Full Chapter Handbook of Health Decision Science 1St Edition Michael A Diefenbach PDF
Full Chapter Handbook of Health Decision Science 1St Edition Michael A Diefenbach PDF
Full Chapter Handbook of Health Decision Science 1St Edition Michael A Diefenbach PDF
https://textbookfull.com/product/practical-handbook-of-marine-
science-fourth-edition-michael-j-kennish/
https://textbookfull.com/product/fitness-science-explained-a-
practical-guide-to-using-science-to-optimize-your-health-fitness-
and-lifestyle-1st-edition-michael-matthews/
https://textbookfull.com/product/integrated-behavioral-health-
practice-1st-edition-michael-a-mancini/
https://textbookfull.com/product/routledge-handbook-of-health-
geography-valorie-a-crooks/
Medical Decision Making A Health Economic Primer 2nd
Edition Stefan Felder
https://textbookfull.com/product/medical-decision-making-a-
health-economic-primer-2nd-edition-stefan-felder/
https://textbookfull.com/product/brewing-science-a-
multidisciplinary-approach-1st-edition-michael-mosher/
https://textbookfull.com/product/the-decision-makers-handbook-to-
data-science-a-guide-for-non-technical-executives-managers-and-
founders-2nd-edition-stylianos-kampakis/
https://textbookfull.com/product/handbook-of-thermal-science-and-
engineering-1st-edition-francis-a-kulacki/
https://textbookfull.com/product/integrated-behavioral-health-
practice-michael-a-mancini/
Michael A. Diefenbach
Suzanne Miller-Halegoua
Deborah J. Bowen Editors
Handbook
of Health
Decision
Science
Handbook of Health Decision Science
Michael A. Diefenbach
Suzanne Miller-Halegoua
Deborah J. Bowen
Editors
Handbook of Health
Decision Science
123
Editors
Michael A. Diefenbach Deborah J. Bowen
Behavioral Research, Department of Department of Bioethics and Humanities
Medicine and Urology Seattle
Northwell Health University of Washington
Manhasset, NY Seattle, WA
USA USA
Suzanne Miller-Halegoua
Fox Chase Cancer Center
Philadelphia, PA
USA
v
Contents
vii
viii Contents
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
Introduction
xi
xii Introduction
Reference
Tversky, A. & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases.
Science, 185(4157), 1124–1131.
Part I
Basics First
What Are Values, Utilities,
and Preferences? A Clarification 1
in the Context of Decision Making
in Health Care, and an Exploration
of Measurement Issues
In health-related decision-making and the uncertainty and expected utility theory applies.
evaluation of health care delivery in particular, An important application of utilities is the
more specific definitions of values are usually QALY, or quality adjusted life year, in which
considered, which generally depend on the levels each year spent in a health state is multiplied by
of decision-making. Three levels are generally its utility, and the thus adjusted life years are
distinguished (Sutherland and Till 1993). The first summed. QALYs are mostly used in
is the health care or macro-level, where in the case cost-effectiveness ratios, based on expected util-
of limited resources budget allocation choices ity decision analyses in which the numerator is
have to be made among programs. The second, or expressed in costs (Dollars or Euros) and the
meso-level, pertains to policy making at the patient denominator (effectiveness) in QALYs.
group or hospital level, at which decisions have to
be made for defined groups of patients with the
same symptoms or disease, and for which Preferences
evidence-based guidelines or protocols are to be
developed. The third is the micro-level and applies There is no consistent definition of preferences in
to decision-making for an individual patient. The health care, but there is convergence in the notion
term value refers to different entities in these that health care preferences can be defined as
contexts, resulting in different elicitation pro- “statements made by individuals regarding the
cesses. For each of these levels, we prefer to relative desirability of a range of health experi-
reserve the term “values” for abstract, ences, treatment options or health states” (Bren-
trans-situational judgments. nan and Strombom 1998, p. 259). Individual
preferences exist as the relatively enduring con-
sequences of values (Brennan and Strombom
Utilities 1998). Differently from values, preferences are
object-focused and relate to specific options, or
Utility is a summary measure of the extent to attributes of options, in a specific decision
which each outcome of each choice option context.
achieves each of our ultimate goals (Baron 2008, Health-related preferences have been descri-
p. 233). Health state utilities play an important bed in relation to a variety of domains. In recent
role in health care decision-making and health studies, the term has been used to describe, for
economics. The most important applications of example, choice among a set of treatment options
utilities are in expected utility decision analysis, (Alolabi et al. 2011), treatment aspects (Pfützner
in which the expected utility for each possible et al. 2012), or health professionals (Bishop et al.
strategy is calculated by combining the utilities 2013); the desirability of procedural aspects of
for all possible resulting health states (outcomes) screening (Blom et al. 2012) or treatment (Vela
with the probabilities of these states occurring. et al. 2012); the desirability of sources (Gaglio
The utility of a health state is a cardinal measure et al. 2012), amount (Ter Hoeven et al. 2011), or
of the strength of an individual’s preference for kind of information (Ormond et al. 2009); and
particular outcomes when faced with uncertainty the desirability of participating in health-related
(Torrance and Feeny 1989). This concept of decision-making (Davison and Breckon 2012).
utilities dates back to the 1940s when a norma- Patient preferences—and this is true for health
tive model for decision-making under uncer- care provider or significant others’ preferences
tainty, expected utility theory, was developed too—vary further widely with respect to stability
(Von Neumann and Morgenstern 1944). In most and clarity (Street et al. 2012). Individual pref-
decisions in health care, outcomes may occur erences can be quite steady but need not. Pref-
with a certain probability, and the decision erences can vary as a function of disease severity,
problem is thus a problem of choice under can evolve as individuals learn new information
1 What Are Values, Utilities, and Preferences? A Clarification … 5
or gain new experiences, or have had more of life (Epstein and Peters 2009). For health care
opportunity to explore thoughts, feelings, and decision-making, the basic values paradigm thus
values relevant to the clinical situation. For seems most appropriate. Utilities and preferences
example, Feldman-Stewart et al. (2004) found are usually being constructed as a function of the
that 82 % of early stage prostate cancer patients specific decision options and the context in
who had already discussed their condition with which the utility or preference is being elicited
their oncologist and who were thinking through (Payne et al. 1992).
their treatment options, changed which attributes
affected their decision, and 72 % changed how
much they valued the treatment options as a Measurement of Utilities
whole, as they were going through a patient at the Macro- and Meso-Level
decision aid (see section “Values Clarification
Methods” on decision aids). Utilities are mostly used at the macro- and
meso-levels of health care decision-making, and
the level determines whether they should be
From Basic Values to Articulated assessed from the general public or from patients.
