data for analysis and the total number of randomly selected study participants eligible for
inclusion in the study that were approached by the study investigators.
Age range of participants included: 18 to 80 years
Source: Dorobantu M, Tautu O-F, Dimulescu D, Sinescu C, Gusbeth-Tatomir P, ArsenescuGeorgescu C, et al. Perspectives on hypertension’s prevalence, treatment and control in a high
cardiovascular risk East European country: data from the SEPHAR III survey. J Hypertens.
2018;36(3):690–700.
South Africa: SANHANES 2012
“The survey applied a multi-stage disproportionate, stratified cluster sampling approach. A total
of 1000 census enumeration areas (EAs) from the 2001 population census were selected from a
database of 86,000 EAs and mapped in 2007 using aerial photography to create the 2007
HSRC master sample to use as a basis for sampling of households. The selection of EAs was
stratified by province and locality type. In the formal urban areas, race was also used as a third
stratification variable (based on the predominant race group in the selected EA at the time of the
2001 census). The allocation of EAs to different stratification categories was disproportionate, in
other words, over-sampling or over-allocation of EAs occurred in areas that were dominated by
Indian, coloured or white race groups to ensure that the minimum required sample size in those
smaller race groups were obtained. Based on the HSRC 2007 Master Sample, 500 Enumerator
Areas (EAs) representative of the sociodemographic profile of South Africa were identified and a
random sample of 20 visiting points (VPs) were randomly selected from each EA, yielding an
overall sample of 10 000 VPs. EAs were sampled with probability proportional to the size of the
EA using the 2001 census estimate of the number of VPs in the EA database as a measure of
size (MOS). One of the tasks of SANHANES-1 was to recruit and establish a cohort of 5 000
households to be followed up over the coming years. The sampling consisted of: Multi-stage
disproportionate, stratified cluster sampling approach; 500 EAs within which 20 VPs/households
per EA were sampled; Main reporting domains: sex (male, female), age-group (< 2 years, 2–5
years, 6–14 years, 15–24 years, 25–49 years, 50 years and older), race group (black African,
white, coloured, Indian), locality type (urban formal, urban informal, rural formal [including
commercial farms] and rural informal], and province (Western Cape, Eastern Cape, Northern
Cape, Free State, KwaZulu-Natal, North West, Gauteng, Mpumalanga, Limpopo).”
Age range of participants included: all ages; biomarker information collected on participants 6
years or older
Source: Human Sciences Research Council. SANHANES: Health and Nutrition. 2015. Available
at: http://www.hsrc.ac.za/en/research-areas/Research_Areas_PHHSI/sanhanes-health-andnutrition
Additional reference: Stokes A, Berry KM, McHiza Z, Parker WA, Labadarios D, Chola L, et al.
Prevalence and unmet need for diabetes care across the care continuum in a national sample of
South African adults: evidence from the SANHANES-1, 2011–2012. PLoS ONE. 2017;
12(10):e0184264. https://doi.org/10.1371/journal.pone.0184264 PMID: 28968435.
21
Sudan: STEPS 2016
A four-stage cluster sampling design was implemented. The four sampling stages were; 1)
selection of states from the six regions 2) selection of clusters (a cluster was a Popular
Administrative unit), 3) selection of households and 4) selection of eligible individuals. First
Stage (State): Administratively Sudan is divided into 18 states which are grouped in six regions,
(North, East, Khartoum, Central, Kordofan and Darfur region (Table 1). States were randomly
selected from each region. No geographical areas or populations were excluded from the
sampling frame. Thus 11 states were selected, probability proportional to the size, to represent
the six regions. A list of the selected states is shown in Table 2.1. Second Stage (Cluster PAU):
The Popular Administrative Units (PAU) is the smallest geographically border unit. These were
defined as the ‘cluster’ in the region. Clusters were randomly sampled from all PAUs, from both
urban and rural strata, according to probability proportional to size in each state, and urban/rural
distribution. The PAUs inaccessible due to security conditions were not excluded from the
sampling frame, because within certain areas the security status was continuously changing.
However, it was planned that if a PAU was found to be inaccessible at survey time, it should be
replaced. However, no replacement was required during this survey. Third Stage (Household):
Within the selected PAUs, all households (HH) were included in the sampling frame.
Accordingly (HH) were selected using systematic random methods.
Fourth Stage (Individual): The members of the household were first listed in the mobile
application (customized software). The inclusion criteria for the listed members were: all
individuals aged between 18 to 69 years, from both sexes, irrespective of his health status and
living in the selected household for a minimum of 6 weeks. The application was then run and it
randomly selected the individual who will be selected to participate in the study.
Age of participants included: 18-69 years.
Source: Sudan STEPS 2016 report. Available at:
https://www.who.int/ncds/surveillance/steps/Sudan_STEPwise_SURVEY_final_2016.pdf?ua=1
Tanzania: STEPS 2012
"The STEPS survey in the United Republic of Tanzania was a population-based survey of adults
aged 25-64. The study used both multistage cluster and random probability sampling
procedures. Fifty of 119 total districts were randomly selected as primary sampling units
(PSUs). Within these PSUs, enumeration areas (EAs) of > 50 households were randomly
selected. Any EA with < 50 households was merged with a neighboring EA. Within the EAs,
households were randomly selected from a list of all eligible households in the EA. A total of
5762 adults participated in the Tanzania STEPS survey. Within each selected household, the
Kish method was used to select the STEPS participant. This procedure was followed until the
predetermined sample was obtained for the enumeration area. The response rate for this survey
was 94.7%.”
Age range of participants included: 25 to 64 years
Source: Tanzania STEPS Survey Report. Available at:
http://www.who.int/chp/steps/UR_Tanzania_2012_STEPS_Report.pdf?ua=1
Additional reference: Mayige M, Kagaruki G. Tanzania STEPS survey report. Dar es Salaam:
National Institute of Medical Research; 2013.
Togo: STEPS 2010
“Those included in this survey are male or female subjects, living in urban or rural areas, aged
15 to 64 on the day of the survey, residing in the enumeration area for at least 6 months and
having given their informed consent to participate in this study. [...] Three hundred clusters were
randomly selected in a systematic draw with probability proportional to the size of the cluster
(number of households) in the 4620 areas of enumeration of the DGSCN (General Directorate
22
of Statistics and National Accounts) sampling frame. In order to obtain the 4,800 households at
the rate of 1 individual / household, 16 households per cluster were randomly selected at the
second stage of survey. In each of the selected households, one individual was selected as a
survey participant via the Kish Method. A household was defined as the group of persons, who
regularly share the main meal (regardless of their relationship). Households were not replaced
in the event of a refusal or two unsuccessful visits to the eligible person selected by Kish's
method. If the selected person was unwell or not present at the time of the interview, the
investigators either tried to find a new appointment or searched for the respondent.”
Age range of participants included: 15 to 64 years
Source: Translated from WHO: The Final Report on the Togo STEPS Survey 2010. Available at:
http://www.who.int/chp/steps/2010STEPS_Report_Togo_FR.pdf?ua=1.
Turkmenistan: STEPS 2018
Sample
The main purpose of the sample design for STEPS research in Turkmenistan - nationwide
coverage and reflection of the situation in the country as a whole for measurable indicators.
The survey was conducted among adults in Turkmenistan aged 18-69 years. (target
population), who gave written informed consent, for exceptions: persons in the ranks of the
National Armed Forces; population WHO STEPS Non-communicable disease risk assessment
26 www.who.int/chp/steps permanently residing (staying) in specialized institutions social and
rehabilitation assistance, hospitals and other institutions health care, correctional facilities.
Method of sampling and stratification
The STEPS study was used to generate a sample set two-stage probability sampling method
using stratification procedures and selection at each of the sampling stages. Geographical
coverage - all regions of Turkmenistan: Akhal, Balkan, Dashoguz, Lebap and Mary provinces
and the city of Ashgabat (the capital), which corresponds national administrative-territorial
division. To ensure the uniformity of the distribution of the sample set across the country was
stratification. Taking into account the division of each province into urban and rural
The total population was determined by 11 streets (the city of Ashgabat - only the city street, in
velayatakh - 10 strat). The total sample size was distributed in proportion to the number
households on the streets.
Age range of participants included: 18 to 69 years
Source: Translated from 2018 STEPS Survey Report. Available at:
https://www.who.int/ncds/surveillance/steps/turkmenistan/en/
Uganda: STEPS 2014
Sample Design
The study methodology followed the World Health Organization's (WHO) STEP wise approach
to surveillance (STEPS) which provides a standardized method for analyzing and disseminating
data on risk factors for non-communicable diseases (NCDs). The sample for the Uganda NCDs
was designed to provide Cardiovascular Diseases (CVD) prevalence’s, smoking and tobacco
use and alcohol consumption estimates for the country as a whole and for urban and rural areas
separately. A two stage sampling design was used to draw the sample. At the first stage,
Enumeration Areas (EAs) were drawn with Probability Proportional to Size (PPS), and at the
second stage, households which were the ultimate sampling units were drawn using Simple
Random Sampling (SRS). A total of 350EAs were selected from 2014 Uganda Population and
Housing Census Mapping Frame. At the EA level, the target was 14 households.
Sample frame
23
The 2014 Uganda NCD survey used a sampling frame of the 2014 Population Census Mapping
listing provided by the Uganda Bureau of Statistics (UBOS). The UBOS has an electronic file
consisting of 78,950 Enumeration Areas (EAs) created for the 2014 Population and Housing
Census. An EA is a geographic area consisting of a convenient number of dwelling units that
serve as counting units for the census. Tables A.1 provides information on the distribution of
EAs and households in the sampling frame by region and residence. The table shows that
among the 78,950 EAs, 13,087 (22%) are in urban areas and 65,863 (78%) are in rural areas.
The average size of an EA, measured in number of households, is 95 in an urban EA and 77 in
a rural EA, with an overall average of 79.
Age range of participants included: 18 to 69 years
Source: Ministry of Health. Non-Communicable Disease Risk Factor Baseline Survey: Uganda
2014 Report. Available at:
https://www.who.int/ncds/surveillance/steps/Uganda_2014_STEPS_Report.pdf
Zx
Vietnam: STEPS 2015
At the same time of STEP survey, MOH also conduct the Global Adult Tobacco Survey (GATS)
at the same scale, location, and study subjects (>15 years for GATS and 18-69 for STEPS).
The sampling of STEPS was done in as part of the sampling for the (GATS) conducted in
combination manner to save time and resources for these two surveys. Applied the multi-stages
complex sampling process, the sampling process done by GSO was as follow: • Sampling of
clusters (EA) In the first stage of sampling, the primary sampling unit (PSU) was an enumeration
area (EA). There are about 170,000 EAs in the whole Viet Nam and the average number of
households in each EA is different between urban and rural areas. An average number of
households in an urban EA and a rural EA is 133 households and 120 households, respectively.
Sample of EAs were selected from the master sample frame. The master sample frame was a
cluster frame made by the GSO based on the frame of Population and Housing Census 2009
and updated with data of 2014. Based on the Population and Housing Census data 2009, GSO
prepared a 15% of master sample to serve as a national survey sampling frame. The master
sample frame contains 25,500 enumeration areas (EAs) from 706/708 districts of Viet Nam (2
island districts were excluded from the GSO master sample frame). The master sample frame of
GSO was divided by two stratification variables: urbanization (1 = urban; 2 = rural) and district
group (1 = district/town/city of province; 2 = plain and coastal district; 3 = mountainous, island
district). It means that the master sample frame was divided into 6 sample frames or 6 strata.
The probability proportional to size (PPS) sampling method was used to select sample of EAs
from 6 strata of master sample frame. The final sample of GATS included 315 EAs in the urban
and 342 EAs for the rural. From these 657 EAs, 315 EAs were systematically selected for
STEPS.
Sampling of households At the second stage of sampling, 10% households in each EA were
selected. Thus, 15 households from the selected urban EA and 14 households from the
selected rural EA were chosen using simple systematic random sampling. The total households
for STEPS 2015 were 4,651 households.
Sampling of individuals: One eligible person is then randomly selected from each selected
household for the STEPS 1 interview. The selection of individual is automatically done by the
PDA program after eligible household members are entered into the PDA. The selection
probability of an eligible individual was calculated as a product of selection probability for each
stage. The sampling base weight for an eligible individual was the inverse of the selection
probability shown above.
Age range of participants included: 18 to 69 years
24
Source: National Survey on the Risk Factors of Non-communicable diseases (STEPS) Viet Nam
Report 2015. Available at: https://www.who.int/ncds/surveillance/steps/viet_nam/en/
Zambia: STEPS 2017
To ensure that the sample reflected the entire country of Zambia, a multi-stage cluster sampling
technique was used to select a nationally representative sample of adults in Zambia aged 18 to
69 years. It was decided to utilize the household listing from the Zambia PopulationBased HIV
Impact Assessment (ZAMPHIA) - a household-based national survey that was conducted
between March and August 2016 in order to measure the status of Zambia’s national HIV
response. ZAMPHIA offered the most pragmatic up to date and accessible national household
listing to be used as the sampling frame for this survey. The ZAMPHIA survey included 60,581
households drawn from 1,103 clusters referred to in this report as standard enumeration area
(SEA) (Table 2.4.1). Thus the sample drawn for the STEPS survey was a subsample of the
households selected for the ZAMPHIA survey. In the first stage of sampling, SEAs were
selected from each province using probability proportional to size (PPS). In the second stage,
15 households in rural SEAs and 20 households in urban SEAs were selected systematically
using appropriate sampling interval based on the number of households in that SEA. These
households constituted the final list of households for the STEPS survey prepared for the field
investigators (FI). In the third stage, while the FI approached the household and sought consent,
all eligible members in the household were entered into the Android-based devise used for the
survey. The device then selected one member from the eligible members using a simple
random sampling technique. The selected member was then interviewed having gone through
the ethical process of consent after being provided with information on the survey. If the
selected member was not available, a scheduled visit was made. If the selected member could
not be reached after two scheduled visits he or she was considered as non-response. There
was no replacement strategy so as to maintain the integrity and representativeness of the
sample.
Age range of participants included: 18 to 69 years
Source: STEPS 2017 Report. Available at:
https://extranet.who.int/ncdsmicrodata/index.php/catalog/620
Zanzibar: STEPS 2011
“The survey took place in June and July 2011, followed by data cleaning and analysis. One
Principal Investigator and five assistant researchers coordinated the survey on site, checked
completed questionnaires daily, and organized logistics. The six data collection teams consisted
each of six interviewers, one supervisor, one laboratory technician and one driver. Interviewers
were either health care workers or professional interviewers familiar with household surveys
such as DHS. The sample size was calculated to be 2800 participants. Each interviewer did on
average 3 – 4 interviews a day and was assisted on site by local village guides.
