Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
Third Revised Draft: Please Do Not Quote or Cite Without Permission
Bounded Rationality and the Conceptual Underpinnings of Health Policy:
A Rationale and Roadmap for Addressing the Challenges of Choice in Medical Settings
Mark Schlesinger, Ph.D.*
Professor at the School of Public Health and Fellow at the Institution for Social and Policy Studies, Yale University
Brian Elbel, M.P.H.
Doctoral Candidate, Yale University
September 15, 2006
* Corresponding Author: Please contact Dr. Schlesinger at Room 304 LEPH, School of Public Health, Yale University, 60 College St. New Haven CT 06520. (203) 785-4619; [email protected]
Acknowledgements: The ideas expressed in this manuscript benefited from a thoughtful reading by
David Adler and feedback from the seminar in health policy and services research in our school’s Division of Health Policy and Administration. We retain responsibility for any persisting errors, misrepresentations or biases.
0
Abstract
The implementation of Medicare Part D illustrates some of the foibles of maximizing the number
of choice options in an effort to empower consumers. Findings from research on bounded
rationality suggest that people frequently make choices on the basis of decision heuristics that
cause them to: (a) ignore relevant information about uncertain prospects, (b) use information in
ways that systematically bias their expectations about future events, (c) partition their choices in
ways that obscure vital trade-offs, and (d) be perversely influenced and at times adversely
affected by expanding their choice options. Yet these findings, along with the early experience
with Medicare Part D, appear to be lost on policymakers who continue to rely on expanded
choices and more information for consumers as a means to improve program performance.
In this paper, we explore both the promise and potential pitfalls of adapting health policy to
reflect bounded rationality in consumer choice. We identify some 40 “anomalies” in the
literature on consumer behavior in health settings. These suggest that the experience with
Medicare Part D is but one example of a more general pattern: that increasing the number of
choices or enhancing information about those options often does not yield the expected results.
Policies that fail to account for bounded rationality, we argue, are likely to be ineffective and in
some cases counterproductive. But to adapt health policies to more directly deal with decision
heuristics and the biases they entail, we suggest that reformers must confront three barriers to
policy change: lagged learning on the part of policymakers, blame avoidance among elected
officials, and concerns about excessive government paternalism in American health care.
1
Bounded Rationality and the Conceptual Underpinnings of Health Policy: A Rationale and Roadmap for Future Research and Policy Development
The implementation of Medicare Part D yielded some paradoxical outcomes. In the face of
looming federal deficits, Congress expanded Medicare in 2003 to cover outpatient prescription
drugs, a benefit that advocates for the elderly portrayed as desperately needed. In accord with
polling data indicating that elders wanted a choice of insurance plans, Part D offered a lot of
choice. The typical beneficiary could select among 40 different drug discount cards offered
during the first two years of the program (Hanoch and Rice, 2006). When insurance coverage for
prescription drugs first became available in the fall of 2005, most beneficiaries had an equally
large number of options. Although there was some variation regionally, on average Medicare
recipients could choose from among three dozen different free-standing drug-benefit plans, in
addition to coverage offered by HMOs that participated in Medicare (Pear, 2006).
Beneficiaries’ response to the new drug benefits has been mixed. Fewer than 20 percent obtained
a drug discount card; only 24 percent of beneficiaries from low-income households, for whom
the card was completely free (Hanoch and Rice, 2006). Of beneficiaries who had no other source
of coverage for prescription drug expenses (estimates range from 15 to 18 million), 10.4 million
had signed up for a drug benefit plan by the close of the enrollment period in May of 2006
(Kaiser Family Foundation, 2006a). Remarkably few beneficiaries who already had some form
of coverage took advantage of their expanded opportunities under Part D to switch to a new plan.
These modest take-up rates were surprising in light substantial penalties for delayed enrollment,
ready availability of information on plans, and estimates by the Centers for Medicare and
Medicaid Services (CMS) that showed coverage under Part D offered substantial cost savings for
many elders. Equally surprising was beneficiaries’ assessment of the program. Surveys
suggested that the most elders were quite dissatisfied: when asked to evaluate the new benefit,
two-thirds gave it a grade of C or lower (Kaiser Family Foundation, 2006b).
Limited enrollment and disenchantment were certainly not a consequence of inadequate choice.