Utilities and Preferences: In cost-utility analyses from a societal perspec-
A Constructive Process tive, i.e., for macro-level decision-making, Gold
et al. (1996) have recommended the use of
Conceptualizations of values differ in the extent society’s preferences, that is, from a representa-
to which they are articulated (Fischhoff 1991). tive sample of fully informed members of the
On the one end of the continuum, people are seen general public. In guideline development, the
to hold articulated or well-differentiated, “com- meso-level, the use of utilities obtained from
plete” values that can be elicited if one asks the actual patients is preferred. Members of the
right question (Gregory et al. 1993). On the other public who are asked to imagine experiencing
end, people are seen to hold only basic values, health states assign lower utilities to those states
that is, lack well-differentiated values for all but than the patients who are actually experiencing
the most familiar issues, and that preferences these states (Stiggelbout and De Vogel-Voogt
need to be constructed (constructive preferences) 2008), which resonate with the disability para-
from basic values at the time of decision-making dox; many people with serious disabilities report
(Payne et al. 1999). In this partial perspective, that they experience a good or excellent quality
people could respond with values that are not at of life (Albrecht and Devlieger 1999). At the
stake if they miss nuances of the question asked. micro-level, with an individual patient, utility
Articulated values most often exist when deci- assessment is seldom used and if used, it is done
sions are personally familiar; with few conse- in a constructive way and meant to serve as
quences; implying no conflicting roles; and values clarification (e.g., Unic et al. 1998, and
formulated in a familiar fashion (Fischhoff 1991, section “Values Clarification Methods”).
Table 3). Complex decisions in health care—
such as allocation of resources or choice of
treatment, often are new to decision makers; have Approaches to Utility Measurement
more than a few consequences; and many of
these consequences are not commensurable, such We can distinguish two different approaches to
as trading treatment convenience (e.g., pills ver- measuring utilities. The holistic approach requires
sus injections) for treatment effectiveness. Fur- the participant to assign values to each possible
ther, values may be conflicting because options health state, where a state represents a combination
on offer cannot achieve both the goals of, for of many attributes. The decomposed approach
example, lengthening life and improving quality enables the investigator to obtain values for all
6 A.H. Pieterse and A.M. Stiggelbout
health states without requiring the judge to assign choice between her remaining life in that state
values to each one. It expresses the overall value as and a gamble with a probability of 0.90 that her
a decomposed function of the attributes. This remaining life will be in optimal health and a
approach can also be used specifically to obtain the probability of 0.10 of immediate death, her utility
utilities of the attributes per se, in health services for that health state is 0.90. The utility measured
research. with a SG reflects not only the participants’
preference for life in the health state, but also
their attitude toward risk. The use of probabilities
Holistic Approaches to Utility has proven to be a major drawback of the
Measurement method, since participants have difficulties
relating to probabilities. Moreover, they have
Holistic valuations of health states encompass been shown to transform probabilities; they tend
valuations of the quality of life of those states, to overweight small probabilities and under-
and the valuations are therefore sometimes called weight large probabilities (Tversky and Kahne-
preference-based measures of quality of life, as man 1992). In most examples in health, small
distinct from descriptive measures of quality of probabilities of bad outcomes (such as death)
life, using questionnaires such as the SF-36. The occur, which thus tend to be overweighted,
methods can be used either to have participants leading to extremely risk averse answers, and too
value hypothetical health states, or to have high utilities for the states under evaluation.
patients rate their own health. In the former case, Ceiling effects subsequently limit the ability of
the health states are described in a scenario, the SG to discriminate between health states.
generally framed in terms of physical, emotional, This has led researchers to use an alternative
and social functioning. Several methods exist to method, the time tradeoff method (TTO) (Wak-
assess utilities for health states holistically ker and Stiggelbout 1995).
(Stiggelbout and De Haes 2001). The Standard In the TTO, a participant is asked to choose
Gamble (SG) has long been seen as the gold between her remaining life expectancy in the
standard, since it adheres to the axioms of health state to be valued and a shorter life span in
expected utility theory. It is based on the prin- optimal health. In other words, she is asked whe-
ciple that a person will be willing to accept a risk ther she would be willing to trade years of her
in order to obtain good health, if he or she feels remaining life expectancy for an improved health.
that the health state under evaluation is undesir- As an example, let us say a 65-year-old woman has
able. The participant is offered the hypothetical a remaining life expectancy (according to national
choice between the sure outcome (the health state life tables) of 15 years. She is asked what length of
to be valued, for one’s remaining life expec- time (X) in optimal health would be equivalent to
tancy) and a gamble, with probability p of 15 years in her state of rheumatoid arthritis,
obtaining the best possible outcome, set at 1 assuming that in each case death would follow
(generally optimal health, for one’s remaining immediately. The simplest and most common way
life expectancy) and a probability (1 − p) of the to transform this optimal-health equivalent X into a
worst possible outcome, set at 0 (usually imme- utility (ranging from 0 to 1) is to divide X by 15.
diate death). By varying p, the value is obtained Thus, if she is willing to trade 3 years to obtain
at which the participant feels the sure outcome optimal health, her utility is 0.80 (12/15).
and the gamble to be equivalent. The utility for Both for the SG and TTO, elicitation becomes
the sure outcome, the health state to be valued, is more complex when temporary states are to be
equal to the value of p at this point of indiffer- valued (see, e.g., Jansen et al. 1998 for the details
ence (U = p x 1 + (1 − p) x 0 = p). Thus, for on the procedure).