The study was a cross-sectional population based survey with a sample of a sufficient size with
a power to determine the proportion of adults that are exposed to selected risk factors
associated with NCDs; including those having raised BP, FBG or blood lipids, had experienced
injuries or traumas in recent times, and/or were mentally unwell (anxiety, depression), as well as
linking these conditions with one another and with the sociodemographic and economic
information obtained. People reported to be permanent residents (spending on average
maximum 3 nights per week outside the house, and not holding an address in another place) in
the selected households and fulfilled the inclusion criteria were enrolled into the survey. A
person could only appear once in the study. Therefore we classified a husband practicing
polygamy to be listed in the household of his first wife but not to be a member in the household
of the following wives. Inclusion criteria was age between 25 - 64 years, able to understand the
25
information given by the interviewer about the study prior to the beginning of the interview,
signing of the informed consent for accepting participation. Exclusion criteria was inability to
understand or comprehend the information given by data collector, inability to communicate
through verbal expression for consent and for responding to the questionnaires, severe/terminal
illness that hinders participation in the survey.
The target population is the entire population in Zanzibar whereby the whole of Zanzibar was
selected as the survey site, and hence all districts included. The total population is estimated to
be 1.2 million distributed unevenly between 10 districts. The sampling frame represented the
entire population in Zanzibar. The sampling strategy used is a multi-stage cluster sampling with
stratification. The ten districts are considered as different strata, and the total number of primary
sampling units, PSU, is allocated proportionately across all strata. Each district is divided into
smaller clusters. These clusters are the geographical and administrative units called Shehia11.
The Shehia are divided into smaller clusters called zones (also called mitaa, vitongoji, or vijiji)
which typically consist of 100-300 households. Zones smaller than that were merged to make
up one larger cluster, and zones much larger were split in smaller clusters.
At the first stage clusters were selected using Simple Random Selection, SRS, from the list of
clusters (Shehia) within each district. At the second stage clusters (zones) were randomly
selected using probability proportionate to size (PPS). At the third stage households were
randomly selected from the household lists provided by the administrative leader of the Shehia.
The two last stages of sampling were done using the software STEPSsampling.xls from WHO.
Finally participants were selected from the household using Kish method. The household lists
were complete and included households with no eligible participants for the survey. Therefore
an extra 7 households were sampled at third stage in each cluster for replacement in case a
selected household had no eligible participants and had to be changed. This was done before
data collectors went to the cluster.
Resources allowed for 100 PSU which was why 2800/100 = 28 households were selected from
each PSU (and disproportionate from each SSU). A structured questionnaire was used, based
on WHO STEPwise approach to chronic diseases risk factor surveillance.. After getting
behavioural and socio-demographic information, anthropometric measurements (BP, height,
weight, waist and hip circumference) was done the same day. Answers were recorded
electronically during interview using a Personal Digital Assistant (PDA). Biochemical
measurements (fasting blood glucose, triglyceride, and cholesterol levels) were done the next
day at a central place in each study site according to appointment and were done by Laboratory
technicians using dry chemistry for rapid and convenient results and to avoid suspicion
surrounding sending away blood samples. Results were recorded electronically on site using a
PDA, and participants received a paper copy of the results.
Every study site was visited one day for interviews. Sampled households/ participants were
visited at least three times before recorded as non-respondent. The following day the site was
visited for biochemical measurements. Laboratory technicians called participants who did not
show up to ask them to set up appointment for the following day (at a new study site). After all
study sites had been visited call-backs were made to all eligible participants (non-respondents)
who’s number we had obtained. A time and place near the participants was identified for data
collection. Participants met fasting and started with having blood sample drawn, afterwards the
interviews and anthropometric measurements were conducted. Laboratory technicians
continued biochemistry measurements for another few days.
Age range of participants included: 25 to 69 years
26
Source: Zanzibar STEPS Survey Report, [online]
https://www.who.int/ncds/surveillance/steps/2011_Zanzibar_STEPS_Report.pdf
27
Appendix 5: Diabetes biomarker devices by country
Diabetes Biomarker
Country
Point-of-care fasting capillary glucose
Accutrend® Plus (Roche, Basel,
Cambodia, Chile, Guyana, Malawi,
Switzerland)
Togo, Zanzibar
CardioCheck® PA (pts Diagnostics, Afghanistan, Belarus, Benin,
Indianapolis, Indiana, USA)
Bhutan, Burkina Faso, Kenya,
Moldova, Morocco, Nepal, Sudan,
Turkmenistan, Uganda, Vietnam,
Zambia
FreeStyle Optium H glucometer
India
HemoCue® Glucose 201 Analyzer
(HemoCue, Brea, California, USA)
MultiCare-in© (Biochemical
Systems International, Arezzo, Italy)
Prima home test
Unknown
Post Hoc
Adjustment*
Multiplied by
1.11
None
Namibia, Tanzania
Multiplied by
1.11
None
Georgia
None
Mongolia
Algeria, Armenia, Azerbaijan, El
Salvador, Kyrgyzstan, Laos
Laboratory-based assessment of fasting plasma glucose
Central laboratory was used for
Bangladesh, Mexico
processing
Cobas 6000 and C311 analyzer
Iran, Romania
(Roche Diagnostics, Indianapolis,
Indiana, USA)
Enzymatic assay (glucose oxidase) Iraq
Hitachi 7600 modular chemistry
China
analyzer (Hitachi, Tokyo, Japan)
CardioCheck PA Analyser
Ethiopia, Jordan
Hemoglobin A1c (HbA1c)
Capillary sample DCA 2000+
Fiji
analyzer (Siemens/Bayer, Munich,
Germany)
Dried blood spots using the
Indonesia
Hemocue system
Plasma sample by Cobas C311
Iran
auto-analyzer (Roche kits)
Central laboratory
Mexico
Unknown
Guyana
Venous blood Cobas 6000
Romania
28
None
None
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
Venous blood using automated
high-performance liquid
chromatography
Whole blood using Bio-Rad HLC723 G7/D10/PDQ A1c
*Post
South Africa
N/A
China
N/A
hoc adjustment to convert from capillary to plasma equivalents. N/A=Not available.
29
Appendix 6: Blood pressure measurement devices by country
Country
Measurement device
Number of
measurements
Interval between
measurements
Afghanistan
Calibrated
sphygmomanometer
No report available
No report available
Riester Ri-Champion
Automatic Digital Monitor1715
Life Source UA-767 Plus
Digital Monitor
3
3 minutes
El Salvador
Ethiopia
Boso-Medicus Uno
Boso Medicus Uno
Omron digital upper arm
meter (model not specified)
Omron Digital Monitor
HEM-705CP
NISSEI Digital Blood
Pressure Monitor (Model DS500)
Omron Digital Monitor
HEM-742
Manual mercury
sphygmomanometer
Not specified
Boso-Medicus Uno
Fiji
Not applicable
Georgia
Guyana
Algeria
Armenia
Azerbaijan
Bangladesh
Belarus
Benin
Bhutan
Burkina Faso
Cambodia
Chile
China
India
No report available No report available
No report available No report available
3
10 minutes
3
10 minutes
3
3
3
3 minutes
3 minutes
5 minutes
3
10 minutes
3
N/A
3
2 minutes
3
10 minutes
Not specified
3
Not specified
3 minutes
Boso Medicus Uno
3
3 minutes
Omron digital upper arm
meter (model not specified)
Omron Digital Monitor HEM8712
3
3 minutes
3
5 minutes
Indonesia
Omron Digital Monitor HEM7203
3
Iran
Beurer BM 20
3
First measurement
taken at beginning of
interview, subsequent
two taken during the
course of the interview
5
Iraq
Not specified
Not specified
Not specified
Jordan
Omron M3
Not specified
Not specified
Kenya
Omron M2 Digital Monitor
3
3-5 minutes
30
Country
Measurement device
Number of
measurements
Kyrgyzstan
No report available
No report available No report available
Laos
No report available
No report available No report available
Malawi
Omron M4-I
3
3-5 minutes
Mexico
Omron HEM-907 XL
“AHA protocol”
“AHA protocol”
Moldova
Boso-Medicus Uno
3
3 minutes
Mongolia
Not specified
Not specified
Not specified
Morocco
Spengler® ES 60
3
“a few minutes”
Namibia
3
Not specified
3
3 minutes
3
1 minute
South Africa
Life Source Digital Monitor
Model UA-767
Omron digital upper arm
meter (model not specified)
A&D Medical UA-767PC
Automatic Monitor
Omron M2 Digital Monitor
3
5 minutes
Sudan
Boso-Medicus Uno
3
3 minutes
Tanzania
3
Not specified
3
5 minutes
Turkmenistan
Omron digital upper arm
meter (model not specified)
Omron digital upper arm
meter (model not specified)
OMRON device
Uganda
Boso Medicus Uno
3
3-5 minutes
Vietnam
BOSO Device
Not specified
Not Specified
Zambia
Not specified
3
3-5 minutes
Zanzibar
Omron M2 Digital Monitor
3
5 minutes
Nepal
Romania
Togo
Interval between
measurements
No report available No report available
N/A=Not available.
31
Appendix 7: Cholesterol measurement devices by country
Measurement
Country
CardioCheck PA
Afghanistan, Belarus, Benin, Bhutan, Burkina Faso, Ethiopia, Jordan,
Kenya, Moldova, Mongolia*, Morocco, Nepal, Sudan, Turkmenistan,
Uganda, Vietnam, Zambia
Bangladesh, Guyana, Iraq, Iran, Mexico, Romania
Mongolia*
Central laboratory
SD LipidoCare
Analyzer
Unknown
Algeria, Armenia, Azerbaijan, Georgia, Kyrgyzstan
*The 2019 Mongolia STEPS survey reports using both CardioCheck PA and SD LipidoCare Analyzer to measure
cholesterol (https://extranet.who.int/ncdsmicrodata/index.php/catalog/836).
32
Appendix 8: National definitions of area residence
Country
Urban and city definitions collated by United Nations6
Afghanistan
“Sixty-six localities and provincial centres.”
Algeria
“For 1998 and 2008, agglomerations with 5,000 inhabitants or more, nonagricultural economic activity, connection to water supply network,
connection to electricity network, connection to network of sanitation and
additional conditions.”
“Cities and urban-type localities, officially designated as such, usually
according to the number of inhabitants and predominance of nonagricultural workers and their families.”
“Cities and urban-type localities, officially designated as such, usually
according to the criteria of number of inhabitants and predominance of
non-agricultural workers and their families.”
“Localities having a municipality (pourashava), town (shahar) committee
or cantonment board. In general, urban areas are a concentration of 5,000
inhabitants or more in a continuous collection of houses where the
community sense is well developed and the community maintains public
utilities, such as roads, street lighting, water supply, sanitary
arrangements, etc. These places are generally centres of trade and
commerce where the labour force is mostly non-agricultural and literacy
levels are high. An area that has urban characteristics but has fewer than
5,000 inhabitants may, in special cases, be considered urban.”
“Cities and urban-type localities (towns, semi-urban centres, industrial
communities and health resort communities), officially designated as
such.”
“Localities with 10,000 inhabitants or more.”
Armenia
Azerbaijan
Bangladesh
Belarus
Benin
Bhutan
Burkina Faso
Cambodia
Chile
China
“Areas satisfying at least 4 out of the following 5 conditions: (1) 1,500
inhabitants or more; (2) 1,000 inhabitants or more per square kilometre;
(3) more than 50 per cent of the population depends on economic activity
outside of the primary (e.g., agriculture, livestock and forestry) sector; (4)
area of the urban centre is 1.5 square kilometres or larger; and (5)
identified potential for future growth of the urban centre, particularly in
terms of its revenue base. As of 2005, there were 28 declared urban
centres and 26 satellite towns.”
“Cities and urban-type localities (communes), officially designated as
such, according to socio-economic characteristics such as a nonagricultural economy.”
“For 1998 and later, communes that meet at least one of the following
criteria: (1) population density exceeding 200 persons per square
kilometre, (2) percentage of male employment in agriculture below 50 per
cent, or (3) 2,000 inhabitants or more.”
“Populated centres with defined urban characteristics, such as certain
public and municipal services.”
“For 1982 and earlier, total population of cities and towns. Cities had
100,000 inhabitants or more or commanded special administrative,
strategic, or economic importance. Towns were either settlements with
33
Country
Urban and city definitions collated by United Nations6
Ethiopia
3,000 inhabitants or more, of whom more than 70 per cent were
registered as non-agricultural, or settlements with between 2,500 and
3,000 inhabitants, of whom more than 85 per cent were registered as nonagricultural. For 1990, all residents of urban districts in provincial and
prefectural-level cities, the resident population of streets (jiedao) in
county-level cities, and the population of all resident committees in towns.
For 2000, population of city districts with average population density of at
least 1,500 persons per square kilometre, population of suburban-district
units and township-level units meeting certain criteria, such as having
contiguous built-up area, being the location of the local government, or
being a street (jiedao) or having a resident committee. For 2010, urban
residents meeting the criterion defined by the National Bureau of Statistics
of China in 2008, i.e., the criteria used in the 2000 census plus residents
living in villages or towns in outer urban and suburban areas that are
directly connected to municipal infrastructure and that receive public
services from urban municipalities.”
“For 2007, the head of the municipality, where the primary civil, religious
and military authorities reside, and those areas having a continuous
cluster of at least 500 dwellings, with street lighting service, basic schools,
regular transportation service, paved or cobbled streets and telephone
services. For 1971, areas where authorities of the municipality reside, as
determined by those authorities.”
“Localities with 2,000 inhabitants or more.”
Fiji
“Places with 1,000 inhabitants or more.”
Georgia
“Cities and urban-type localities, officially designated as such, usually
according to criteria surrounding the number of inhabitants and the
predominance of non-agricultural workers and their families.”
“City of Georgetown (capital), and four other towns.”
El Salvador
Guyana
India
Indonesia
Iran
Iraq
Jordan
Kenya
“Statutory places with a municipality, corporation, cantonment board or
notified town area committee and places satisfying all of the following
three criteria: (1) 5,000 inhabitants or more; (2) at least 75 per cent of
male working population engaged in non-agricultural pursuits; and (3) at
least 400 inhabitants per square kilometre.”
“Municipalities (kotamadya), regency capitals (kabupaten) and other
places with urban characteristics.”
“For 1986 and later, districts with a municipality. Prior to 1986, all county
centres (shahrestan) regardless of size and places with 5,000 inhabitants
or more.”
“Municipality councils (Al-Majlis Al- Baldei).”
“Localities with 5,000 inhabitants or more as well as the district and subdistrict centres of each governorate irrespective of population size.”
“Municipalities, town councils, and other urban centres with 2,000
inhabitants or more. Due to substantial changes in the 1999 census
delineations of urban areas, only the population of the “urban core” is
considered to ensure consistency with previous censuses.”
34
Country
Urban and city definitions collated by United Nations6
Kyrgyzstan
Malawi
“Cities and urban-type localities, officially designated as such, usually
according to criteria based on the number of inhabitants and
predominance of non-agricultural workers and their families.”
“For 2005, areas within municipal vicinity with the centre of that
municipality having 600 inhabitants or more, or at least 100 households.
Further, the areas must have certain urban characteristics (roads,
electricity, market function, tap water supply).”
“Townships, town planning areas and district centres.”
Mexico
“Localities with 2,500 inhabitants or more.”
Moldova
“Cities and urban-type localities, officially designated as such, usually
according to criteria based on the number of inhabitants and the
predominance of non-agricultural workers and their families.”
“Ulaanbaatar (capital) and district centres.”