Quite the contrary, many individuals involved with the selection process saw the challenges of
excessive choice as the heart of the problem. Surveys of the health care professionals most often
involved in assisting elders in their drug coverage choices, revealed that 55 percent of physicians
1
and 74 percent of pharmacists felt that their patients had too many coverage options to make
effective choices (Kaiser Family Foundation, 2006c; 2006d).1 Their assessment was echoed by
beneficiaries’ own judgments about the challenges that they faced in making choices. Of those
who failed to select a plan, 37 percent attributed this to the program being “too complicated”, 28
percent to difficulties deciding “which plan to chose” (Kaiser Family Foundation, 2006b).
These outcomes and attributions will come as little surprise to academics who study choices
among uncertain prospects, particularly those working in the fields of bounded rationality,
decision heuristics, and behavioral economics (Bernheim and Rangel, 2005; Camerer and
Lowenstein, 2004; Schwartz, 2004; Anderson, 2003). Research in this vein explores the ways in
which limited time and attention cause people to make choices using various forms of decision-
heuristics, shortcuts that reduce the cognitive burdens of choice but which can often bias
decisions in systematic ways. Research in this vein documents how large choice sets can leave
many people less satisfied with their choices and deter others from making any choices at all,
how an abundance of information can cause people to be information-avoidant, how choice sets
with more differentiated products can lead to choices that more poorly match consumers’
preferences, and how default arrangements can shape the choices that consumers make.
Though the checkered track record of choice under Medicare Part D may therefore have seemed
quite explicable to some scholars, it appears to have been essentially invisible to administrators
implementing the program and politicians overseeing its performance. Rather than respond to
concerns about excessive choice, CMS continues to promote increased options. During the next
open enrollment period (selecting benefits for the calendar year 2007), the number of drug plans
offered throughout the country will grow from 9 to 17. In every state except Hawaii and Alaska,
beneficiaries will have more than 50 drug plans to choose from; in 23 states, they will have more
than 55 options (Pear, 2006). Mark McClellan, administrator for CMS, lauds these changes as all
to the good. “As a result of robust competition and smart choices by seniors, plans are adding
drugs, removing options that were not popular, and providing more options with enhanced
coverage” (Pear, 2006, p.25). Conditions that many academics would see as undermining
effective choice are viewed by many policymakers as hallmarks of choice triumphant.
2
In this paper, we explore this gap between academics’ understanding and policymakers’
aspirations for choice in health care contexts. Our central contention is that the Medicare Part D
experience is far from unique. It is easy to dismiss its shortcomings as aberrations of a
beneficiary population that is too old or frail to take on conventional responsibilities and
decision-making tasks.2 We argue that the health care system is in fact rife with comparable
examples, affecting decision-makers of all ages and cognitive capacities:
Disseminating more accurate or more complete information about the threats to public health and well-being may actually induce the public to make less-informed choices,3 Providing the public with additional options for risk protection may actually leave them less
able to make a sound choices, or induce them to defer choices so that their risks increase,4 Expanding the choices may lead to less effective matching of consumer preferences with
provider styles or health plan characteristics and discourage dissatisfied consumers from switching to alternative sources of care or coverage,5 and Enhancing awareness of disparities in health and well-being may cause the groups most
affected by those disparities to be less responsive to episodes of mistreatment.6
An abundance of evidence can be found in support of each of these predictions. But it’s import
for health policy has been obscured because (a) the findings are scattered across a variety of
disciplines (psychology, behavioral economics, marketing, political psychology, medicine), (b)
the relevant findings are often by-products of studies focusing on other matters, so that they are
neither highlighted by the investigators nor easily identified in literature searches, and (c) many
of the findings relevant to health behaviors or health care appear in journals that are rarely read
by scholars who focus on health concerns (Frank, 2004). We hope that assembling this evidence
in a coherent manner will better convey the salience of these accumulated findings.
Our aspirations for this paper, however, go beyond documenting that health-related choices are
often made in ways inconsistent with policymakers’ expectations. We further argue that
conventional approaches for enhancing choice, by offering larger choice sets, more abundant
information, and stronger incentives for individuals to engage in choice, will be ineffective at
improving health system performance. We draw upon findings from the general literature on
heuristics and biases to explain why this is likely to be the case.