example, a woman is asked to rate the state In the TTO no uncertainty is involved, and it
“rheumatoid arthritis”. If she is indifferent to the therefore does not adhere to expected utility
1 What Are Values, Utilities, and Preferences? A Clarification … 7
this system. The descriptive system consists of a Most commonly, two cases or options (treatments
set of attributes, and a health state is described by or health states) are seen at a time (hence the name
indicating the appropriate level of functioning on conjoint analysis) and a choice is made between
each attribute. For instance, in the EQ-5D the them. Adaptive conjoint analysis cases are paired
attributes are mobility, self-care, usual activities, according to a set of stated attribute weights and
pain/discomfort, and anxiety/depression. Each responses to previous options—using special
attribute is divided into three levels of severity software (Pieterse et al. 2010). Analysis of the
(no problem, some problems, and extreme prob- data is based on random utility theory. These
lems). By combining one level from each of the methods have predominantly been used in health
five attributes, a total of 35, that is, 243 EQ-5D services research to assess correlates of prefer-
health states are defined. The formula for ences, such as sociodemographic characteristics
assigning utilities to these states is based on of (potential) service users and to influence policy
utilities that have been obtained in a sample from decision-making. The adaptive methods are
the general public, in part from direct measure- finding their way in micro-level decision-making,
ment and in part from application of MAUT (as in to support values clarification, as described in the
the Health Utilities Index) or statistical inference next section.
(as in the EQ-5D), to fill in values not measured
directly. Based on this formula (for the EQ-5D,
e.g., see Dolan 1997), premeasured utilities from Measurement of Preferences
the general public are thus available for these at the Micro Level
systems (Russell et al. 1996). In a cost-
effectiveness study it suffices to map the treat- Assessments of preference for specific options,
ment outcomes (the health states) onto the rather than outcome states, are tailored to the
descriptive system—using a patient questionnaire clinical problem at hand and will reflect the real-
based on the descriptive system—and to use the life situation more than does the utility assessment.
scoring formula to obtain utilities from the gen- In health services research, at the meso-level,
eral public for the health states indicated by the assessment of treatment preferences informs cut-
patients. In this way, standardization over studies offs in guidelines above or below which treatment
is obtained. All researchers use the same utility is indicated. For example, patient preferences were
set, and cost-effectiveness ratios are comparable. incorporated in the decision to recommend
Whereas the aim of these decomposed tech- chemotherapy at a benefit in overall 10-year sur-
niques is mostly to assess holistic valuations of vival of 5 % in the Dutch breast cancer treatment
health states or treatments via decomposition, guidelines (Bontenbal et al. 2000). Alternatively,
other techniques, such as conjoint analysis and preferences can be assessed to define profiles of
discrete choice experiments, aim to measure how patients for whom a particular option is more
treatment or health state attributes are valued germane than for others. At the micro or individ-
per se. ual patient level, decisions about treatment
Conjoint Analysis, developed to examine and health care management ought to reflect
consumer preferences in marketing is increas- individual patients’ preferences (Kassirer 1994).
ingly used in health to assess attribute prefer-
ences. Similar to the decomposition techniques
described above, participants judge hypothetical Treatment Tradeoff Method
cases (health states or treatments) that are
described in terms of combinations of attributes at The treatment or probability tradeoff method was
particular levels. Statistical analysis reveals the developed to assess participants’ strength of
attribute level utilities (Ryan and Farrar 2000). preference for one health management option
1 What Are Values, Utilities, and Preferences? A Clarification … 9
relative to another, usually treatments. In this management preferences. The name is confusing
method, preferences for combined process- as these interventions really are aimed at eliciting
and-outcome paths are elicited in the following and clarifying preferences. VCM include any
way. The patient is usually presented with two methods “that are intended to help patients
clinical options, for example, treatments A (e.g., evaluate the desirability of options or attributes
no adjuvant treatment) and B (e.g., adjuvant of options within a specific decision context, in
chemotherapy), which are described with respect order to identify which option he/she prefers”
to (probabilities of) benefits (e.g., additional (Fagerlin et al. 2013). These VCM can also be
probability of 5-year survival) and side-effects used to measure individual preferences (Fraenkel
(e.g., nausea, hair loss, and fatigue), and is asked et al. 2006).
to state a preference for an option. If treatment A Many and very different types of VCM exist.
is preferred, the interviewer systematically either In treatment-related decision-making, interven-
increases the probability of benefit from treatment tions described as VCM include balance scales
B, or reduces the probability of benefit from (O’Connor et al. 1998); rating (Feldman-Stewart
treatment A (and vice versa if treatment B is et al. 2006) or ranking (Sheridan et al. 2010) the
preferred at the outset). Which treatment aspects importance of risks or benefits of options; indi-
are altered and in which direction, is decided cating whether each piece of information pushes
upon beforehand, according to the relevant clin- one toward or away from a given choice (Smith
ical characteristics and the research question et al. 2010); or listing reasons (Abhyankar et al.
(Llewellyn-Thomas et al. 1996). The patient’s 2011). They can also consist of having an open
willingness to accept side-effects of one treatment discussion about attributes of interest
or forego benefits of the alternative treatment (Matheis-Kraft and Roberto 1997). Evidence on
determines the patient’s relative strength of the effects of using VCM in the context of patient
preference. This general approach has been decision aids is still limited, but there are indi-
adapted to a variety of treatment decisions, cations that it improves decision processes
including adjuvant chemotherapy in breast cancer (Fagerlin et al. 2013).