Laos
Mongolia
Morocco
Namibia
Nepal
Romania
South Africa
Sudan
Tanzania
Togo
“Localities officially designated as urban according to administrative
divisions and entities that satisfy the quantitative criteria (minimum
population threshold) and qualitative criteria (density of equipment,
predominance of non-agricultural activities, etc.)”
“The district headquarters and other settlements of rapid population
growth with facilities that encourage people to engage in non- agricultural
activities.”
“For 1999 and later, a complex set of rules varying by ecological zones
and based on annual revenue, population size and infrastructure is used.
For 1981 and 1991, localities (panchayats) with 9,000 inhabitants or more.
For 1961 and 1971, localities (panchayats) with 5,000 inhabitants or
more.”
“Municipalities and towns with certain urban socio-economic
characteristics.”
“A classification based on dominant settlement type and land use. Cities,
towns, townships, suburbs, etc., are typical urban settlements.
Enumeration areas comprising informal settlements, hostels, institutions,
industrial and recreational areas, and smallholdings within or adjacent to
any formal urban settlement are classified as urban. The 1996 estimate
was adjusted to comply with the 2001 census definition...”
“Localities of administrative and/or commercial importance or with 5,000
inhabitants or more.”
“For 1978 and later, all regional and district headquarters, as well as all
wards with urban characteristics (i.e., exceeding certain minimal level of
size-density criteria and/or with many of their inhabitants in nonagricultural occupations). No specific numerical values of size and density
are identified, and wards are defined as urban based on the decision of
the District/Regional Census Committees. For 1957 and 1967, 16
gazetted townships.”
“For 1981 and later, 21 administrative centres of prefectures. For 1970
and earlier, seven urban communes.
35
Country
Urban and city definitions collated by United Nations6
Turkmenistan
“Cities and urban-type localities, officially designated as such, usually
according to criteria based on the number of inhabitants and the
predominance of non-agricultural workers and their families.”
“For 2002 and later, gazetted cities, municipalities and towns with 2,000
inhabitants or more. For 1991 and earlier, cities, municipalities, towns,
town boards and all trading centres with 1,000 inhabitants or more.”
“Places with 4,000 inhabitants or more.”
Uganda
Vietnam
Zambia
Zanzibar
“Localities with 5,000 inhabitants or more and with a majority of the labour
force not in agricultural activities.”
See Tanzania above.
36
Appendix 9: Summary of diabetes performance measures
Numerator
Denominatora
Ever tested
Individuals who ever had glucose
measured by a health worker
Individuals with diabetes
35
Awareness of
diagnosis
Individuals ever told by a health
worker that they have diabetes
Individuals with diabetes
42
Glucose-lowering Individuals using an oral glucosemedication
lowering medication or insulin
Individuals with diabetes who
have HbA1c ≥8.0% (FBG ≥9.2
mmol/L) or use an oral glucoselowering medication or insulin
42
Blood pressurelowering
medication
Individuals using an
antihypertensive medication
Individuals with diabetes and
hypertensionb
40
Statin
Individuals using a statin
Individuals age ≥40 years with
diabetes
28
Performance
measure
Number of
countries
Diagnosis
Treatment
Control
Glycemic control
Individuals with HbA1c <8.0% (FBG Individuals with diagnosed
<9.2 mmol/L)
diabetes
42
Blood pressure
control
Individuals with SBP <140 and DBP Individuals with diagnosed
<90 mmHg
diabetes
41
Cholesterol
control
Individuals (1) age <40 years with
total cholesterol <190 mg/dL or (2)
age ≥40 years and using statin
Individuals with diagnosed
diabetes
28
Combined ABC
control
Individuals with glycemic
and blood pressure control
Individuals with diagnosed
diabetes
41
Combined AB
control
Individuals with glycemic, blood
pressure, and cholesterol control
Individuals with diagnosed
diabetes
28
aDiabetes
was defined as use of a glucose-lowering drug (oral glucose-lowering medication or insulin) or an elevated
biomarker meeting the WHO’s criteria for diabetes: fasting plasma glucose (FPG) ≥7.0 mmol/l (126 mg/dl), random
plasma glucose ≥11.1 mmol/l (200 mg/dl), or glycated hemoglobin (HbA1c) ≥6.5%. bHypertension was defined as
systolic blood pressure of 140 mmHg or higher, diastolic blood pressure of 90 mmHg or higher, or current use of an
antihypertensive medication. Performance measures are generally consistent with recommendations in the WHO
Package of Essential Noncommunicable Disease Interventions for Primary Health Care.7 AB=glycemic and blood
pressure control. ABC=glycemic, blood pressure, and cholesterol. DBP=diastolic blood pressure. FBG=fasting blood
glucose. HbA1c=Glycated hemoglobin. SBP=systolic blood pressure.
37
Appendix 10: Unavailability of performance measures by country
Country
Ever
Awareness GlucoseBPStatin Glycemic Blood
Lipid
Combined
tested of diagnosis lowering
lowering
control
pressure control AB control
medication medication
control
Combined
ABC
control
Afghanistan
Algeria
Armenia
Azerbaijan
Bangladesh
Belarus
Benin
Bhutan
Burkina Faso
Cambodia
X
X
X
Chile
X
X
X
China
X
X
X
X
El Salvador
X
X
X
X
Ethiopia
Fiji
X
X
X
X
X
X
X
Georgia
Guyana
India
X
X
X
X
Indonesia
X
X
X
X
Iran
Iraq
38
Country
Ever
Awareness GlucoseBPStatin Glycemic Blood
Lipid
Combined
tested of diagnosis lowering
lowering
control
pressure control AB control
medication medication
control
Combined
ABC
control
Jordan
Kenya
Kyrgyzstan
Laos
X
X
X
Malawi
X
X
X
X
X
X
X
X
X
Tanzania
X
X
X
Togo
X
X
X
X
X
X
Mexico
X
Moldova
Mongolia
Morocco
Namibia
Nepal
Romania
X
X
South Africa
Sudan
Turkmenistan
Uganda
Vietnam
Zambia
Zanzibar
Countries
with data
35
42
42
40
28
42
41
The “X” refers to a diabetes performance measure that is unavailable in the country’s survey. BP=blood pressure.
39
28
41
28
Appendix 11: Details on missing data by country
Statina
Glycemic
control
Blood
pressure
control
Lipid
control
0
Blood
pressurelowering
medicationa
0
0
0
0.6
0.6
0.4
0.4
0.4
0.4
0
0.9
0.2
0
0
0
0
0
0
2.7
0
0
0
0
0
0
0
0
0
0
Bangladesh
0
0
0
0
0
0
0
0
0
Belarus
0
0
0
0
0
0
0
0
0
Benin
0
0
0
0
0
0
0
0
3.3
Bhutan
0
0
0
0
0
0
0
0
0
Burkina Faso
0
0
0
0
12.1
0
0
0
0
Cambodia
0
0
0
0
0
N/A
0
0
N/A
Chile
0
0.5
0.5
0.5
0.5
N/A
0
0.4
N/A
China
0
N/A
0.2
0.2
0.2
N/A
0
0
N/A
El Salvador
0
N/A
0
0
0
N/A
0
0.4
N/A
Ethiopia
0
0
0
0
0
0
0
0
0
Fiji
0
N/A
0
0
N/A
N/A
0
N/A
N/A
Georgia
0
0
0
0
0
0
0
0
0
Guyana
0
0
0
0
0
0
0
1.2
0
India
0
N/A
1.4
0
0.2
N/A
0
0.1
N/A
Indonesia
0.1
N/A
0.7
0.7
0.7
N/A
0
1
N/A
Iran
0
1.4
1.4
1.4
1.3
1.4
0
0.6
0.2
Iraq
0
21.7
21.7
21.7
12.7
0
0
0.3
0
Country
Rural
Testing Awareness of Glucosediagnosis
lowering
medicationa
Afghanistan
0
0
0
Algeria
0
0.4
Armenia
0
Azerbaijan
40
Statina
Glycemic
control
Blood
pressure
control
Lipid
control
0
Blood
pressurelowering
medicationa
0
0
0
2.7
0
0
0
0.9
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
N/A
0
0
N/A
Malawi
0
0
0
0
0
N/A
0
0
N/A
Mexico
0
N/A
0
4.9
0
0
0
0.8
0.3
Moldova
0
0.9
0.9
0.9
0.9
0.9
0
0
0
Mongolia
0
0
0
0
0
0
0
0
0.6
Morocco
0
0
0
0
0
0
0
0
0
Namibia
0
0
0
0
0
N/A
0
0
N/A
Nepal
0
0
0
0
0
0
0
0
0
Romania
0
N/A
0
0
0
0
0
0
0
South Africa
2
8.6
8.6
8.6
8.6
N/A
0
6
N/A
Sudan
0
0
0
0
0
0
0
0
0
Tanzania
0
0.7
0.7
0.7
0.7
N/A
0
0
N/A
Togo
0
1.1
1.1
1.1
0
N/A
0
14.3
N/A
Turkmenistan 0
0
0
0
0
0
0
0
0
Uganda
0
0
0
0
0
0
0
7.1
0
Vietnam
0
0.9
0.9
0.9
0.9
0.9
0
0
0
Zambia
0
0
0
0
0
0
0
0
0
Zanzibar
0
0
0
0
0
N/A
0
0
N/A
Overall
0
1.8
1.4
0.8
0.6
0.2
0
0.4
0.1
Country
Rural
Testing Awareness of Glucosediagnosis
lowering
medicationa
Jordan
0
0
0
Kenya
0
0
Kyrgyzstan
0
Laos
aRefers
to missingness among all individuals with diabetes in the sample. Note that missingness does not include data unavailability. N/A=Not available.
41
Appendix 12: Map of included countries
Countries included in the analysis but not easily visible on this map include Benin, El Salvador, Fiji, and Zanzibar.
42
Appendix 13: Sample characteristics
Characteristic
Total sample
Sample with diabetes
Sample with diagnosed
diabetes
Weighted % (95% CI)a
n
Weighted % (95% CI)a n
Weighted % (95% CI)a
n
<30 years
320,277
30.3 (29.6-30.9)
3,863
11.0 (9.8-12.4)
1,026
4.3 (3.2-5.8)
30-39 years
246,851
23.8 (23.4-24.2)
7,513
15.4 (14.0-16.9)
2,810
9.2 (7.2-11.7)
40-49 years
208,369
21.4 (21.1-21.7)
14,226
24.4 (23.2-25.6)
6,862
22.2 (20.4-24.0)
50-59 years
41,975
15.5 (15.3-15.8)
5,565
28.8 (27.5-30.1)
3,209
37.5 (35.0-40.1)
60-69 years
22,638
9.0 (8.4-9.6)
4,237
28.8 (27.5-30.1)
2,787
26.8 (24.5-29.2)
Male
171,331
49.3 (48.8-49.8)
9,962
48.1 (46.5-49.7)
4,556
45.4 (42.6-48.2)
Female
668,779
50.7 (50.2-51.2)
25,442
51.9 (50.3-53.5)
12,138
54.6 (51.8-57.4)
No schooling
229,041
17.5 (16.6-18.4)
8,995
18.6 (17.2-20.0)
3,617
15.4 (13.5-17.5)
Primary education
145,989
31.0 (29.5-32.5)
8,083
33.5 (31.5-35.6)
4,084
31.5 (28.7-34.4)
Secondary or above 463,202
51.5 (49.6-53.4)
18,094
47.9 (45.6-50.3)
8,865
53.2 (50.1-56.3)
283,216
45.7 (44.5-46.9)
17,588
53.0 (51.3-54.6)
9,478
61.5 (59.4-63.5)
Rural
556,810
54.3 (53.1-55.5)
17,803
47.0 (45.4-48.7)
7,210
38.5 (36.5-40.6)
Overall
840,110
100
35,404
100
16,694
100
Ageb
Sex
Education
Rural vs. urban
residence
Urban
aEstimates
account for survey design and equal country weighting.
regression analyses in this study.
bAge
is depicted in categories in this table but is maintained as a continuous variable in all
43
Appendix 14: Rural versus urban residence among study sample
Country
Rural
sample, n
Urban
sample, n
1,731
Rural sample of
total sample,
unweighted %
48.1
Rural sample of
total sample,
weighted %
43.7
Afghanistan
1,605
Algeria
1,943
3,925
33.1
33.5
Armenia
606
1,140
34.7
33.4
Azerbaijan
1,177
1,450
44.8
46.3
Bangladesh
3,682
3,265
53.0
79.7
Belarus
2,377
2,359
50.2
45.8
Benin
2,490
2,320
51.8
48.5
Bhutan
1,860
807
69.7
69.3
Burkina Faso
3,155
790
80.0
75.5
Cambodia
4,136
890
82.3
83.0
Chile
593
3,457
14.6
12.6
China
5,352
2,216
70.7
70.7
El Salvador
1,978
2,125
48.2
43.1
Ethiopia
5,662
2,049
73.4
81.8
Fiji
658
531
55.3
55.3
Georgia
1,596
1,559
50.6
52.4
Guyana
596
228
72.3
73.4
India
462,075
196,634
70.1
63.5
Indonesia
2,271
3,182
41.6
48.3
Iran
6,222
11,772
34.6
29.3
Iraq
764
2,758
21.7
24.0
Jordan
623
2,703
18.7
15.5
Kenya
2,040
1,934
51.3
61.2
Kyrgyzstan
1,485
997
59.8
66.2
Laos
1,655
738
69.2
69.4
Malawi
2,464
341
87.8
88.8
Mexico
3,937
7,464
34.5
21.6
Moldova
1,680
1,986
45.8
57.1
Mongolia
2,135
3,861
35.6
36.8
Morocco
1,669
2,611
39.0
35.5
Namibia
1,735
1,509
53.5
53.4
Nepal
4,420
641
87.3
90.9
44
Country
Rural
sample, n
Urban
sample, n
996
Rural sample of
total sample,
unweighted %
40.9
Rural sample of
total sample,
weighted %
40.9
Romania
689
South Africa
1,351
2,431
35.7
30.1
Sudan
4,333
2,119
67.2
63.3
Tanzania
3,622
1,001
78.3
68.9
Togo
2,473
711
77.7
62.0
Turkmenistan
1,912
1,833
51.1
52.0
Uganda
2,502
906
73.4
81.1
Vietnam
1,669
1,346
55.4
65.1
Zambia
2,148
1,183
64.5
54.0
Zanzibar
1,470
717
67.2
53.4
Overall
556,810
283,216
66.3
54.3 (53.1 to 55.5)a
aEstimate
(95% CI) using equal country weights.