3
But an effective case for policy change must also address the factors that currently reinforce
conventional approaches to consumer choice. Efforts to identify policy-relevant insights from
bounded rationality have been previously pursued in three contexts: legal scholarship (Sunstein,
2002; Korobkin and Ulen, 2000; Jolls et al, 1998), research on the policy implications of risk
perceptions (Noll and Krier, 1990; Camerer and Kunreuther, 1989; Viscusi et al, 1987) and for
some limited aspects of health policy (Hanoch and Rice, 2006; Frank, 2004; Milstein and Adler,
2003; Hibbard et al, 2002; Sage, 1999). None of these efforts have examined the barriers that
could inhibit policymakers’ acceptance of findings from this literature or constrain the ways in
which policies could be adapted.
Identifying these barriers, and exploring ways in which they might be overcome, is our focus in
the second half of this paper. We attribute the persistence of conventional policy approaches to
three factors. The first involves the inevitable lag that exists between changes in prevailing
academic paradigms and their translation into policy discourse. As Keynes famously lamented
with regard to macro-economic policies, the guiding ideas for policymakers typically reflect the
favored notions of academics long dead (literally or figuratively), the ideas that were the norms
in higher education when those policymakers were themselves students. Because mainstream
economics has only recently begun to embrace the insights from behavioral economics (Rabin
and Thaler, 2001), the ideas that shape most policymakers’ understanding of consumer choice
remain imprinted by a neoclassical paradigm of decision-making. Bridging this paradigmatic gap
becomes one important challenge for promoting policy change: we propose a strategy for so
doing that relies on demonstrating that existing policies are not only behaviorally unrealistic and
ineffective, but likely to prove counterproductive, thwarting the very objectives that
policymakers endeavor to promote.
A second barrier to policymakers’ adapting to bounded rationality involves the political risks of
so doing. Health care issues fall into a general class of policy concerns in which politicians are
strongly motivated by the imperatives of “blame avoidance” (McGraw, 1991; Weaver, 1986). A
policy paradigm that “empowers” personal choice implicitly assigns blame for poor outcomes or
inadequate health system performance to flawed decisions made by individual consumers
(Iyengar, 1991). Conversely, a paradigm that acknowledges bounded rationality and requires
policymakers to assume some responsibility for improving choices carries with it the implication
4
that policymakers must share blame for health system failures. To induce policymakers to accept
this risk, policies based on bounded rationality must embody a suitable return in terms of
political credit, to balance out the increased potential of future blame.
A third constraint involves contemporary American doubts about the efficacy of collective
action. The notion that individuals are unable to make choices that promote their own well-being
seemingly opens wide the door for government intrusion. Such paternalistic policies seem
inconsistent with Americans’ general distaste for government in their lives. They appear
particular inflammatory in health care, where fears ham-handed government involvement have
derailed a number of past reform initiatives (Gordon, 2003; Skocpol, 1996). At the elite level,
bounded rationality seems to clash with the prevailing ideological consensus within the Beltway
that favors individual choice regarding health benefits over collective judgments about the
efficacy or appropriateness of medical care (Schlesinger, 2002).
These concerns about excessive government intrusion can be somewhat mollified if interventions
are designed to be consistent with notions of “libertarian paternalism.” Such policies could
beneficially alter the choices of boundedly rational consumers, while being effectively
transparent to decision-makers who are better informed. By not impairing the autonomy of this
second group, interventions of this sort reduce some of the ideological tensions that emerge from
bounded rationality. But they do not entirely eliminate them, particularly in the context of the
health care choices likely to be most consequential in American medicine. We explore here how
libertarian paternalism might work in American health policy and consider some responses to the
limitations of this approach in health care contexts.
Our goals for this paper are thus to make the case that health care choices are often inconsistent
with the assumptions of current market-oriented health policies because they emerge from low-
information decision heuristics, that conventional policies will therefore prove counterproductive
in their effects, and that an alternative approach to policy design that takes bounded rationality
into account could achieve more appropriate outcomes. To make this alternative more feasible,
we explore and propose responses to the challenges of delayed policymaker learning, blame
avoidance, and fears of excessive paternalism in American medicine.
5
In addressing these issues, it is important to acknowledge that there are certain aspects of
bounded rationality that we are not addressing. We consider here implications of bounded
rationality for the choices made by individuals as patients in medical care and consumers in the
health care system, as they evaluate and select among uncertain prospects (treatments, health
care professionals, health insurance plans). There are other important ways in which consumer
choices deviate from neoclassical assumptions – in terms discount rates on future streams of
costs and benefits (Caplin and Leahy, 2003), as well as through the influence of social norms on
individual choices (Jolls et al, 1998), But we set these aside to simplify our analysis.