(Levine et al. 1992), treatment of Lupus Nephritis There is a little evidence suggesting how
(Fraenkel et al. 2002), and radiotherapy for rectal patients actually clarify the personal importance
cancer (Pieterse et al. 2007). In all cases, prefer- they associate with different health management
ence strength is idiosyncratic to the original options, such as how they weigh pros and cons
decision problem, that is, relative to the specific within a decision, and thus how best to support the
alternatives that were presented. The method can process. Further, since preferences in health are
be used to support individual treatment deemed constructive, there is no way to measure
decision-making and has been applied “at the “true” preferences since they are formed in the
bedside” using decision boards as visual aids process of elicitation. From a cognitive psycho-
(Levine et al. 1992). logical perspective, VCM should aim to facilitate
one or more of the following processes: help
optimize individuals’ mental representations of
Values Clarification Methods the decision and the options; encourage individ-
uals to consider all potentially appropriate
At the micro-level, so-called patient decision aids options; delay the selection of an initially favored
have been developed to help individuals facing option; facilitate the retrieval of relevant values
challenging health decisions make specific and from memory; facilitate the comparison of options
deliberative choices (Stacey et al. 2011). As a and their attributes; and offer time to decide
part of these interventions, components referred (Pieterse et al. 2013). These recommendations
to as “values clarification methods” (VCM) can were formulated based on commonalities between
be included to help elucidate individuals’ health the four process theories of decision-making
10 A.H. Pieterse and A.M. Stiggelbout
(differentiation and consolidation theory, image method, “lead time TTO” is currently under study
theory, parallel constraint satisfaction theory, as a way to possibly overcome the problem
fuzzy-trace theory), for which evidence has been (Augustovski et al. 2013).
gathered though mostly outside of the health care At the micro-level, research revolves around
context. the evaluation of how effective the VCM are at
clarifying preferences. A challenge at the
micro-level for future research lies, therefore, in
Key Directions for Future Research designing theory-based VCM and outcome mea-
sures—where the theory chosen should help in
At the macro-level, most of the researches that are selecting outcome measures that the intervention
currently performed in utility assessment relate to is expected to affect (Pieterse et al. 2013).
the classification systems, such as the EQ-5D.
This is likely because these have the most direct
practical application in cost-utility analyses, Conclusion
which in turn are mandatory for reimbursement
decisions in many health care systems around the Preferences refer to very different entities at the
world. The assessment of holistic utilities, for macro-, meso-, and micro-levels of health-related
example using the TTO, is typically seen in decision-making. At each of these levels, we
purely scientific work, without direct practical recommend to save the term value for abstract,
application. The challenges for the EQ-5D mostly trans-situational judgments. The most adequate
lie in improving the descriptive systems, for process of preference elicitation is a function of
example, by adding levels to the attributes. A re- the goal of assessing individuals’ health-related
curring issue is the actual content of the classifi- priorities and depends on the level of health care
cation systems, and whether the traditional decision-making. Particularly at the level of the
dimensions, generally based on the WHO defi- individual patient, more research is needed on the
nition of health and incorporating physical, psy- clarification of patient preferences.
chological, and social functioning, should not
be replaced by a capability approach or by Box 1. Definition of values, utilities
dimensions of subjective well-being (Coast et al. and preferences in a health care deci-
2008). sion context
The elicitation of utilities is quite an abstract Values
task, with which participants have been found to Abstract, trans-situational judgments about
have difficulties (Edelaar-Peeters et al. 2014). intermediate or terminal goals that guide the
Interviewer help is therefore generally needed, evaluation of states or selection of behav-
even though web-based administration would iors and are ordered by relative importance
highly reduce costs. Future research should find Utilities
ways to mimic the help that interviewers give as Summary measures of how health states
part of web-based administration. realize our ultimate values or goals; should
Moreover, conventional approaches to the TTO be measured in specific ways resulting in a
are problematic when evaluating health states that number between 0 and 1 and are most
are perceived to be worse than death. The TTO often applied in expected utility decision
requires fundamentally different tradeoffs tasks for analyses and in cost-utility analyses at the
the valuation of states better and worse than death macro and meso-decision-making level
(Tilling et al. 2010). An alternative elicitation
1 What Are Values, Utilities, and Preferences? A Clarification … 11
Gregory, R., Lichtenstein, S., & Slovic, P. (1993). Valuing analysis to assess patient preferences: An application
environmental resources—A constructive approach. in oncology. Health Expectations, 13, 392–405.
Journal of Risk and Uncertainty, 7, 177–197. Rohan, M. J. (2000). A rose by any name? The values
Jansen, S. J., Stiggelbout, A. M., Wakker, P. P., Vliet construct. Personality and Social Psychology Review,
Vlieland, T. P., Leer, J. W., Nooy, M. A., et al. (1998). 4, 255–277.
Patients’ utilities for cancer treatments: A study of the Russell, L. B., Gold, M. R., Siegel, J. E., Daniels, N., &
chained procedure for the standard gamble and time Weinstein, M. C. (1996). The role of
tradeoff. Medical Decision Making, 18, 391–399. cost-effectiveness analysis in health and medicine.
Kassirer, J. P. (1994). Incorporating patients’ preferences JAMA, 276, 1172–1177.
into medical decisions. New England Journal of Ryan, M., & Farrar, S. (2000). Using conjoint analysis to
Medicine, 330, 1895–1896. elicit preferences for health care. BMJ, 320, 1530–1533.
Levine, M. N., Gafni, A., Markham, B., & MacFarlane, Schwartz, S. H. (1994). Are there universal aspects in the
D. (1992). A bedside decision instrument to elicit a structure and contents of human values. Journal of
patient’s preference concerning adjuvant chemother- Social Issues, 50, 19–45.
apy for breast cancer. Annals of Internal Medicine, Schwartz, S. H., & Bilsky, W. (1987). Toward a universal
117, 53–58. psychological structure of human-values. Journal of
Llewellyn-Thomas, H. A., Williams, J. I., Levy, L., & Personality and Social Psychology, 53, 550–562.
Naylor, C. D. (1996). Using a trade-off technique to Sheridan, S. L., Griffith, J. M., Behrend, L., Gizlice, Z.,
assess patients’ treatment preferences for benign Jianwen, C., & Pignone, M. P. (2010). Effect of
prostatic hyperplasia. Medical Decision Making, 16, adding a values clarification exercise to a decision aid
262–282. on heart disease prevention: A randomized trial.