45
Appendix 15: Proportion of diabetes population living in rural or urban areas
Country
Afghanistan
Rural
sample with
diabetes, n
119
Urban
sample with
diabetes, n
279
Proportion of diabetes
population who are
rural, unweighted %
29.9
Proportion of diabetes
population who are
rural, weighted %
33.2
Algeria
162
547
22.8
23.7
Armenia
38
95
28.6
24.5
Azerbaijan
81
188
30.1
29.3
Bangladesh
257
415
38.2
67.8
Belarus
124
140
47.0
42.2
Benin
113
190
37.3
30.8
Bhutan
42
33
56.0
55.0
Burkina Faso
70
29
70.7
68.9
Cambodia
91
61
59.9
63.3
Chile
60
337
15.1
14.6
China
318
181
63.7
63.7
El Salvador
141
217
39.4
33.5
Ethiopia
119
105
53.1
70.5
Fiji
277
232
54.4
54.4
Georgia
127
135
48.5
52.5
Guyana
102
27
79.1
79.5
India
12,137
8,397
59.1
53.6
Indonesia
167
260
39.1
44.5
Iran
376
1,030
26.7
22.5
Iraq
119
493
19.4
24.2
Jordan
77
407
15.9
14.8
Kenya
43
64
40.2
48.6
Kyrgyzstan
93
60
60.8
61.6
Laos
88
42
67.7
66.5
Malawi
20
6
76.9
76.3
Mexico
578
1,252
31.6
17.7
Moldova
136
183
42.6
53.3
Mongolia
186
393
32.1
32.7
Morocco
162
410
28.3
25.2
Namibia
80
138
36.7
36.2
Nepal
247
86
74.2
83.7
46
Country
Romania
Rural
sample with
diabetes, n
87
Urban
sample with
diabetes, n
103
Proportion of diabetes
population who are
rural, unweighted %
45.8
Proportion of diabetes
population who are
rural, weighted %
45.8
South Africa
120
347
25.7
29.3
Sudan
268
285
48.5
46.3
Tanzania
93
49
65.5
60.2
Togo
69
20
77.5
60.7
Turkmenistan 127
134
48.7
53.3
Uganda
21
21
50.0
62.7
Vietnam
40
69
36.7
49.5
Zambia
173
89
66.0
56.3
Zanzibar
55
39
58.5
41.1
Overall
17,803
17,588
50.3
47.0 (45.4 to 48.6)b
aThe
sum of the rural and urban diabetes population in these columns (n=35,391) differs from the total sample with
diabetes in the analysis (n=35,404) because there were n=29 respondents with diabetes who were missing the
variable for area of residence. bEstimate (95% CI) using equal country weights.
47
Appendix 16: Number of respondents with diabetes and diabetes prevalence by country
Country
Sample with
diabetes, n
398
Prevalence of diabetes
among rural population,
weighted %
9.0 (6.6-12.1)
Prevalence of diabetes
among urban
population, weighted %
14.2 (11.0-18.0)
Afghanistan
Algeria
709
7.1 (6.1-8.3)
11.6 (10.7-12.6)
Armenia
133
5.2 (3.6-7.5)
7.5 (5.7-9.8)
Azerbaijan
269
4.6 (3.6-5.8)
9.5 (8.0-11.3)
Bangladesh
672
6.9 (5.8-8.2)
12.6 (10.8-14.8)
Belarus
264
4.3 (3.5-5.3)
5.0 (4.1-6.0)
Benin
303
3.9 (3.1-5.0)
8.5 (7.0-10.2)
Bhutan
75
1.8 (1.2-2.8)
3.4 (2.1-5.4)
Burkina Faso
99
2.5 (1.9-3.3)
3.5 (2.3-5.4)
Cambodia
152
1.9 (1.5-2.4)
5.3 (4.0-7.0)
Chile
397
9.8 (6.4-14.8)
8.6 (7.3-10.2)
China
499
5.9 (5.3-6.6)
8.2 (7.1-9.4)
El Salvador
358
7.4 (6.1-8.9)
11.2 (9.7-12.9)
Ethiopia
224
1.9 (1.5-2.4)
3.5 (2.7-4.7)
Fiji
509
42.1 (38.4-45.9)
43.7 (39.5-48.0)
Georgia
262
5.9 (4.7-7.3)
6.0 (4.9-7.3)
Guyana
129
14.4 (11.4-17.9)
10.2 (6.4-15.9)
India
20,534
3.4 (3.3-3.6)
5.2 (5.0-5.5)
Indonesia
428
7.3 (6.2-8.7)
9.9 (8.7-11.2)
Iran
1,406
6.4 (5.8-7.0)
8.7 (8.2-9.3)
Iraq
612
14.7 (11.5-18.6)
14.5 (13.0-16.2)
Jordan
484
12.1 (8.8-16.3)
12.5 (10.9-14.4)
Kenya
107
1.5 (1.0-2.1)
2.5 (1.4-4.2)
Kyrgyzstan
153
5.1 (4.0-6.4)
6.2 (4.0-9.3)
Laos
130
3.7 (2.8-4.7)
5.3 (3.8-7.4)
Malawi
26
0.8 (0.5-1.3)
1.9 (0.8-4.5)
Mexico
1,830
14.2 (12.8-15.7)
17.9 (16.6-19.3)
Moldova
319
5.7 (4.6-6.9)
6.6 (5.4-8.0)
Mongolia
579
7.9 (6.7-9.3)
9.4 (8.4-10.6)
Morocco
572
7.9 (6.6-9.3)
12.7 (11.5-14.1)
Namibia
218
4.2 (3.3-5.3)
8.4 (6.9-10.2)
Nepal
333
5.6 (4.7-6.6)
11.1 (8.0-15.2)
48
Country
Sample with
diabetes, n
190
Prevalence of diabetes
among rural population,
weighted %
12.6 (10.3-15.3)
Prevalence of diabetes
among urban
population, weighted %
10.3 (8.6-12.4)
Romania
South Africa
479
8.6 (7.0-10.6)
10.8 (9.3-12.5)
Sudan
553
4.9 (4.2-5.8)
9.8 (8.5-11.3)
Tanzania
142
2.5 (1.9-3.3)
3.6 (1.9-6.8)
Togo
89
2.7 (2.0-3.4)
2.8 (1.7-4.6)
Turkmenistan
261
6.0 (4.9-7.3)
5.7 (4.6-6.9)
Uganda
42
1.1 (0.7-1.9)
2.8 (1.7-4.7)
Vietnam
109
2.1 (1.5-3.0)
4.3 (3.3-5.6)
Zambia
262
7.1 (6.0-8.3)
6.4 (4.9-8.4)
Zanzibar
94
2.7 (2.0-3.7)
4.5 (2.8-6.9)
Overall
35,404
6.0 (5.5-6.4)a
9.4 (8.9-9.9) a
aEstimate
(95% CI) using equal country weights. Note that the age range of the underlying surveys differs by country;
these estimates are not age adjusted, and thus they are not directly comparable among countries.
49
Appendix 17: Age-adjusted proportion of individuals with diabetes achieving
performance measures
Estimate, % (95% CI)
Goal
Urban
Rural
Testing
64.7 (62.1 to 67.3)
49.1 (46.5 to 51.7)
Awareness
48.5 (45.9 to 51.2)
39.0 (35.7 to 42.4)
Glucose-lowering medication
61.6 (57.9 to 65.2)
53.1 (49.7 to 56.5)
Blood pressure-lowering medication
44.8 (41.7 to 47.9)
37.2 (33.8 to 40.5)
Statin
9.4 (8.2 to 10.6)
6.7 (5.2 to 8.1)
Glycemic control
56.0 (52.5 to 59.5)
48.2 (43.7 to 52.6)
Blood pressure control
48.7 (45.2 to 52.1)
45.7 (41.3 to 50.1)
Cholesterol control
23.2 (19.1 to 27.3)
17.9 (14.2 to 21.6)
AB control
29.0 (25.5 to 32.5)
23.1 (19.4 to 26.8)
ABC control
9.5 (5.9 to 13.2)
3.8 (2.2 to 5.3)
50
Appendix 18: Population of individuals achieving and not achieving goal
Population size, thousands (95% CI)
Goal
Testing
Urban,
achievement
293 (282 to 303)
Urban, no
achievement
136 (126 to 147)
Rural, achievement Rural, no
achievement
149 (140 to 157)
175 (166 to 183)
Awareness
221 (210 to 233)
208 (196 to 219)
117 (107 to 128)
206 (195 to 217)
Glucose-lowering medication
174 (161 to 186)
92 (81 to 103)
88 (82 to 94)
73 (66 to 81)
Blood pressure-lowering medication
119 (111 to 127)
126 (117 to 135)
58 (52 to 64)
98 (91 to 106)
Statin
37 (33 to 42)
282 (272 to 293)
13 (11 to 16)
209 (199 to 220)
Glycemic control
121 (110 to 131)
101 (91 to 111)
56 (49 to 65)
61 (54 to 68)
Blood pressure control
102 (93 to 112)
113 (106 to 121)
56 (48 to 65)
58 (53 to 64)
Cholesterol control
51 (45 to 58)
176 (164 to 189)
22 (18 to 26)
101 (89 to 114)
AB control
57 (50 to 65)
162 (152 to 172)
28 (22 to 35)
88 (81 to 96)
ABC control
19 (15 to 24)
206 (193 to 218)
6 (4 to 8)
117 (105 to 129)
51
Appendix 19: Age-adjusted proportion of individuals with diabetes achieving performance measures by sex
Estimate, % (95% CI)
Goal
Urban men
Rural men
Urban women
Rural women
Testing
59.8 (56.1 to 63.5)
48.0 (44.0 to 52.0)
68.9 (65.6 to 72.2)
50.6 (47.5 to 53.7)
Awareness
45.8 (42.4 to 49.3)
37.5 (33.5 to 41.5)
51.1 (47.9 to 54.2)
40.5 (36.7 to 44.4)
Glucose-lowering medication
57.3 (52.2 to 62.3)
52.4 (47.5 to 57.3)
65.9 (61.7 to 70.1)
54.4 (50.2 to 58.7)
Blood pressure-lowering medication
38.1 (33.9 to 42.4)
30.8 (26.4 to 35.1)
51.6 (47.0 to 56.3)
42.2 (37.7 to 46.7)
Statin
9.7 (7.6 to 11.8)
6.3 (4.5 to 8.2)
9.2 (7.9 to 10.6)
6.5 (4.5 to 8.5)
Glycemic control
51.6 (46.6 to 56.7)
48.6 (42.2 to 55.1)
58.4 (54.1 to 62.7)
49.1 (43.6 to 54.5)
Blood pressure control
45.0 (39.5 to 50.6)
49.5 (43.4 to 55.6)
50.6 (45.7 to 55.5)
44.6 (39.1 to 50.1)
Cholesterol control
24.8 (19.0 to 30.6)
19.4 (13.6 to 25.1)
20.6 (16.4 to 24.8)
17.5 (12.6 to 22.4)
AB control
22.4 (18.1 to 26.8)
28.3 (23.2 to 33.5)
32.6 (27.5 to 37.7)
22.4 (16.9 to 28.0)
ABC control
8.5 (4.8 to 12.2)
7.4 (3.1 to 11.7)
8.2 (5.3 to 11.1)
2.7 (1.2 to 4.3)
52
Appendix 20: Differences in achievement of ever tested among rural versus urban
(reference category) populations with diabetes by country
Country
Risk ratio
P value
Average marginal
effect (%)
Urban adjusted
proportion (%)
Rural adjusted
proportion (%)
Afghanistan
Azerbaijan
Chile
Iraq
Moldova
Jordan
Laos
Iran
Georgia
Belarus
Togo
Algeria
Mongolia
Guyana
Bangladesh
Armenia
South Africa
Vietnam
Cambodia
Morocco
Sudan
Kyrgyzstan
Malawi
Turkmenistan
Nepal
Namibia
Tanzania
Bhutan
Kenya
Zanzibar
Benin
Ethiopia
Uganda
Zambia
Burkina Faso
1.12 (0.83 to 1.51)
1.04 (0.88 to 1.23)
1.03 (0.80 to 1.33)
1.01 (0.93 to 1.10)
0.98 (0.85 to 1.14)
0.98 (0.87 to 1.11)
0.98 (0.72 to 1.33)
0.97 (0.93 to 1.02)
0.96 (0.83 to 1.10)
0.94 (0.89 to 1.01)
0.93 (0.48 to 1.82)
0.92 (0.83 to 1.03)
0.88 (0.71 to 1.09)
0.88 (0.77 to 1.01)
0.87 (0.76 to 1.01)
0.87 (0.64 to 1.17)
0.86 (0.68 to 1.08)
0.78 (0.53 to 1.15)
0.77 (0.62 to 0.96)
0.75 (0.64 to 0.89)
0.72 (0.60 to 0.86)
0.72 (0.52 to 1.00)
0.64 (0.16 to 2.61)
0.61 (0.47 to 0.78)
0.61 (0.42 to 0.89)
0.59 (0.45 to 0.78)
0.57 (0.27 to 1.19)
0.48 (0.25 to 0.91)
0.47 (0.29 to 0.76)
0.33 (0.18 to 0.59)
0.29 (0.13 to 0.67)
0.25 (0.14 to 0.48)
0.24 (0.08 to 0.78)
0.22 (0.11 to 0.42)
0.11 (0.03 to 0.35)
0.471
0.664
0.803
0.854
0.837
0.781
0.872
0.208
0.535
0.081
0.841
0.155
0.234
0.065
0.066
0.357
0.202
0.213
0.020
<0.001
<0.001
0.050
0.520
<0.001
0.010
<0.001
0.133
0.026
0.003
<0.001
0.004
<0.001
0.019
<0.001
<0.001
6.7 (-11.6 to 25.0)
2.6 (-9.3 to 14.5)
2.6 (-17.8 to 22.9)
0.7 (-7.1 to 8.6)
-1.2 (-12.9 to 10.4)
-1.6 (-12.8 to 9.6)
-1.6 (-20.7 to 17.6)
-2.5 (-6.3 to 1.4)
-3.7 (-15.3 to 8.0)
-5.5 (-11.5 to 0.6)
-1.9 (-20.9 to 17.1)
-5.9 (-13.9 to 2.0)
-6.1 (-15.9 to 3.7)
-11.3 (-23.2 to 0.6)
-9.3 (-19.0 to 0.3)
-9.1 (-28.0 to 9.8)
-9.2 (-22.9 to 4.5)
-15.3 (-38.4 to 7.7)
-18.4 (-33.6 to -3.2)
-17.8 (-27.2 to -8.3)
-19.1 (-28.9 to -9.2)
-19.7 (-39.9 to 0.6)
-12.9 (-58.6 to 32.9)
-26.4 (-39.0 to -13.9)
-21.2 (-39.4 to -3.1)
-28.2 (-41.4 to -15.1)
-23.9 (-60.4 to 12.6)
-33.4 (-59.5 to -7.3)
-39.4 (-61.3 to -17.5)
-47.1 (-70.7 to -23.5)
-12.6 (-19.8 to -5.4)
-40.2 (-55.1 to -25.3)
-49.4 (-79.4 to -19.4)
-24.8 (-37.1 to -12.6)
-35.8 (-55.1 to -16.5)
57.5 (45.0 to 70.0)
69.0 (60.7 to 77.4)
77.3 (70.1 to 84.6)
91.4 (87.9 to 94.9)
81.0 (72.9 to 89.1)
90.7 (85.7 to 95.7)
62.5 (47.0 to 77.9)
90.4 (88.6 to 92.2)
81.9 (73.9 to 89.8)
98.8 (96.6 to 100.0)
29.3 (12.7 to 45.8)
78.8 (75.1 to 82.5)
51.0 (44.9 to 57.2)
93.8 (85.7 to 100.0)
73.4 (67.5 to 79.3)
69.3 (56.1 to 82.5)
65.7 (58.2 to 73.3)
70.6 (56.4 to 84.8)
81.4 (70.0 to 92.8)
71.6 (66.5 to 76.7)
68.2 (61.1 to 75.2)
69.9 (53.1 to 86.7)
36.0 (0.0 to 79.4)
67.7 (57.9 to 77.5)
54.0 (37.0 to 71.0)
69.4 (60.1 to 78.6)
55.3 (21.3 to 89.2)
64.2 (43.6 to 84.9)
73.9 (57.1 to 90.6)
70.0 (49.2 to 90.8)
17.8 (11.7 to 24.0)
53.8 (39.5 to 68.1)
65.2 (39.0 to 91.4)
31.8 (20.0 to 43.5)
40.1 (21.2 to 59.0)
64.2 (50.2 to 78.2)
71.7 (60.4 to 82.9)
79.9 (60.8 to 98.9)
92.1 (85.0 to 99.2)
79.8 (71.3 to 88.3)
89.1 (79.0 to 99.2)
60.9 (49.0 to 72.8)
87.9 (84.5 to 91.3)
78.2 (69.5 to 86.9)
93.3 (87.9 to 98.7)
27.4 (15.1 to 39.7)
72.9 (65.7 to 80.0)
44.9 (36.9 to 52.9)
82.4 (73.4 to 91.4)
64.1 (55.7 to 72.5)
60.2 (43.3 to 77.1)
56.5 (45.1 to 67.9)
55.3 (36.7 to 73.9)
63.0 (51.9 to 74.1)
53.8 (45.5 to 62.1)
49.1 (41.6 to 56.7)
50.2 (38.4 to 62.0)
23.1 (3.5 to 42.7)
41.3 (32.0 to 50.5)
32.8 (25.2 to 40.4)
41.2 (30.6 to 51.8)
31.4 (19.0 to 43.7)
30.8 (12.7 to 49.0)
34.4 (18.7 to 50.2)
22.9 (11.1 to 34.8)
5.2 (1.2 to 9.2)
13.6 (5.0 to 22.3)
15.8 (0.0 to 33.9)
6.9 (3.2 to 10.7)
4.3 (0.0 to 9.1)
0.01
0.1
1
10
Risk ratio
See availability by survey in Table 1 and Appendix 7. Sample sizes also were insufficient to run the models for the
following performance measures and surveys: BP lowering meds (Malawi excluded), glycemic control (Malawi and
Togo excluded), and BP control and combined AB control (Malawi, Togo, and Uganda excluded).