Although there is also considerable evidence that similar heuristics and perceptual biases affect
the choices of health care professionals and policy elites (Frank, 2004; Slovic et al, 2002; Rabin,
1998; Redelmeier et al, 1995; Mondak, 1994; Baumann et al, 1991; Tversky and Kahneman,
1984), these comparisons are not pursued here. Similarly, there is ample evidence that cognitive
constraints affect the ways in which citizens decide which public policies they consider
legitimate (Lau and Redlawsk, 2001; Schlesinger and Lau, 2000). The heuristics they apply can
also entail biases: the risk-reducing policies that citizens most value may actually undermine
their true security by increasing fragmentation of public and private insurance programs,7 the
policy priorities the public favors may shift government spending away from some of the most
serious contemporary threats to health.8 All these other implications of bounded rationality merit
further exploration but must remain outside the scope of our current considerations.
6
PART I: EVIDENCE OF ANOMALIES IN HEALTH (CARE)-RELATED CHOICES
In this first section of the paper, we unpack the conventional policy approach pursued under the
rubric of “market-oriented” health care reform. Policies of this sort rely on three forms of
intervention, related to the choice set itself, information about the alternatives in that set, or the
inducements for consumers to actively engage in health-related choices. After illustrating these
interventions, we identify evidence on health-care-related choices that is inconsistent with the
assumptions on which these interventions depend. We then link these anomalies to findings from
the broader literature on heuristics and biases under bounded rationality.
Conventional Assumptions Underpinning Market-Oriented Health Care Reforms
Policies intended to make medical care more market-like take three forms. The first involve
interventions expanding consumers’ choice set by increasing the number of options from which
they can choose or diversifying the alternatives that are available. A second set of interventions
involves the provision of performance data, intended to make consumers’ choices more fully
informed. The third set of interventions is designed to increase consumer engagement in the
process of choice, either through financial inducements or the assignment of decision rights that
require consumer participation in health care decisions.
Expanding the Choice Set: This first strategy for intervention is well-grounded in neoclassical
economic models of consumer behavior (Suen, 1991), which suggest that when consumers are
uncertain about future preferences, needs, or personal circumstances that their well-being is
increased by having a larger and more diverse set of options from which to choose. In health care
settings, the rationale for these interventions presumes that offering people a larger array of
choices will allow them to better match the distinctive strengths of particular health plans or
health care providers to their personal preferences or medical needs.
Examples of health policies designed to increase consumers’ choices include: (1) state-level
managed care regulations designed to increase the number of physicians from which patients can
choose by requiring health plans to affiliate with “any willing provider” (Noble and Brennan,
1999), (2) the Balanced Budget Act of 1997, which expanded the number of private insurance
options available to Medicare beneficiaries as alternatives to the conventional fee-for-service
7
insurance program (Oberlander, 2000), and (3) the fore-mentioned implementation of Medicare
Part D, under which CMS has resolutely pursued efforts to maximize the number of insurance
plans covering prescription drugs.
Policies designed to diversify consumers’ choices in medical settings have been equally
common. At the state level, they date back to the 1970s, when states enacted “freedom of
choice” laws that required insurance companies to pay clinical psychologists and psychiatric
social workers for the outpatient care of patients with mental illness, payments that had
previously been restricted to psychiatrists (Lambert and McGuire, 1990). At the federal level,
diversification interventions are perhaps best illustrated by other recent reforms to Medicare
(including provisions in the Medicare Modernization Act) that were designed to expand the types
of private insurance alternatives to conventional Medicare coverage, giving beneficiaries the
options to enroll in private fee-for-service insurance, preferred provider organizations of local or
regional scope, and health maintenance organizations that specialize in the treatment of
chronically ill beneficiaries (Gold, 2006).
Providing Information to Enhance Consumer Decision-making: A second category of market-
oriented policies is predicated on the notion that choice has value only to the extent that it is
based on reasonably complete and accurate information. These interventions can also be divided
into two categories: (1) those intended to assist consumers evaluating their options prior to first
seeking medical care and (2) those that provide feedback on the ongoing performance of
consumers’ health care treatments, providers or insurers
The first set of policies includes regulatory policies associated with advertising, including FTC
decisions in the late 1970s to overturn the legality of health professional restrictions on
advertising (Schlesinger, 2002) and policy changes at the FDA in the early 1990s that promoted
direct-to-consumer advertising of pharmaceuticals (Wilkes et al, 2000). Other policies with the
same intent took the form of creating report cards measuring the performance of health plans,
hospitals, and nursing homes, an approach pursued by both CMS for the Medicare program and a
number of states (Werner and Asch, 2005).