Matheis-Kraft, C., & Roberto, K. A. (1997). Influence of Medical Decision Making, 30, E28–E39.
a values discussion on congruence between elderly Smith, S. K., Trevena, L., Simpson, J. M., Barratt, A.,
women and their families on critical health care Nutbeam, D., & McCaffery, K. J. (2010). A decision
decisions. Journal of Women & Aging, 9, 5–22. aid to support informed choices about bowel cancer
O’Connor, A. M., Tugwell, P., Wells, G. A., Elmslie, T., screening among adults with low education: Ran-
Jolly, E., Hollingworth, G., et al. (1998). A decision domised controlled trial. BMJ, 341, c5370.
aid for women considering hormone therapy after Stacey, D., Bennett, C. L., Barry, M. J., Col, N. F., Eden,
menopause: Decision support framework and evalua- K. B., Holmes-Rovner, M., et al. (2011). Decision aids
tion. Patient Education and Counseling, 33, 267–279. for people facing health treatment or screening
Ormond, K. E., Banuvar, S., Daly, A., Iris, M., Minogue, decisions. Cochrane Database of Systematic Reviews,
J., & Elias, S. (2009). Information preferences of high 10, CD001431.
literacy pregnant women regarding informed consent Stiggelbout, A. M., & De Haes, J. C. J. M. (2001). Patient
models for genetic carrier screening. Patient Educa- preference for cancer therapy: An overview of mea-
tion and Counseling, 75, 244–250. surement approaches. Journal of Clinical Oncology,
Payne, J. W., Bettman, J. R., & Johnson, E. J. (1992). 19, 220–230.
Behavioral decision research: A constructive process- Stiggelbout, A. M., & De Vogel-Voogt, E. (2008).
ing perspective. Annual Review of Psychology, 43, Health-state utilities: A framework for studying the
87–131. gap between the imagined and the real. Value in
Payne, J. W., Bettman, J. R., & Schkade, D. (1999). Health, 11, 76–87.
Measuring constructed preferences: Towards a building Street, R. L, Jr., Elwyn, G., & Epstein, R. M. (2012).
code. Journal of Risk and Uncertainty, 19, 243–270. Patient preferences and healthcare outcomes: An
Pfützner, A., Schipper, C., Niemeyer, M., Qvist, M., ecological perspective. Expert Reviews in PharmacoE-
Löffler, A., Forst, T., et al. (2012). Comparison of conomics and Outcomes Research, 12, 167–180.
patient preference for two insulin injection pen devices Sutherland, H. J., & Till, J. E. (1993). Quality of life
in relation to patient dexterity skills. Journal of assessments and levels of decision making: Differenti-
Diabetes Science and Technology, 6, 910–916. ating objectives. Quality of Life Research, 2, 297–303.
Pieterse, A. H., De Vries, M., Kunneman, M., Stiggel- Ter Hoeven, C. L., Zandbelt, L. C., Fransen, S., De Haes,
bout, A. M., & Feldman-Stewart, D. (2013). H., Oort, F., Geijsen, D., et al. (2011). Measuring
Theory-informed design of values clarification meth- cancer patients’ reasons for their information prefer-
ods: A cognitive psychological perspective on patient ence: Construction of the Considerations Concerning
health-related decision making. Social Science and Cancer Information (CCCI) questionnaire.
Medicine, 77, 156–163. Psycho-Oncology, 20, 1228–1235.
Pieterse, A. H., Stiggelbout, A. M., Baas-Thijssen, M. C., Tilling, C., Devlin, N., Tsuchiya, A., & Buckingham, K.
Van de Velde, C. J., & Marijnen, C. A. (2007). Benefit (2010). Protocols for time tradeoff valuations of health
from preoperative radiotherapy in rectal cancer treat- states worse than dead: A literature review. Medical
ment: Disease-free patients’ and oncologists’ prefer- Decision Making, 30, 610–619.
ences. British Journal of Cancer, 97, 717–724. Torrance, G. W. (1976). Preferences for health states: An
Pieterse, A. H., Stiggelbout, A. M., & Marijnen, C. A. empirical evaluation of three measurement techniques.
(2010). Methodologic evaluation of adaptive conjoint Socio-Economic Planning Sciences, 10, 129–136.
1 What Are Values, Utilities, and Preferences? A Clarification … 13
Torrance, G. W., & Feeny, D. (1989). Utilities and Unic, I., Stalmeier, P. F., Verhoef, L. C., & Van Daal, W.
quality-adjusted life years. Journal of Technological A. (1998). Assessment of the time-tradeoff values for
Assessment in Health Care, 5, 559–575. prophylactic mastectomy of women with a suspected
Torrance, G. W., Feeny, D., & Furlong, W. (2001). Visual genetic predisposition to breast cancer. Medical Deci-
analog scales: Do they have a role in the measurement sion Making, 18, 268–277.
of preferences for health states? Medical Decision Vela, K. C., Walton, R. E., Trope, M., Windschitl, P., &
Making, 21, 329–334. Caplan, D. J. (2012). Patient preferences regarding
Torrance, G. W., Feeny, D. H., Furlong, W. J., Barr, R. 1-visit versus 2-visit root canal therapy. Journal of
D., Zhang, Y., & Wang, Q. (1996). Multiattribute Endodontics, 38, 1322–1325.
utility function for a comprehensive health status Von Neumann, J., & Morgenstern, O. (1944). Theory of
classification system: Health utilities index mark 2. games and economic behavior. Princeton: Princeton
Medical Care, 34, 702–722. University Press.
Tversky, A., & Kahneman, D. (1992). Advances in Wakker, P., & Stiggelbout, A. (1995). Explaining distor-
prospect-theory—Cumulative representation of tions in utility elicitation through the rank-dependent
uncertainty. Journal of Risk and Uncertainty, 5, model for risky choices. Medical Decision Making,
297–323. 15, 180–186.