53
Appendix 21: Differences in achievement of awareness of diagnosis among rural versus
urban (reference category) populations with diabetes by country
Country
Risk ratio
P value
Average marginal
effect (%)
Urban adjusted
proportion (%)
Rural adjusted
proportion (%)
Afghanistan
Chile
El Salvador
Moldova
Armenia
Romania
Jordan
Laos
Belarus
Georgia
Iraq
Algeria
Mexico
Vietnam
Iran
Fiji
Azerbaijan
Mongolia
South Africa
Sudan
India
Guyana
Togo
Bangladesh
Morocco
Cambodia
China
Kyrgyzstan
Kenya
Malawi
Indonesia
Tanzania
Namibia
Bhutan
Nepal
Turkmenistan
Benin
Uganda
Zanzibar
Ethiopia
Zambia
Burkina Faso
1.26 (0.88 to 1.79)
1.15 (0.85 to 1.56)
1.14 (0.98 to 1.32)
1.11 (0.84 to 1.47)
1.10 (0.69 to 1.75)
1.10 (0.94 to 1.28)
1.04 (0.88 to 1.23)
1.03 (0.72 to 1.48)
1.02 (0.86 to 1.21)
1.01 (0.83 to 1.25)
1.01 (0.86 to 1.20)
0.97 (0.84 to 1.12)
0.97 (0.87 to 1.08)
0.96 (0.62 to 1.49)
0.96 (0.89 to 1.02)
0.94 (0.73 to 1.20)
0.92 (0.72 to 1.18)
0.90 (0.64 to 1.27)
0.84 (0.62 to 1.15)
0.84 (0.69 to 1.02)
0.83 (0.78 to 0.88)
0.81 (0.64 to 1.03)
0.78 (0.33 to 1.82)
0.74 (0.58 to 0.93)
0.71 (0.57 to 0.88)
0.71 (0.51 to 0.98)
0.67 (0.53 to 0.84)
0.66 (0.43 to 1.00)
0.56 (0.23 to 1.36)
0.56 (0.14 to 2.29)
0.56 (0.36 to 0.88)
0.55 (0.26 to 1.17)
0.52 (0.35 to 0.77)
0.52 (0.24 to 1.10)
0.50 (0.32 to 0.76)
0.49 (0.29 to 0.82)
0.43 (0.17 to 1.07)
0.38 (0.12 to 1.16)
0.32 (0.17 to 0.61)
0.26 (0.14 to 0.51)
0.24 (0.11 to 0.51)
0.21 (0.05 to 0.89)
0.202
0.352
0.089
0.450
0.684
0.228
0.639
0.855
0.832
0.898
0.877
0.689
0.530
0.857
0.192
0.605
0.502
0.556
0.280
0.071
<0.001
0.084
0.561
0.010
0.002
0.036
<0.001
0.052
0.200
0.403
0.011
0.122
0.001
0.087
0.001
0.007
0.069
0.087
<0.001
<0.001
<0.001
0.035
12.2 (-6.4 to 30.8)
9.3 (-11.2 to 29.7)
9.4 (-1.5 to 20.2)
5.0 (-8.0 to 18.1)
4.3 (-16.7 to 25.2)
7.3 (-4.5 to 19.0)
3.2 (-10.4 to 16.8)
1.7 (-16.8 to 20.2)
1.4 (-11.2 to 13.9)
0.9 (-12.9 to 14.7)
0.9 (-10.3 to 12.1)
-1.8 (-10.6 to 7.0)
-2.1 (-8.8 to 4.5)
-2.1 (-25.1 to 20.9)
-3.4 (-8.6 to 1.7)
-2.1 (-10.3 to 6.0)
-4.5 (-17.6 to 8.5)
-2.3 (-10.0 to 5.3)
-7.8 (-21.4 to 5.9)
-8.5 (-17.6 to 0.5)
-8.3 (-11.2 to -5.4)
-14.8 (-31.8 to 2.3)
-5.1 (-22.8 to 12.6)
-15.0 (-25.6 to -4.4)
-15.9 (-24.8 to -6.9)
-18.9 (-36.6 to -1.2)
-14.9 (-23.7 to -6.2)
-20.5 (-42.3 to 1.3)
-16.5 (-43.9 to 10.9)
-15.2 (-57.2 to 26.7)
-13.9 (-23.2 to -4.6)
-24.0 (-59.9 to 11.8)
-25.5 (-38.9 to -12.0)
-24.2 (-50.8 to 2.3)
-23.0 (-39.6 to -6.4)
-14.0 (-24.0 to -3.9)
-5.7 (-11.2 to -0.1)
-27.5 (-52.9 to -2.1)
-36.3 (-56.3 to -16.2)
-36.1 (-50.9 to -21.4)
-17.9 (-29.0 to -6.7)
-14.5 (-30.9 to 1.9)
47.3 (34.2 to 60.3)
60.3 (51.9 to 68.7)
67.6 (60.4 to 74.9)
44.1 (34.8 to 53.3)
42.4 (29.2 to 55.5)
74.6 (66.1 to 83.1)
80.5 (74.4 to 86.5)
50.8 (34.8 to 66.8)
71.5 (63.1 to 79.9)
66.4 (57.0 to 75.8)
66.7 (60.8 to 72.6)
62.8 (58.5 to 67.0)
62.0 (57.7 to 66.3)
53.5 (39.4 to 67.5)
78.6 (76.1 to 81.2)
34.2 (28.1 to 40.2)
55.6 (47.1 to 64.1)
24.1 (19.2 to 28.9)
49.9 (41.8 to 58.1)
51.6 (44.9 to 58.3)
48.2 (45.8 to 50.5)
78.9 (64.7 to 93.1)
23.1 (8.0 to 38.2)
57.1 (50.2 to 63.9)
54.2 (48.8 to 59.5)
64.3 (50.0 to 78.6)
45.3 (38.0 to 52.6)
59.8 (40.9 to 78.6)
37.7 (13.1 to 62.4)
34.7 (0.0 to 74.6)
31.6 (25.7 to 37.6)
53.7 (20.2 to 87.2)
52.9 (42.9 to 62.8)
50.3 (28.8 to 71.8)
45.6 (29.5 to 61.6)
27.5 (19.0 to 36.1)
9.9 (5.3 to 14.4)
44.1 (22.7 to 65.5)
53.3 (34.0 to 72.6)
49.1 (34.9 to 63.3)
23.4 (12.7 to 34.1)
18.2 (2.3 to 34.2)
59.4 (45.3 to 73.5)
69.6 (50.4 to 88.7)
77.0 (68.9 to 85.1)
49.1 (39.1 to 59.1)
46.6 (28.5 to 64.7)
81.8 (73.7 to 90.0)
83.7 (71.6 to 95.8)
52.6 (40.2 to 64.9)
72.9 (63.7 to 82.1)
67.3 (57.4 to 77.2)
67.6 (56.8 to 78.4)
61.0 (53.2 to 68.8)
59.9 (54.4 to 65.3)
51.4 (32.4 to 70.3)
75.2 (70.7 to 79.6)
32.0 (26.5 to 37.5)
51.1 (39.3 to 62.8)
21.7 (15.6 to 27.9)
42.2 (31.1 to 53.3)
43.1 (35.8 to 50.4)
39.9 (38.2 to 41.6)
64.1 (53.0 to 75.2)
18.0 (6.6 to 29.4)
42.0 (33.3 to 50.8)
38.3 (30.7 to 45.9)
45.5 (34.0 to 56.9)
30.4 (25.4 to 35.4)
39.3 (27.9 to 50.6)
21.3 (7.9 to 34.6)
19.4 (0.7 to 38.1)
17.7 (10.5 to 24.9)
29.7 (17.4 to 41.9)
27.4 (17.5 to 37.3)
26.1 (9.1 to 43.0)
22.6 (16.0 to 29.1)
13.5 (7.7 to 19.4)
4.2 (0.8 to 7.7)
16.6 (0.0 to 35.2)
17.0 (6.6 to 27.4)
13.0 (4.4 to 21.6)
5.5 (2.1 to 9.0)
3.8 (0.0 to 8.3)
0.01
0.1
1
10
Risk ratio
See availability by survey in Table 1 and Appendix 7. Sample sizes also were insufficient to run the models for the
following performance measures and surveys: BP lowering meds (Malawi excluded), glycemic control (Malawi and
Togo excluded), and BP control and combined AB control (Malawi, Togo, and Uganda excluded).
Appendix 22: Differences in achievement of glucose-lowering medication among rural
versus urban (reference category) populations with diabetes by country
Country
Risk ratio
P value
Average marginal
effect (%)
Urban adjusted
proportion (%)
Rural adjusted
proportion (%)
El Salvador
Ethiopia
Armenia
Moldova
Georgia
Guyana
Fiji
Chile
Tanzania
Romania
Iran
Algeria
Togo
India
Jordan
Iraq
Mongolia
Azerbaijan
Burkina Faso
Bangladesh
Belarus
Vietnam
Namibia
South Africa
Laos
Morocco
Nepal
Sudan
Afghanistan
China
Indonesia
Kyrgyzstan
Cambodia
Benin
Turkmenistan
Bhutan
Kenya
Zambia
Zanzibar
Uganda
1.49 (0.88 to 2.52)
1.26 (0.85 to 1.88)
1.23 (0.90 to 1.68)
1.20 (0.88 to 1.63)
1.19 (0.97 to 1.47)
1.17 (0.78 to 1.74)
1.15 (0.91 to 1.46)
1.10 (0.82 to 1.49)
1.10 (0.56 to 2.13)
1.07 (0.67 to 1.72)
1.04 (0.97 to 1.12)
1.02 (0.90 to 1.15)
1.02 (0.40 to 2.57)
0.99 (0.97 to 1.00)
0.97 (0.82 to 1.14)
0.94 (0.77 to 1.15)
0.92 (0.65 to 1.32)
0.91 (0.72 to 1.15)
0.91 (0.11 to 7.21)
0.91 (0.77 to 1.07)
0.91 (0.80 to 1.02)
0.90 (0.62 to 1.32)
0.86 (0.67 to 1.11)
0.86 (0.68 to 1.08)
0.85 (0.60 to 1.22)
0.84 (0.69 to 1.02)
0.83 (0.47 to 1.46)
0.78 (0.64 to 0.96)
0.77 (0.50 to 1.21)
0.75 (0.63 to 0.90)
0.72 (0.44 to 1.20)
0.70 (0.47 to 1.05)
0.67 (0.47 to 0.96)
0.67 (0.24 to 1.93)
0.66 (0.42 to 1.06)
0.55 (0.27 to 1.13)
0.46 (0.23 to 0.93)
0.43 (0.16 to 1.19)
0.37 (0.09 to 1.61)
0.33 (0.05 to 2.28)
0.134
0.248
0.189
0.240
0.089
0.447
0.242
0.526
0.783
0.765
0.280
0.749
0.966
0.119
0.687
0.563
0.658
0.427
0.920
0.244
0.098
0.598
0.243
0.196
0.373
0.080
0.509
0.020
0.258
0.001
0.212
0.082
0.029
0.456
0.083
0.101
0.032
0.101
0.181
0.215
9.4 (-3.2 to 22.0)
16.9 (-13.2 to 46.9)
10.7 (-5.4 to 26.8)
10.6 (-7.0 to 28.3)
13.2 (-1.7 to 28.0)
11.0 (-16.0 to 38.0)
7.9 (-5.2 to 20.9)
7.0 (-15.1 to 29.0)
4.7 (-28.7 to 38.1)
2.9 (-16.5 to 22.3)
3.2 (-2.6 to 8.9)
1.5 (-7.7 to 10.7)
0.6 (-27.9 to 29.1)
-1.1 (-2.5 to 0.3)
-2.9 (-17.0 to 11.2)
-4.2 (-18.2 to 9.8)
-3.9 (-21.2 to 13.3)
-6.2 (-21.1 to 8.8)
-2.9 (-64.8 to 58.9)
-5.9 (-15.7 to 3.9)
-8.8 (-19.0 to 1.4)
-7.0 (-33.2 to 19.1)
-10.7 (-28.4 to 7.1)
-11.1 (-27.2 to 5.1)
-9.8 (-32.0 to 12.3)
-11.6 (-23.9 to 0.7)
-9.2 (-38.3 to 19.9)
-14.7 (-26.6 to -2.8)
-14.6 (-38.6 to 9.3)
-18.5 (-29.6 to -7.3)
-10.7 (-26.1 to 4.7)
-21.9 (-46.3 to 2.5)
-24.0 (-44.2 to -3.8)
-7.9 (-27.1 to 11.2)
-16.3 (-34.3 to 1.7)
-25.4 (-53.7 to 3.0)
-39.9 (-68.9 to -10.9)
-26.1 (-55.8 to 3.7)
-25.7 (-56.5 to 5.1)
-45.3 (-114.8 to 24.2)
19.1 (11.9 to 26.3)
64.1 (48.4 to 79.7)
46.5 (31.9 to 61.1)
53.3 (40.6 to 66.1)
67.5 (56.5 to 78.6)
66.1 (41.0 to 91.1)
51.5 (42.0 to 61.1)
68.2 (57.4 to 79.1)
48.7 (17.6 to 79.8)
39.6 (25.3 to 54.0)
74.5 (71.5 to 77.6)
75.7 (71.3 to 80.1)
30.5 (7.0 to 54.0)
97.4 (96.5 to 98.2)
90.6 (85.2 to 96.1)
74.5 (67.8 to 81.1)
51.1 (40.6 to 61.6)
68.1 (58.3 to 78.0)
31.8 (0.0 to 67.0)
63.1 (55.7 to 70.5)
93.1 (88.0 to 98.2)
74.0 (57.8 to 90.2)
77.7 (65.6 to 89.9)
77.1 (68.2 to 85.9)
66.3 (46.5 to 86.1)
70.9 (64.9 to 76.9)
53.1 (25.9 to 80.2)
67.9 (60.0 to 75.9)
65.0 (51.3 to 78.8)
74.2 (65.9 to 82.6)
38.9 (30.6 to 47.2)
73.5 (55.0 to 91.9)
73.4 (59.3 to 87.4)
24.2 (12.2 to 36.2)
48.5 (34.3 to 62.8)
56.5 (28.5 to 84.5)
73.9 (53.4 to 94.3)
45.8 (21.8 to 69.8)
41.0 (15.1 to 66.9)
67.2 (12.8 to 100.0)
28.5 (18.1 to 38.9)
80.9 (55.0 to 100.0)
57.2 (40.6 to 73.8)
64.0 (51.2 to 76.8)
80.7 (71.0 to 90.4)
77.1 (66.1 to 88.0)
59.4 (50.4 to 68.4)
75.2 (56.7 to 93.7)
53.4 (26.7 to 80.1)
42.5 (27.3 to 57.8)
77.7 (72.8 to 82.6)
77.2 (69.2 to 85.2)
31.1 (5.8 to 56.4)
96.3 (95.2 to 97.4)
87.7 (74.6 to 100.0)
70.3 (57.0 to 83.5)
47.2 (32.7 to 61.7)
61.9 (48.4 to 75.5)
28.9 (0.0 to 76.4)
57.2 (47.9 to 66.5)
84.3 (75.5 to 93.0)
67.0 (46.5 to 87.5)
67.1 (53.8 to 80.4)
66.0 (51.8 to 80.2)
56.5 (40.5 to 72.5)
59.3 (48.4 to 70.3)
43.9 (32.1 to 55.7)
53.2 (43.9 to 62.5)
50.4 (30.6 to 70.2)
55.8 (48.0 to 63.5)
28.2 (15.0 to 41.3)
51.5 (35.3 to 67.8)
49.4 (34.3 to 64.5)
16.3 (0.7 to 31.9)
32.2 (19.7 to 44.7)
31.1 (7.5 to 54.7)
34.0 (10.5 to 57.4)
19.7 (3.0 to 36.5)
15.3 (0.0 to 37.3)
21.9 (0.0 to 58.3)
0.01
0.1
1
10
Risk ratio
See availability by survey in Table 1 and Appendix 7. Sample sizes also were insufficient to run the models for the
following performance measures and surveys: BP lowering meds (Malawi excluded), glycemic control (Malawi and
Togo excluded), and BP control and combined AB control (Malawi, Togo, and Uganda excluded).
Appendix 23: Differences in achievement of blood pressure-lowering medication among
rural versus urban (reference category) populations with diabetes by country
Country
Risk ratio
P value
Average marginal
effect (%)
Urban adjusted
proportion (%)
Rural adjusted
proportion (%)
Ethiopia
Chile
Guyana
Iran
El Salvador
Mexico
Armenia
Jordan
Azerbaijan
Belarus
Iraq
Mongolia
Kyrgyzstan
South Africa
Namibia
Nepal
Indonesia
Algeria
Benin
Georgia
India
Zambia
Turkmenistan
Cambodia
Afghanistan
Laos
Moldova
China
Togo
Zanzibar
Sudan
Kenya
Bhutan
Bangladesh
Morocco
Tanzania
Burkina Faso
Vietnam
Uganda
2.68 (1.07 to 6.70)
1.77 (1.21 to 2.59)
1.56 (0.77 to 3.15)
1.15 (1.01 to 1.32)
1.11 (0.90 to 1.39)
1.09 (0.92 to 1.28)
1.08 (0.50 to 2.29)
1.05 (0.84 to 1.33)
1.05 (0.77 to 1.43)
1.00 (0.82 to 1.24)
0.98 (0.73 to 1.31)
0.93 (0.73 to 1.18)
0.91 (0.50 to 1.65)
0.88 (0.66 to 1.18)
0.88 (0.67 to 1.15)
0.86 (0.42 to 1.77)
0.82 (0.44 to 1.53)
0.82 (0.64 to 1.05)
0.82 (0.35 to 1.90)
0.81 (0.60 to 1.10)
0.81 (0.69 to 0.94)
0.81 (0.35 to 1.87)
0.81 (0.56 to 1.17)
0.76 (0.45 to 1.28)
0.75 (0.48 to 1.18)
0.75 (0.36 to 1.55)
0.74 (0.55 to 1.00)
0.74 (0.57 to 0.96)
0.67 (0.08 to 5.57)
0.67 (0.16 to 2.75)
0.66 (0.45 to 0.96)
0.61 (0.15 to 2.59)
0.48 (0.20 to 1.17)
0.48 (0.31 to 0.75)
0.42 (0.27 to 0.67)
0.42 (0.11 to 1.68)
0.42 (0.06 to 2.92)
0.39 (0.14 to 1.06)
0.13 (0.01 to 1.23)
0.035
0.003
0.212
0.034
0.331
0.336
0.848
0.650
0.767
0.963
0.874
0.562
0.753
0.400
0.355
0.675
0.530
0.112
0.636
0.177
0.007
0.618
0.251
0.295
0.214
0.427
0.052
0.025
0.694
0.568
0.031
0.501
0.105
0.001
<0.001
0.217
0.365
0.064
0.071
28.2 (-2.4 to 58.8)
34.6 (9.8 to 59.5)
23.2 (-8.7 to 55.1)
8.2 (0.4 to 16.0)
7.7 (-8.0 to 23.5)
4.3 (-4.5 to 13.0)
2.8 (-26.3 to 31.9)
4.1 (-14.0 to 22.2)
2.6 (-14.8 to 20.0)
0.4 (-14.4 to 15.1)
-1.3 (-16.7 to 14.2)
-4.1 (-17.8 to 9.6)
-3.0 (-22.2 to 16.2)
-6.7 (-22.1 to 8.7)
-9.0 (-27.7 to 9.8)
-3.9 (-22.9 to 15.1)
-4.3 (-17.1 to 8.6)
-9.5 (-20.6 to 1.5)
-3.4 (-17.0 to 10.1)
-12.1 (-29.3 to 5.1)
-6.4 (-11.0 to -1.8)
-4.2 (-21.1 to 12.7)
-10.1 (-27.1 to 6.9)
-15.5 (-44.2 to 13.3)
-16.9 (-42.8 to 8.9)
-12.6 (-44.3 to 19.1)
-14.0 (-27.8 to -0.2)
-14.1 (-26.8 to -1.4)
-6.5 (-40.7 to 27.8)
-7.3 (-34.2 to 19.6)
-11.8 (-22.0 to -1.6)
-9.1 (-38.9 to 20.7)
-29.3 (-61.0 to 2.5)
-31.1 (-46.5 to -15.7)
-20.9 (-29.8 to -11.9)
-18.2 (-44.1 to 7.7)
-14.5 (-42.0 to 13.1)
-28.4 (-53.0 to -3.8)
-39.8 (-72.7 to -7.0)
16.8 (5.4 to 28.2)
44.9 (33.3 to 56.4)
41.6 (14.6 to 68.5)
53.4 (49.3 to 57.6)
67.7 (57.7 to 77.6)
49.7 (44.1 to 55.3)
36.5 (21.5 to 51.5)
75.3 (67.3 to 83.4)
54.2 (43.7 to 64.7)
70.6 (61.1 to 80.0)
54.0 (46.5 to 61.6)
59.8 (50.9 to 68.6)
33.2 (16.6 to 49.7)
57.8 (48.2 to 67.5)
75.2 (63.4 to 87.0)
27.4 (10.8 to 44.0)
23.7 (15.8 to 31.6)
52.2 (46.6 to 57.8)
18.6 (10.3 to 27.0)
64.3 (53.0 to 75.6)
33.5 (29.9 to 37.1)
22.0 (8.4 to 35.6)
52.0 (40.1 to 64.0)
64.0 (43.5 to 84.4)
68.6 (50.5 to 86.6)
49.7 (24.6 to 74.7)
54.5 (43.8 to 65.2)
54.5 (43.9 to 65.2)
19.9 (0.0 to 49.5)
21.8 (0.0 to 44.9)
34.8 (27.1 to 42.4)
23.7 (0.0 to 52.5)
56.8 (32.1 to 81.5)
59.9 (49.9 to 69.9)
36.2 (29.8 to 42.7)
31.6 (10.0 to 53.2)
25.1 (2.3 to 47.9)
46.6 (27.5 to 65.6)
45.8 (15.1 to 76.4)
45.0 (16.1 to 73.8)
79.5 (57.6 to 100.0)
64.8 (49.0 to 80.5)
61.6 (55.0 to 68.3)
75.4 (62.9 to 87.9)
53.9 (46.6 to 61.2)
39.3 (14.6 to 64.0)
79.4 (63.0 to 95.9)
56.8 (42.0 to 71.6)
70.9 (59.7 to 82.1)
52.8 (38.5 to 67.1)
55.7 (44.4 to 66.9)
30.1 (18.2 to 42.1)
51.1 (38.5 to 63.7)
66.3 (51.4 to 81.1)
23.5 (14.1 to 32.8)
19.5 (9.1 to 29.9)
42.7 (32.8 to 52.5)
15.2 (3.5 to 26.9)
52.2 (39.5 to 64.9)
27.1 (24.2 to 30.1)
17.8 (7.4 to 28.2)
41.9 (29.2 to 54.6)
48.5 (28.5 to 68.5)
51.6 (32.1 to 71.2)
37.1 (16.5 to 57.7)
40.5 (30.3 to 50.8)
40.4 (33.1 to 47.7)
13.4 (0.0 to 38.4)
14.5 (0.8 to 28.3)
23.0 (15.5 to 30.4)
14.5 (0.0 to 29.7)
27.5 (5.4 to 49.6)
28.8 (16.9 to 40.7)
15.4 (8.9 to 21.9)
13.4 (0.0 to 30.5)
10.6 (0.0 to 30.5)
18.2 (1.0 to 35.3)
5.9 (0.0 to 18.8)
0.01
0.1
1
10
Risk ratio
See availability by survey in Table 1 and Appendix 7. Sample sizes also were insufficient to run the models for the
following performance measures and surveys: BP lowering meds (Malawi excluded), glycemic control (Malawi and
Togo excluded), and BP control and combined AB control (Malawi, Togo, and Uganda excluded).
Appendix 24: Differences in achievement of glycemic control among rural versus urban
(reference category) populations with diabetes by country
Country
Risk ratio
P value
Average marginal
effect (%)
Urban adjusted
proportion (%)
Rural adjusted
proportion (%)
Nepal
Kyrgyzstan
Fiji
Jordan
Indonesia
Burkina Faso
Sudan
Namibia
Laos
Romania
Turkmenistan
Ethiopia
Zambia
Armenia
Bangladesh
Iran
Belarus
China
South Africa
Moldova
Georgia
Algeria
El Salvador
Tanzania
Iraq
Guyana
Azerbaijan
Benin
India
Cambodia
Vietnam
Afghanistan
Morocco
Chile
Kenya
Mongolia
Bhutan
Uganda
Zanzibar
1.47 (0.67 to 3.22)
1.38 (0.56 to 3.41)
1.35 (0.84 to 2.17)
1.16 (0.98 to 1.37)
1.13 (0.58 to 2.20)
1.09 (0.63 to 1.89)
1.09 (0.78 to 1.54)
1.06 (0.70 to 1.61)
1.05 (0.52 to 2.10)
1.05 (0.86 to 1.27)
1.05 (0.49 to 2.24)
1.03 (0.44 to 2.42)
1.02 (0.65 to 1.61)
1.01 (0.62 to 1.65)
1.00 (0.77 to 1.31)
0.96 (0.84 to 1.11)
0.96 (0.80 to 1.15)
0.95 (0.77 to 1.19)
0.95 (0.58 to 1.57)
0.95 (0.71 to 1.26)
0.94 (0.70 to 1.26)
0.92 (0.75 to 1.13)
0.92 (0.50 to 1.69)
0.89 (0.52 to 1.53)
0.88 (0.64 to 1.20)
0.88 (0.29 to 2.63)
0.88 (0.60 to 1.29)
0.87 (0.27 to 2.80)
0.86 (0.39 to 1.87)
0.85 (0.56 to 1.29)
0.78 (0.52 to 1.17)
0.75 (0.53 to 1.07)
0.70 (0.51 to 0.97)
0.66 (0.33 to 1.32)
0.62 (0.18 to 2.09)
0.58 (0.35 to 0.98)
0.56 (0.29 to 1.07)
0.19 (0.02 to 1.96)
0.07 (0.00 to 1.08)
0.330
0.484
0.217
0.089
0.711
0.635
0.605
0.783
0.887
0.622
0.904
0.943
0.932
0.973
0.974
0.607
0.631
0.668
0.850
0.706
0.667
0.422
0.781
0.662
0.423
0.814
0.502
0.815
0.699
0.434
0.227
0.110
0.030
0.238
0.428
0.041
0.077
0.127
0.057
18.0 (-14.2 to 50.2)
7.7 (-13.0 to 28.4)
8.5 (-4.8 to 21.7)
11.3 (-2.2 to 24.7)
4.6 (-20.6 to 29.7)
7.7 (-39.9 to 55.3)
3.5 (-9.8 to 16.7)
3.5 (-22.2 to 29.3)
2.1 (-26.6 to 30.8)
3.6 (-10.7 to 17.9)
2.2 (-34.2 to 38.6)
1.8 (-48.8 to 52.4)
1.4 (-31.8 to 34.6)
0.5 (-30.3 to 31.3)
0.3 (-14.7 to 15.2)
-1.8 (-8.7 to 5.1)
-3.4 (-17.1 to 10.3)
-3.1 (-17.2 to 11.1)
-2.1 (-24.2 to 19.9)
-3.5 (-21.8 to 14.8)
-4.0 (-22.3 to 14.3)
-4.8 (-16.4 to 6.7)
-1.7 (-13.4 to 10.0)
-6.7 (-36.9 to 23.6)
-7.1 (-23.6 to 9.5)
-3.6 (-35.6 to 28.3)
-6.7 (-25.8 to 12.4)
-7.3 (-67.9 to 53.2)
-0.3 (-1.9 to 1.3)
-9.4 (-33.0 to 14.2)
-18.3 (-46.0 to 9.5)
-21.3 (-45.2 to 2.5)
-17.3 (-31.1 to -3.4)
-18.8 (-45.6 to 7.9)
-22.4 (-76.0 to 31.3)
-24.0 (-43.4 to -4.7)
-39.6 (-75.0 to -4.3)
-62.4 (-150.7 to 25.8)
-54.1 (-91.9 to -16.3)
38.1 (10.4 to 65.9)
20.5 (4.1 to 36.9)
24.3 (15.1 to 33.5)
71.0 (64.2 to 77.7)
34.5 (23.2 to 45.8)
81.1 (41.9 to 100.0)
37.2 (28.2 to 46.1)
59.0 (45.9 to 72.1)
40.7 (17.6 to 63.7)
72.0 (61.9 to 82.0)
46.5 (26.5 to 66.5)
58.2 (32.8 to 83.7)
71.9 (49.3 to 94.5)
62.3 (45.2 to 79.4)
56.5 (46.3 to 66.7)
50.5 (47.0 to 54.1)
77.4 (68.4 to 86.4)
66.3 (56.0 to 76.7)
45.6 (33.2 to 58.0)
66.1 (53.0 to 79.3)
65.2 (53.5 to 77.0)
59.4 (54.1 to 64.7)
20.3 (13.1 to 27.4)
59.6 (31.0 to 88.1)
58.8 (51.4 to 66.2)
29.9 (0.9 to 58.8)
54.5 (44.0 to 65.1)
58.3 (33.5 to 83.1)
2.1 (0.6 to 3.5)
62.1 (45.1 to 79.1)
83.6 (64.4 to 100.0)
85.9 (74.5 to 97.2)
58.2 (51.3 to 65.1)
54.9 (43.3 to 66.6)
58.4 (20.8 to 96.1)
57.4 (46.8 to 68.1)
89.2 (69.4 to 100.0)
77.1 (0.0 to 100.0)
58.4 (24.4 to 92.3)
56.1 (39.8 to 72.4)
28.2 (13.8 to 42.6)
32.7 (22.8 to 42.6)
82.3 (70.5 to 94.0)
39.1 (16.7 to 61.5)
88.8 (35.5 to 100.0)
40.7 (31.0 to 50.4)
62.6 (40.6 to 84.6)
42.7 (26.0 to 59.4)
75.5 (65.3 to 85.8)
48.7 (19.5 to 77.9)
60.0 (22.7 to 97.4)
73.3 (46.5 to 100.0)
62.8 (36.4 to 89.3)
56.7 (44.9 to 68.6)
48.7 (42.8 to 54.6)
74.1 (63.3 to 84.8)
63.2 (53.5 to 72.9)
43.5 (25.3 to 61.7)
62.6 (49.9 to 75.4)
61.2 (47.3 to 75.1)
54.6 (44.4 to 64.8)
18.6 (9.1 to 28.1)
52.9 (24.9 to 80.8)
51.7 (36.9 to 66.6)
26.2 (12.6 to 39.8)
47.8 (31.7 to 63.9)
50.9 (0.0 to 100.0)
1.8 (1.1 to 2.4)
52.7 (35.8 to 69.6)
65.4 (38.2 to 92.5)
64.5 (43.5 to 85.5)
40.9 (28.8 to 53.1)
36.1 (11.9 to 60.2)
36.0 (0.1 to 72.0)
33.4 (17.0 to 49.7)
49.6 (17.6 to 81.6)
14.7 (0.0 to 43.4)
4.3 (0.0 to 15.0)
0.01
0.1
1
10
Risk ratio
See availability by survey in Table 1 and Appendix 7. Sample sizes also were insufficient to run the models for the
following performance measures and surveys: BP lowering meds (Malawi excluded), glycemic control (Malawi and
Togo excluded), and BP control and combined AB control (Malawi, Togo, and Uganda excluded).