Decision Architectures
2
Bradford W. Hesse
published in 1952 (Daniels and Hill 1952). ways in which clinical teams and their patients
Staying abreast of the medical literature in 1952 make ongoing judgments related to care (Institute
when scientific publications from randomized of Medicine and McClellan 2008).
controlled trials were few and far between may
have been a reasonable task. By 2003, though,
the U.S. National Library of Medicine estimated Failing to Support Patient Engagement
that it had been adding almost 10,000 new arti- in Decision-Making
cles per week within its own online archives—a
number that represented only about 40 % of all Not only have strains in the system made it dif-
medical articles published worldwide in ficult to support physicians’ decision needs, but it
biomedical and clinical journals. To stay abreast may also be failing to support patients’ informed
of the exploding research base just within one participation in their own care. In other sectors of
specialty would require practitioners to read the economy, consumers have been learning to
upwards of 20 articles per day, 365 days a year interact with complex information systems in
(Shaneyfelt 2001). That is an impossible task. ways that are responsive, user-centered, and
To complicate matters further, medical empowering. Whether relying on a seamless
decision-making is by its nature becoming much network of Automated Teller Machines and
more data intensive (Topol 2012). By one online bill paying to access their own money in
account, the average number of facts a physician any currency around the world, benefiting from
would need to bear in mind when making a the unparalleled safety records of an integrated
decision about a patient’s treatment in 1990 was air traffic control system, or simply reaching their
around five (Smith et al. 2012). These were own personal destinations with the help of a
decisions based on an evaluation of the clinical user-friendly Global Positioning System (i.e.,
phenotype only; that is, decisions in which the GPS), consumers have come to depend on
medical practitioner’s task would primarily be to state-of-the-art information architectures to nav-
evaluate the signs and symptoms accompanying igate the choices of their daily lives (Obama
a chief complaint and then to apply a type of 2012). This has not necessarily been the case in
one-size-fits-all formulary to match the hypoth- health care.
esized diagnosis with a population-based guess Consider an all too familiar scenario. A pa-
on treatment. Medicine is becoming much more tient waits until a medical problem has pro-
predictive and personalized than that today gressed to the point of extreme discomfort or
(Culliton 2006; Collins 2010). Genomic indica- reduced functioning before visiting a physician.
tors, functional expressions of DNA transcrip- If this is the patient’s first visit to a physician’s
tions, and molecularly precise assessments of office, the receptionist may ask her to fill out a
treatment efficacy will create a reality in which form listing all of the medications she may be
the treating physician must take into account currently taking; all of the medications she had
1000 or more facts over the course of a patient’s previously taken; any remembered side effects or
treatment. Cognitive research suggests that the deleterious interactions from previous treatments;
number of facts a human information processor any persistent complaints or recollections of
can manage at any given time hovers at around diagnosed conditions; blood type or other rele-
7 (±2) (Miller 1956). Under the projections of vant biologic assessments; and a cursory expla-
precision medicine, it is difficult to consider a nation for the purpose of her visit. More often
future that does not explicitly include the design than not, these patients will not have brought
of efficient decision architectures to improve the their own records with them, and they will likely
2 Decision Architectures 17
find it impossible to remember the technical services—while at the same time the patient’s
names of pharmaceuticals they might have taken sense of decisional control grows more tenuous
previous to the visit. They may even find it dif- (Taplin and Rodgers 2010).
ficult to remember the names and prescribed
dosages of medications they are taking currently.
After completing the intake forms, the patient The Consequences
is then escorted to an examination area where a
few cursory measures of weight, height, blood The consequences of these systemic strains on
pressure, and temperature are added to the the decision systems in medicine and health have
patient’s newly initiated file. The results are become severe, especially when considered at the
recorded, and the patient is instructed to disrobe population level. Up until 1999, hospital staff
and wait until the doctor is ready. After a few intuitively knew that avoidable errors were
seemingly interminable moments the physician occurring with some regularity within their sys-
comes into the room, quickly peruses the chart tem. In 1999, the Institute of Medicine put a
and briefly asks a few clarifying questions, per- population level count on the consequences of
forms a focused physical examination, and then those errors by estimating that some 48,000–
writes a prescription. To the patient, the pre- 98,000 deaths occurred annually due to some
scription appears to be written in code, with type of avoidable medical error (Kohn et al.
abbreviated Latin terms for mode and frequency 2000). That figure exceeded the number of
of administration (e.g., “p.o.” for per os, or by individuals dying from AIDS, breast cancer, or
mouth; “q 3 h” for quaque 3 hora, or “every automobile accidents at the time it was reported.
three hours”). If there is accompanying literature, Actuarial data from 2008 suggested that the
the formatted small print and technical jargon annual cost of measurable medical errors result-
will look more like a legal disclaimer than a set ing in direct harm to patients as assessed through
of coherent, easy-to-follow instructions medical claims was $17.1 billion (Van Den Bos
(McClellan 2008). et al. 2011). The practice of medicine used to be
In the event the physician is not able to reach “simple, largely ineffective, and relatively safe,”
a firm diagnosis within the 15 min customarily argued Sir Cyril Chantler in an oft-quoted Lancet
allotted for a clinical encounter, the patient may article. With advances in modern medical tech-
be given instructions for visiting a laboratory or nology, medicine is rapidly becoming much
specialist. Responsibility for the handoff is fre- more effective, but it is also becoming much
quently put on the shoulders of the patient with more complex and extraordinarily dangerous
instructions to make the follow-up appointment, (Chantler 1999).
to request that the appropriate records be trans- Contrast these numbers with the number of
ferred, and to ensure that insurance will cover the avoidable deaths from an equally
extra expense. The patient will then proceed from technology-dependent, complex system: avia-
one appointment to the next, repeating the chief tion. On April 27, 2012, the U.S. National
complaint and brief history along the way. Office Transportation Safety Board (NTSB) reported
staff will take new notes and record them into an that there were zero fatalities involving U.S. air
expanding chart, though the patient may not have carriers or commuter operations in 2011.
any idea what information or how much infor- According to the press release, 2011 was the
mation has been transferred between offices. The second straight year in which no fatalities were
patient’s files grow—with more insurance forms observed from air travel among U.S. carriers
and more technically framed descriptions of (National Transportation Safety Board 2012).