Appendix 25: Differences in achievement of blood pressure control among rural versus
urban (reference category) populations with diabetes by country
Country
Risk ratio
P value
Average marginal
effect (%)
Urban adjusted
proportion (%)
Rural adjusted
proportion (%)
Guyana
Laos
Zambia
Chile
Bhutan
Ethiopia
Romania
Indonesia
Namibia
El Salvador
Nepal
Algeria
Iraq
Bangladesh
Cambodia
Jordan
Tanzania
Sudan
Mexico
South Africa
Morocco
India
Georgia
Vietnam
Iran
Mongolia
China
Kyrgyzstan
Turkmenistan
Zanzibar
Azerbaijan
Moldova
Belarus
Afghanistan
Armenia
Benin
Kenya
Burkina Faso
1.95 (0.89 to 4.24)
1.83 (0.89 to 3.77)
1.62 (0.78 to 3.37)
1.54 (1.04 to 2.28)
1.43 (0.48 to 4.24)
1.27 (0.66 to 2.47)
1.27 (0.97 to 1.66)
1.21 (0.66 to 2.21)
1.15 (0.65 to 2.05)
1.13 (0.93 to 1.38)
1.09 (0.63 to 1.86)
1.08 (0.89 to 1.31)
1.07 (0.71 to 1.63)
1.05 (0.76 to 1.45)
1.05 (0.77 to 1.43)
1.05 (0.79 to 1.40)
1.04 (0.56 to 1.95)
1.04 (0.76 to 1.41)
1.03 (0.89 to 1.18)
0.99 (0.60 to 1.64)
0.98 (0.71 to 1.35)
0.98 (0.92 to 1.04)
0.91 (0.56 to 1.47)
0.89 (0.52 to 1.52)
0.88 (0.77 to 1.02)
0.88 (0.61 to 1.27)
0.84 (0.62 to 1.14)
0.83 (0.35 to 2.00)
0.82 (0.29 to 2.34)
0.80 (0.16 to 4.08)
0.66 (0.36 to 1.21)
0.64 (0.25 to 1.68)
0.60 (0.30 to 1.18)
0.57 (0.30 to 1.07)
0.45 (0.13 to 1.56)
0.26 (0.04 to 1.90)
0.24 (0.05 to 1.23)
0.18 (0.00 to 141.80)
0.092
0.099
0.189
0.030
0.506
0.469
0.078
0.539
0.626
0.205
0.762
0.425
0.743
0.749
0.742
0.733
0.892
0.810
0.695
0.983
0.896
0.500
0.688
0.665
0.082
0.488
0.265
0.681
0.706
0.785
0.174
0.366
0.135
0.081
0.203
0.177
0.085
0.474
35.6 (3.7 to 67.4)
28.1 (0.1 to 56.0)
27.7 (-12.2 to 67.6)
28.1 (0.5 to 55.8)
15.9 (-32.2 to 64.0)
11.3 (-19.7 to 42.3)
14.1 (-1.4 to 29.7)
7.4 (-17.5 to 32.3)
6.5 (-20.9 to 33.9)
8.8 (-4.9 to 22.6)
4.5 (-24.5 to 33.5)
4.1 (-6.2 to 14.5)
2.9 (-15.1 to 20.9)
2.9 (-14.8 to 20.6)
3.6 (-17.8 to 25.0)
3.0 (-14.7 to 20.7)
2.3 (-31.8 to 36.5)
1.6 (-11.5 to 14.8)
1.7 (-6.7 to 10.1)
-0.2 (-18.3 to 17.9)
-1.0 (-15.6 to 13.6)
-1.4 (-5.7 to 2.8)
-3.5 (-20.7 to 13.7)
-6.6 (-36.5 to 23.2)
-6.3 (-13.3 to 0.6)
-7.6 (-28.3 to 13.2)
-8.5 (-23.4 to 6.5)
-5.8 (-35.3 to 23.6)
-4.8 (-29.7 to 20.1)
-5.1 (-42.7 to 32.6)
-13.5 (-31.0 to 4.0)
-6.7 (-20.8 to 7.3)
-10.1 (-23.1 to 2.9)
-22.0 (-44.7 to 0.7)
-21.4 (-49.3 to 6.5)
-33.5 (-67.9 to 0.9)
-16.7 (-36.6 to 3.2)
-66.8 (-277.8 to 144.3)
37.6 (9.4 to 65.8)
33.8 (11.1 to 56.5)
44.7 (17.0 to 72.4)
51.7 (39.8 to 63.6)
36.9 (6.0 to 67.8)
41.5 (18.4 to 64.6)
52.0 (41.0 to 63.0)
35.5 (24.1 to 46.9)
42.7 (29.3 to 56.1)
65.5 (56.5 to 74.4)
52.9 (27.5 to 78.3)
51.2 (45.9 to 56.6)
40.6 (33.0 to 48.2)
53.5 (41.6 to 65.4)
67.7 (52.1 to 83.3)
58.8 (50.8 to 66.8)
53.2 (22.5 to 84.0)
42.1 (32.8 to 51.4)
58.6 (53.7 to 63.6)
36.5 (25.2 to 47.9)
46.8 (40.0 to 53.6)
68.3 (65.0 to 71.6)
37.8 (26.0 to 49.6)
60.7 (41.7 to 79.7)
55.2 (51.7 to 58.7)
62.1 (51.9 to 72.4)
53.5 (42.4 to 64.6)
35.4 (7.6 to 63.1)
26.6 (10.0 to 43.2)
25.8 (0.0 to 54.3)
39.4 (29.1 to 49.6)
18.8 (8.3 to 29.4)
25.0 (14.7 to 35.4)
51.3 (32.8 to 69.8)
38.8 (19.3 to 58.3)
45.5 (18.9 to 72.0)
21.9 (2.0 to 41.7)
81.5 (0.0 to 100.0)
73.2 (58.8 to 87.6)
61.9 (45.8 to 78.0)
72.4 (42.9 to 100.0)
79.8 (55.4 to 100.0)
52.8 (15.4 to 90.2)
52.8 (26.9 to 78.7)
66.1 (54.9 to 77.4)
42.9 (20.6 to 65.2)
49.2 (25.3 to 73.1)
74.3 (63.8 to 84.8)
57.5 (41.9 to 73.0)
55.4 (46.2 to 64.5)
43.5 (27.2 to 59.9)
56.4 (43.5 to 69.3)
71.3 (55.8 to 86.7)
61.8 (46.0 to 77.7)
55.5 (21.8 to 89.3)
43.7 (34.2 to 53.1)
60.3 (53.5 to 67.1)
36.3 (20.1 to 52.6)
45.8 (32.6 to 59.1)
66.8 (64.2 to 69.4)
34.3 (21.6 to 47.0)
54.0 (27.6 to 80.5)
48.8 (42.9 to 54.8)
54.6 (36.8 to 72.3)
45.1 (35.1 to 55.0)
29.5 (12.9 to 46.2)
21.8 (4.2 to 39.4)
20.7 (0.0 to 45.8)
25.8 (11.6 to 40.1)
12.1 (2.3 to 22.0)
15.0 (6.6 to 23.3)
29.3 (11.8 to 46.7)
17.5 (0.0 to 37.1)
12.0 (0.0 to 34.8)
5.2 (0.0 to 13.3)
14.7 (0.0 to 88.7)
0.01
0.1
1
10
Risk ratio
See availability by survey in Table 1 and Appendix 7. Sample sizes also were insufficient to run the models for the
following performance measures and surveys: BP lowering meds (Malawi excluded), glycemic control (Malawi and
Togo excluded), and BP control and combined AB control (Malawi, Togo, and Uganda excluded).
Appendix 26: Differences in achievement of AB control among rural versus urban
(reference category) populations with diabetes by country
Country
Risk ratio
P value
Average marginal
effect (%)
Urban adjusted
proportion (%)
Rural adjusted
proportion (%)
Guyana
El Salvador
Indonesia
Nepal
Ethiopia
Laos
South Africa
Romania
Sudan
Jordan
Namibia
Zambia
Cambodia
Algeria
Iraq
Bangladesh
Chile
Bhutan
Azerbaijan
Vietnam
Turkmenistan
India
Kyrgyzstan
Iran
Georgia
China
Moldova
Tanzania
Morocco
Belarus
Armenia
Afghanistan
Benin
Mongolia
Kenya
Burkina Faso
11.47 (1.30 to 101.08)
1.83 (0.89 to 3.73)
1.75 (0.54 to 5.64)
1.69 (0.57 to 5.01)
1.43 (0.23 to 9.01)
1.37 (0.47 to 3.98)
1.36 (0.48 to 3.81)
1.30 (0.93 to 1.81)
1.28 (0.72 to 2.27)
1.24 (0.87 to 1.79)
1.17 (0.52 to 2.65)
1.12 (0.48 to 2.63)
1.06 (0.58 to 1.94)
1.03 (0.73 to 1.47)
1.02 (0.47 to 2.19)
0.98 (0.57 to 1.69)
0.92 (0.42 to 2.02)
0.91 (0.25 to 3.32)
0.88 (0.40 to 1.93)
0.84 (0.46 to 1.55)
0.83 (0.26 to 2.58)
0.80 (0.31 to 2.02)
0.78 (0.11 to 5.73)
0.78 (0.61 to 0.99)
0.77 (0.42 to 1.42)
0.77 (0.49 to 1.20)
0.77 (0.27 to 2.20)
0.76 (0.31 to 1.87)
0.73 (0.39 to 1.40)
0.73 (0.32 to 1.65)
0.61 (0.15 to 2.54)
0.50 (0.22 to 1.10)
0.36 (0.05 to 2.40)
0.34 (0.12 to 0.96)
0.26 (0.03 to 2.00)
0.18 (0.00 to 141.80)
0.029
0.098
0.343
0.340
0.701
0.556
0.558
0.119
0.397
0.234
0.696
0.783
0.837
0.862
0.963
0.940
0.842
0.879
0.747
0.579
0.738
0.632
0.803
0.040
0.397
0.247
0.621
0.551
0.347
0.449
0.491
0.084
0.276
0.042
0.189
0.474
20.5 (6.5 to 34.5)
8.0 (-2.5 to 18.6)
9.3 (-13.3 to 31.9)
12.1 (-9.8 to 33.9)
8.1 (-34.8 to 51.0)
7.9 (-17.1 to 33.0)
5.2 (-13.2 to 23.7)
12.4 (-3.1 to 28.0)
4.4 (-5.8 to 14.6)
10.9 (-8.2 to 30.0)
4.4 (-19.1 to 28.0)
5.5 (-35.3 to 46.3)
2.4 (-20.8 to 25.6)
0.9 (-9.3 to 11.1)
0.4 (-16.8 to 17.6)
-0.6 (-17.0 to 15.7)
-2.4 (-25.4 to 20.6)
-3.3 (-47.3 to 40.7)
-2.4 (-16.7 to 11.9)
-8.6 (-38.3 to 21.1)
-3.8 (-26.0 to 18.3)
-0.3 (-1.8 to 1.1)
-1.4 (-13.3 to 10.4)
-6.8 (-12.9 to -0.7)
-6.6 (-21.6 to 8.4)
-8.3 (-22.5 to 5.9)
-3.4 (-16.7 to 9.9)
-10.1 (-44.7 to 24.4)
-6.4 (-18.5 to 5.7)
-4.7 (-16.7 to 7.4)
-10.2 (-37.3 to 16.9)
-21.5 (-44.1 to 1.0)
-22.8 (-55.3 to 9.7)
-26.0 (-43.3 to -8.6)
-9.4 (-23.6 to 4.8)
-66.8 (-277.8 to 144.3)
2.0 (0.0 to 6.0)
9.7 (5.1 to 14.3)
12.4 (5.0 to 19.7)
17.5 (0.0 to 35.3)
19.0 (0.0 to 40.1)
21.3 (1.8 to 40.9)
14.5 (4.2 to 24.8)
41.4 (30.6 to 52.1)
15.7 (9.0 to 22.5)
44.5 (36.2 to 52.8)
25.6 (15.5 to 35.6)
45.0 (17.2 to 72.8)
37.5 (21.1 to 53.9)
28.3 (23.4 to 33.1)
22.1 (15.3 to 29.0)
30.2 (19.5 to 41.0)
31.2 (20.0 to 42.5)
35.7 (4.6 to 66.8)
19.9 (11.8 to 28.0)
54.9 (35.7 to 74.0)
21.9 (7.2 to 36.6)
1.6 (0.2 to 3.0)
6.4 (0.0 to 16.7)
30.7 (27.4 to 33.9)
28.6 (18.0 to 39.2)
36.0 (25.2 to 46.8)
14.6 (4.9 to 24.4)
43.1 (7.8 to 78.4)
24.1 (18.0 to 30.2)
17.4 (8.1 to 26.7)
26.1 (7.3 to 44.9)
42.6 (23.5 to 61.8)
35.3 (10.1 to 60.6)
39.1 (28.1 to 50.1)
12.7 (0.0 to 26.8)
81.5 (0.0 to 100.0)
22.5 (9.2 to 35.8)
17.7 (8.2 to 27.2)
21.7 (0.7 to 42.6)
29.5 (14.6 to 44.4)
27.2 (0.0 to 57.6)
29.3 (14.2 to 44.3)
19.7 (3.1 to 36.4)
53.8 (42.0 to 65.5)
20.1 (12.5 to 27.8)
55.4 (38.2 to 72.7)
30.0 (8.8 to 51.2)
50.5 (19.1 to 82.0)
39.9 (23.4 to 56.5)
29.2 (20.2 to 38.2)
22.5 (6.7 to 38.4)
29.6 (17.8 to 41.4)
28.8 (8.6 to 49.0)
32.4 (1.2 to 63.6)
17.5 (5.6 to 29.4)
46.3 (20.2 to 72.3)
18.1 (1.0 to 35.2)
1.3 (0.8 to 1.7)
4.9 (0.0 to 10.5)
23.9 (18.7 to 29.0)
22.0 (11.0 to 33.0)
27.7 (18.7 to 36.7)
11.2 (1.5 to 21.0)
33.0 (5.2 to 60.7)
17.7 (7.1 to 28.3)
12.7 (4.5 to 21.0)
15.9 (0.0 to 35.7)
21.1 (5.4 to 36.8)
12.6 (0.0 to 35.7)
13.2 (0.0 to 26.4)
3.3 (0.0 to 9.8)
14.7 (0.0 to 88.7)
0.01
0.1
1
10
Risk ratio
See availability by survey in Table 1 and Appendix 7. Sample sizes also were insufficient to run the models for the
following performance measures and surveys: BP lowering meds (Malawi excluded), glycemic control (Malawi and
Togo excluded), and BP control and combined AB control (Malawi, Togo, and Uganda excluded).