18 B.W. Hesse
This remarkable statistic was reported in spite of all problems lies within people—that the way to
the fact that there was a small increase in the fix problems is through more education, more
number of observed accidents or near misses in rewards, or a culture of “blame and shame.”
the industry overall in 2011. Even when factor- Neither of these worldviews is sufficient on its
ing in deaths from previous years, the number of own to guarantee safety and efficiency in a
fatalities is astonishingly low in the airline technology-dependent industry such as aviation
industry. The NSTB had reported that the overall or health care. Separated, they are leading to a
number of fatalities over the previous decade was hidden epidemic of error and chaos (Vicente
about three deaths per 10 billion passenger miles 2003).
traveled per year (Insurance Information Institute What human factors researchers discovered
2012). That makes a stark contrast to the 48,000– when performing root-cause analyses of acci-
98,000 deaths from medical error estimated to dents and near misses in aviation (i.e., using
occur annually during the same period (Kohn specialized analytic techniques to identify the
et al. 2000). originating cause, rather than symptom, of a
Why is there such a contrast between these critical error) was that the technological and
two sectors of the economy? This is the question human subsystems are inextricably linked and
posed by Donald Berwick, Director of the Cen- must be studied together to improve performance
ters of Medicare and Medicaid from 2009 to within systems. This combined, or transdisci-
2011. He has concluded that there are significant plinary (Stokols et al. 2008), view is based on the
differences in the contextual fabrics of medicine observations: (a) that technical systems have
and aviation that account for these vast differ- social consequences; (b) that social systems have
ences. In medicine, he observed, a perverse technical consequences; (c) that systems engi-
system of “fee-for-service” incentives has created neers do not create technologies, they create
a decentralized medical environment in which sociotechnical systems; and (d) that progress
adherence to evidence-based approaches for within these systems must be gained by under-
treatment is spotty; the use of risky, and often standing how people and technologies interact.
unnecessary, treatments is prevalent; and an In medicine this view is referred to as a
assessment of end-to-end quality control is sociotechnical perspective on health system
infeasible (Berwick 2002; Berwick et al. 2008). redesign (Coiera 2004). Within the National
National health care reform efforts are working to Academy of Sciences, this focus has been
change those incentives while establishing the referred to as “Human System Integration”
data infrastructure needed to track patients across (Committee on Human-System Design, N.R.C.
health systems (President’s Council of Advisors 2007).
on Science and Technology 2010).
Another reason why the aviation system may
be superior in its control of error is the invest- Nudging Best Practice: A Behavioral
ment it has made in understanding the psychol- Economics Approach
ogy of human technology interaction, a field
known historically as “human factors” research. Behavioral economists Thaler and Sunstein
There have been two antiquated cultures in popularized the notion of improving the systemic
modern times, reasoned human factors scientist architectures upon which individuals make
Kim Vincente: one based on a mechanistic, day-to-day decisions in their book “Nudge:
engineering view of the world and one based on Improving Decisions about Wealth, Health, and
a very humanistic, social view of the world. The Happiness.” What these two authors were able to
first view assumes that most problems can be do was bring together decades of research in
solved through new and better technology—that human factors, cognitive psychology, and social
the answer to bad technology is more technol- psychology to dispel the Cartesian notion popu-
ogy; while the second assumes that the locus of lar in classical economics that human judgment
2 Decision Architectures 19
Subjective
Chief complaint Health Information Exchange
Patient Reported Outcomes
Objective
Clinical measures
Laboratory findings Hospital Based Hospital Based
Sensor data EHR Data EHR Data
Assessment
Diagnosis
Categorical reporting
Prognosis
Plan
Treatment planning
Self-care planning Patient
Medical Hospital Medical
Post treatment &
Team System Researcher
Surveillance Family
Fig. 2.1 Decision support within a fully connected, data-driven environment of care
is universally rational, logical, and deliberative; reached 100 % (Jamoom et al. 2012). At the left
while at the same time suggesting how new of Fig. 2.1 is a depiction of the various types of
supports could be constructed within systems to data signals that are available for compilation
overcome those limitations and improve overall within an EHR system. The inputs are organized
decision-making (Thaler and Sunstein 2009). following the traditional SOAP notes format
Hesse et al. 2011 extended the architectural utilized in paper-based-charts; that is, with sub-
theme into the realm of contemporary health care jective and qualitative descriptions of the chief
by examining how a movement toward inter- complaint included in the record along with ob-
connected data systems could be marshaled to jective measures from medical tests and labora-
create a robust foundation for evidence-based tory findings, diagnostic conclusions and other
practice. professional assessments, and ongoing plan for
Figure 2.1 offers an overview of what an treatment and in some cases long-term care or
interconnected data system in health care might vigilance.
offer decision makers. Once just a theoretical As these data are brought together, they can
vision, these types of systems are becoming a be made available—separately or jointly—to
reality in many health care systems around the each of the stakeholders in an expanded view of
world. In the United States, data from the Centers the care team. According to a literature review
for Disease Control and Prevention show that sponsored by the Agency for Health care
adoption of Electronic Health Record Research and Quality (AHRQ), data at the
(EHR) systems has risen above the 50 % pene- practice level can be harnessed for clinical deci-
tration mark among all physicians. Penetration sion support to: “(1) remind clinicians of things
within Health Maintenance Organizations has they need to do, (2) provide information when
Another random document with
no related content on Scribd:
XLIV.—No. 16, LITTLE WILD STREET
(Demolished).
In 1590 William Short, the same who ten years later bought
Rose Field, purchased of John Vavasour two messuages, two gardens
and four acres of land, with appurtenances, in St. Giles.[490] The
precise position of the property is not stated, but from evidence
which will be referred to, it is known that it lay to the west of Drury
Lane, and comprised The Greyhound inn in Broad Street, with land
to the south lying on both sides of what is now Short’s Gardens.