Appendix 27: Sensitivity analyses 1 (unadjusted proportions)
Estimate, % (95% CI)
Goal
Urban
Rural
Testing
68.3 (65.8-70.7)
46.0 (43.3-48.7)
Awareness
51.6 (48.9-54.2)
36.3 (33.0-39.6)
Glucose-lowering medication
63.0 (59.2-66.7)
51.4 (47.8-54.9)
Blood pressure-lowering medication
47.1 (44.0-50.2)
34.8 (31.3-38.4)
Statin
10.7 (9.5-12.1)
5.5 (4.4-6.8)
Glycemic control
55.8 (52.3-59.3)
48.5 (43.7-53.3)
Blood pressure control
46.9 (43.5-50.3)
48.3 (43.6-53.1)
Cholesterol control
21.9 (18.4-26.0)
19.6 (15.1-25.0)
AB control
28.1 (24.8-31.6)
24.3 (20.5-28.5)
ABC control
8.4 (5.8-11.9)
4.5 (3.0-6.8)
Appendix 28: Sensitivity analyses 2 (less strict glycemic target of HbA1c <8.0% [FPG <9.2
mmol/L]) – Differences in achievement of diabetes performance measures among rural
versus urban (reference category) populations
Outcome
Risk ratio
Average marginal
effect (%)
P value
Urban adjusted
proportion (%)
Rural adjusted
proportion (%)
Diagnosis
Ever tested
Awareness
0.76 (0.71 to 0.81)
0.80 (0.74 to 0.87)
-15.6 (-19.2 to -11.9)
-9.5 (-12.7 to -6.2)
<0.001
<0.001
64.7 (62.1 to 67.3)
48.5 (45.9 to 51.2)
49.1 (46.5 to 51.7)
39.0 (35.7 to 42.4)
Treatment
Glucose-lowering med
BP-lowering med
Statin
0.84 (0.77 to 0.90)
0.83 (0.75 to 0.91)
0.71 (0.55 to 0.92)
-8.6 (-12.4 to -4.9)
-7.6 (-11.5 to -3.8)
-2.7 (-4.6 to -0.8)
<0.001
<0.001
0.009
52.4 (49.1 to 55.7)
44.8 (41.7 to 47.9)
9.4 (8.2 to 10.6)
43.7 (40.9 to 46.5)
37.2 (33.8 to 40.5)
6.7 (5.2 to 8.1)
Control
Glycemic control
BP control
Cholesterol control
Combined AB control
Combined ABC control
0.90 (0.79 to 1.03)
0.94 (0.84 to 1.05)
0.77 (0.61 to 0.98)
0.75 (0.59 to 0.94)
0.52 (0.31 to 0.85)
-4.0 (-9.1 to 1.1)
-2.9 (-8.2 to 2.3)
-5.3 (-10.2 to -0.5)
-6.0 (-10.8 to -1.2)
-3.1 (-5.3 to -0.8)
0.127
0.277
0.033
0.014
0.010
41.1 (37.4 to 44.9)
48.7 (45.2 to 52.1)
23.2 (19.1 to 27.3)
23.5 (19.6 to 27.3)
6.3 (4.4 to 8.3)
37.2 (33.0 to 41.3)
45.7 (41.3 to 50.1)
17.9 (14.2 to 21.6)
17.5 (14.4 to 20.6)
3.3 (1.9 to 4.7)
.25
.5
1
2
Risk ratio
Lower in rural
Greater in rural
Appendix 29: Sensitivity analyses 3 (population weights) – Differences in achievement of
diabetes performance measures among rural versus urban (reference category)
populations
Risk ratio
Average marginal
effect (%)
P value
Urban adjusted
proportion (%)
Rural adjusted
Outcome
Diagnosis
Ever tested
Awareness
0.77 (0.71 to 0.85)
0.76 (0.68 to 0.83)
-15.5 (-20.6 to -10.3)
-11.4 (-15.3 to -7.5)
<0.001
<0.001
67.9 (64.8 to 71.0)
46.6 (43.7 to 49.6)
52.4 (48.5 to 56.4)
Treatment
Glucose-lowering med
BP-lowering med
Statin
0.88 (0.82 to 0.95)
0.78 (0.67 to 0.90)
0.77 (0.56 to 1.06)
-9.0 (-14.2 to -3.9)
-9.5 (-15.1 to -4.0)
-2.5 (-5.3 to 0.3)
<0.001
<0.001
0.110
76.5 (72.6 to 80.4)
43.2 (38.8 to 47.5)
10.8 (9.4 to 12.3)
67.5 (63.2 to 71.8)
Control
Glycemic control
BP control
Cholesterol control
Combined AB control
Combined ABC control
0.94 (0.81 to 1.10)
0.94 (0.83 to 1.07)
0.77 (0.60 to 0.97)
0.80 (0.59 to 1.08)
0.37 (0.19 to 0.71)
-2.4 (-8.5 to 3.7)
-3.2 (-9.8 to 3.4)
-5.7 (-10.7 to -0.8)
-4.5 (-10.4 to 1.4)
-6.8 (-11.1 to -2.6)
0.445
0.344
0.029
0.139
0.003
41.6 (37.7 to 45.6)
55.8 (51.6 to 59.9)
24.5 (20.9 to 28.2)
22.4 (18.2 to 26.5)
10.9 (7.3 to 14.4)
39.2 (34.2 to 44.3)
.25
.5
1
2
Risk ratio
Lower in rural
Greater in rural
proportion (%)
35.3 (32.3 to 38.2)
33.6 (29.7 to 37.5)
8.3 (5.7 to 11.0)
52.6 (47.9 to 57.3)
18.8 (14.7 to 22.9)
17.9 (13.9 to 21.9)
4.0 (1.8 to 6.3)
Appendix 30: STROBE checklist
Title and abstract
Item
Recommendation
No
1
(a) Indicate the study’s design with a commonly used term in the title or
the abstract
This information is provided in the Title and Abstract.
(b) Provide in the abstract an informative and balanced summary of
what was done and what was found
This information is provided throughout the Abstract.
Introduction
Background/rationale
2
Explain the scientific background and rationale for the investigation
being reported
This information is provided throughout the Introduction.
Objectives
3
State specific objectives, including any prespecified hypotheses
This information is stated in the final paragraph of the
Introduction.
Study design
4
Setting
5
Participants
6
Variables
7
Data sources/
measurement
8*
Bias
9
Present key elements of study design early in the paper
Study design is presented throughout the Methods section and in
Appendix 1.
Describe the setting, locations, and relevant dates, including periods of
recruitment, exposure, follow-up, and data collection
This information is provided in the first paragraph of the Methods
section and in Appendices 1-2.
(a) Give the eligibility criteria, and the sources and methods of selection
of participants
This information is provided in the second and third paragraph of
the Methods section and in Appendices 1-2.
Clearly define all outcomes, exposures, predictors, potential
confounders, and effect modifiers. Give diagnostic criteria, if applicable
This information is provided in the Methods under the Outcomes
and Statistical Analysis subsections and in Table 1.
For each variable of interest, give sources of data and details of
methods of assessment (measurement). Describe comparability of
assessment methods if there is more than one group
This information is provided in the Methods under the Data
Sources and Outcomes subsections and in Appendices 2-6.
Describe any efforts to address potential sources of bias
This information is described in the Methods under the Statistical
Analysis subsection.
Study size
10
Quantitative variables
11
Methods
Explain how the study size was arrived at
This information is provided in the Methods under the Sample and
definitions subsection and in Appendices 1-2.
Explain how quantitative variables were handled in the analyses. If
applicable, describe which groupings were chosen and why
This information is described in the Methods under the Statistical
Analysis subsection and in Appendix 1.
Statistical methods
12
a) Describe all statistical methods, including those used to control for
confounding
This information is provided in the Methods, throughout the
Statistical Analysis subsection.
(b) Describe any methods used to examine subgroups and interactions
This information is provided in the Methods, throughout the
Statistical Analysis subsection.
(c) Explain how missing data were addressed
This information is provided in the Methods in the second
paragraph of the Statistical Analysis subsection.
(d) If applicable, describe analytical methods taking account of
sampling strategy
This information is provided in the Methods under the Sample
subsection, and in Appendix 1.
(e) Describe any sensitivity analyses
This information is provided in the Methods under the Sensitivity
analysis subsection
Results
Participants
13*
Descriptive data
14*
Outcome data
15*
Report numbers of outcome events or summary measures
This information is provided in Figure 1.
Main results
16
(a) Give unadjusted estimates and, if applicable, confounder-adjusted
estimates and their precision (eg, 95% confidence interval). Make clear
which confounders were adjusted for and why they were included
This information is provided in the Statistical analysis subsection
of the methods, in Results, in Figure 2, and in the Sensitivity
analysis appendices (Appendices 21-24).
(b) Report category boundaries when continuous variables were
categorized
No continuous variables are categorized.
(c) If relevant, consider translating estimates of relative risk into
absolute risk for a meaningful time period
In addition to risk ratios, we present absolute differences (using
average marginal effects) and predictive margins throughout the
Results section and Appendices.
(a) Report numbers of individuals at each stage of study—eg numbers
potentially eligible, examined for eligibility, confirmed eligible, included
in the study, completing follow-up, and analysed
This information is reported in the Results under the Survey and
sample characteristics subsection and in Appendices 1-2.
(b) Give reasons for non-participation at each stage
This information is reported in Appendix 1.
(c) Consider use of a flow diagram
This information is reported in Appendix 1.
(a) Give characteristics of study participants (eg demographic, clinical,
social) and information on exposures and potential confounders
This information is provided in Table 2, in the Results under the
Survey and sample characteristics subsection and in Appendix 10.
(b) Indicate number of participants with missing data for each variable
of interest
This information is provided in Appendix 8.
Other analyses
17
Report other analyses done—eg analyses of subgroups and
interactions, and sensitivity analyses
This information is provided in the Results section under the
Sensitivity analyses subsection, and in Appendix 21-24.
Discussion
Key results
18
Limitations
19
Interpretation
20
Generalisability
21
Summarise key results with reference to study objectives
This information is provided throughout the Discussion.
Discuss limitations of the study, taking into account sources of potential
bias or imprecision. Discuss both direction and magnitude of any
potential bias
This information is provided in the second-to-last paragraph the
Discussion.
Give a cautious overall interpretation of results considering objectives,
limitations, multiplicity of analyses, results from similar studies, and
other relevant evidence
This information is provided throughout the Discussion.
Discuss the generalisability (external validity) of the study results
This information is provided throughout the Discussion.
Other information
Funding
22
Give the source of funding and the role of the funders for the present
study and, if applicable, for the original study on which the present
article is based
This information is provided in the Funding Support and
Disclosures sections
Supplementary references
1.
Riley L, Guthold R, Cowan M, et al. The World Health Organization STEPwise approach
to noncommunicable disease risk-factor surveillance: methods, challenges, and opportunities.
Am J Public Health 2016; 106(1): 74–8.
2.
WHO. HEARTS Technical package for cardiovascular disease management in primary
health care: systems for monitoring. Geneva: World Health Organization, 2018.
3.
WHO. Noncommunicable diseases global monitoring framework: indicator definitions
and specifications. 2014. https://www.who.int/nmh/ncdtools/indicators/GMF_Indicator_Definitions_Version_NOV2014.pdf (accessed January 10,
2022).
4.
World Health Organization. STEPS Country Reports. 2021.
https://www.who.int/ncds/surveillance/steps/reports/en/ (accessed April 13, 2021).
5.
World Health Organization. NCD Microdata Repository. 2021.
https://extranet.who.int/ncdsmicrodata/index.php/catalog (accessed July 19, 2021).
6.
United Nations, Department of Economic and Social Affairs, Population Division. World
Urbanization Prospects: The 2018 Revision. New York: United Nations, 2018.
7.
WHO. WHO package of essential noncommunicable (PEN) disease interventions for
primary health care. Geneva: World Health Organization, 2020.