A portion of this property he leased,[491] in 1623–
4, to Esmé Stuart, Earl of March (afterwards Duke of
Lennox), for a term of 51 years as from Michaelmas,
1617. It is possible to ascertain within a little the
boundaries of this part of the Short estate. In a deed[492]
dated 10th January, 1614–5, relating to Elm Field, the
land lying between Castle Street and Long Acre, the Esmé
Stuart,
northern boundary is stated to be “certain closes called Seigneur
by the name of Marshlands alias Marshlins, and a D’Aubigny,
garden sometime in the tenure of William Short or his Duke of
assignes”; and in a later deed,[493] dated 2nd February, Lennox.
1632–3, relating to a portion of the same field, the
northern boundary, said to be 249 feet distant from
Long Acre, is referred to as “a way or back lane of 20 feet adjoining
the garden wall of the Right Honble. the Duchess of Lenox.”
The distance of the “back lane” from Long Acre corresponds
exactly with that of the present Castle Street, and it is therefore clear
that this was the southern boundary. The property afterwards came
into the possession of the Brownlow family, and an examination of
the leases which were granted in the early part of the 18th century,
shows that it reached as far as Drury Lane on the east and Short’s
gardens on the north. On the west it stretched as far as Marshland.
[494]
Whether the house leased to the Earl of March was one of the
two (the other being The Greyhound) purchased by Short in 1590, or
a house quite recently built, there is no evidence to show.
The Earl, in February, 1623–4, succeeded to the dukedom of
Lennox, and on 30th July of the same year he died. His widow[495]
continued to reside at the house. Letters from her, headed “Drury
Lane,” and dating from 1625 to 1629, are extant,[496] and she also, in
1628, joined with other “inhabitants adjoining the house of the
Countess of Castlehaven, in Drury Lane,” in a petition to the Privy
Council.[497] There is, therefore, ample evidence that she actually
resided at the house.
In 1632 she married James Hamilton, second Earl of
Abercorn, and died on 17th September, 1637, leaving to her husband,
in trust for their son James, “all that my capitall house, scituate in
Drury Lane.”[498]
The Earl sold the remainder of the lease[499] to the Duchess’s
cousin, Adrian Scroope, who apparently let the house, as the Subsidy
Roll for 1646 shows the “Earl of Downe” as occupying the premises.
[500]
In 1647 Sir Gervase Scroope, Adrian’s son, sold the lease to Sir
John Brownlow,[499] who certainly acquired the freehold also, though
no record of the transaction has come to light. Finding the house too
large[501] Sir John divided it in two, and in 1662 Lady Allington[502]
was paying a rent of £50 for the smaller of the two residences.[499] Sir
John died in November, 1679. By his will[503] (signed 10th April,
1673) he left to his wife all the plate, jewels, etc. “which shall be in
her closett within or neare our bedd chamber at London in my house
at Drury Lane ... and the household stuffe in the said house, except
all that shall then be in my chamber where the most part of my
bookes and boxes of my evidences are usually kept, and except all
those in the same house that shall then be in the chamber where I
use to dresse myselfe, both which chambers have lights towardes the
garden.” He also left to his wife “that part of my house in Drury Lane
which is now in my own possession for her life if she continue my
widowe,” together with “that house or part of my house wherein the
Lady Allington did heretofore live, ... by which houses I meane yards,
gardens and all grounds therewith used”; and moreover the furniture
“of two roomes in my house in Drury Lane where I use to dresse
myself, and where my evidences and bookes are usually kept.”
The estate afterwards came into the hands of Sir
John Brownlow, son of his nephew, Sir Richard
Brownlow, who at once took steps to develop the
property, letting plots on building lease for a term of
years expiring in 1728. Except in one case, information is
not to hand as to the date on which these leases were
Brownlow. granted, but in that instance it is stated to be 21st May,
1682,[504] a date which may be regarded as
approximately that of the beginning of the development of the
interior part of the estate by building,[505] though at least a part of the
frontages to Drury Lane and Castle Street had been built on before
1658 (see Plate 3).
At the same time (circ. 1682) apparently Lennox House was,
either wholly or in part, demolished. A deed of 1722[506] relates to the
assignment of two leases of a parcel of ground “lately belonging to
the capital messuage or tenement of Sir John Brownlow then in part
demolished, scituate in Drury Lane, in St. Giles, sometime called
Lenox House.” The description is obviously borrowed from the
original leases, since reference is also made to “a new street there
then to be built, intended to be called Belton Street,” which street
was certainly in existence in 1683.[507] What is apparently Lennox
House is shown in Morden and Lea’s Map of 1682 as occupying a
position in the central portion of the estate, with a wide approach
from Drury Lane, and this is to a certain extent confirmed by the
tradition that the first Lying-In Hospital in Brownlow Street
(occupying the site of the present No. 30) was a portion of the
original building. It is remarkable, however, that no hint of a house
in this position is given either in Hollar’s Plan of 1658 (Plate 3) or in
Faithorne’s Map of the same date (Plate 4).
The name of Brownlow Street was in 1877 altered to Betterton
Street.
XLVII.–XLVIII.—Nos. 24 and 32,
BETTERTON STREET.
General description and date of
structure.
No. 24, Betterton Street, dating from the 18th century, must at
one time have been a fine residence, but there is now nothing in it to
record. The doorcase is illustrated on Plate 35.
No. 32 also dates from the 18th century. Attached to these
premises is a boldly recessed carved wooden doorcase of interesting
design, illustrated on Plate 36. The interior of the house contains a
wood and compo chimney piece of some interest in the front room of
the ground floor, and one of white marble, relieved with a little
carving and red stone inlay, in the corresponding room on the floor
above.
Condition of repair.
The houses are in fair repair.
Biographical notes.
The sewer ratebook for 1718 shows “John Bannister” in occupation of
No. 32. This was probably John Bannister, the younger, “who came from an
old St. Giles’s family, his father having been a musician, composer and
violinist, and his grandfather one of the parish waits. He himself was in the
royal band during the reigns of Charles II., James II., William and Mary,
and Anne, and played first violin at Drury Lane theatre, when Italian operas
were first introduced into England.”[508]
In the Council’s collection are:—
No. 24, Betterton Street—General exterior (photograph).
[509]No. 24, Betterton Street—Entrance doorway (measured
drawing).
[509]No. 32, Betterton Street—Entrance doorway (photograph).