8
A second category of information-based interventions are intended to alert consumers if their
current health care arrangements become problematic. This includes a variety of disclosure
requirements for health care providers (Sage, 1999). Some states require, for example, that health
plans make public the grievances that their enrollees file with them; some go further to require
that summaries of this information are directly distributed to all enrollees (Schlesinger et al,
2002). CMS using survey data from beneficiaries who have left their Medicare health plans to
construct profiles of the reasons for disenrollment.
Increasing Consumer Engagement in Choice: Economists have long been concerned that health
insurance, by muting the financial costs of medical care, creates a problem of moral hazard
among the insured, who have less incentive to limit their use of costly treatments or costly
providers (Nyman, 2005). To the extent that identifying lower cost treatments and providers is
itself a cognitively challenging task, these same incentives may cause health care consumers to
be less actively engaged in choice. At least so goes the reasoning behind consumer-driven health
plans, which incorporate large deductibles as financial inducements for greater consumer
engagement and medical savings accounts (MSAs) , which are intended to make consumers
more sensitive to the out-of-pocket costs of choosing a health insurer and thus become more
focused on this choice ((Rosenthal and Milstein, 2004). Although both reforms are more
common to date in employer-based insurance than public policy, Medicare has for the past few
years incorporated an MSA demonstration project. In addition, the Medicare Modernization Act
authorized CMS to develop a premium-support demonstration, which would pay beneficiaries a
fixed amount thereby making them more financial sensitive to their choice of insurance
arrangements under Medicare auspices, and thus purportedly more committed to carefully select
among alternative plans.
A second approach to enhancing consumer engagement in choice involves the assignment of
decision rights – requirements that choices reflect the active participation of patients or their
surrogates. Examples include the federal Patient Self-Determination Act that mandated that
patients near the end of life be offered the opportunity to sign advance directives. At the state
level, informed consent requirements for physician-patient decision-making have been adopted
by a number of states for particular treatments or diagnostic tests (e.g., AIDS screening).
9
Anomalies Evident in the Literature on Choices Involving Health and Health Care
All three sets of interventions have a commonsense appeal. Expanding choices, enhancing
information about those choices, and increasing consumers’ engagement with the choice process,
should, it would seem, logically lead to better decision-making and outcomes that better promote
consumers’ welfare. Nonetheless, there is considerable evidence suggesting otherwise.
Anomalies Associated with Expanding the Choice Set Expectation #1: Expanding Options Leads to Choices That Better Match Consumer Preferences In reality: As documented above, about half of the elders not enrolling in Medicare Part D explained
their reticence in terms of confusion about their choices; health care providers who assist elders in these choices generally concluded that there are too many options available for beneficiaries to sensibly assess
Similar patterns emerge for choice under Medicare Part C. Case studies comparing
communities with many private plan options (7-11 plans) to those in which there are fewer alternatives (2-3 plans) suggests that providing more choice is not associated with an increase in the percentage (7-9%) of new beneficiaries choosing a private plan (Gold et al, 2001). Nor do beneficiaries in “high-choice” communities report having better coverage, getting more value for their money, or being more willing to recommend their health plan to others.
Larger statistical analyses of enrollment in Medicare Part C (a.k.a. Medicare Advantage)
plans find little evidence that the number of plans available actually increases enrollment. One study found no statistically significant relationship between number of plan options and enrollment (Gold et al, 2003). A second study found mixed effects over time – but a significant negative relationship between the number of plan options and enrollment was most common (Laschober, 2005). A third study also documented a significant negative relationship between the number of plans available in the community and the propensity for Medicare enrollees to enroll in any of the plans (Mello et al, 2002).
Private insured enrollees in HMOs report lower levels of satisfaction with their plan when
they live in communities in which a large number of HMOs operate, compared to communities in which there are fewer such health plans (Scanlon et al., 2006)
Changing the order in which three attributes of health insurance policies are presented to
consumers leads to significant shifts in the valuation of those policies (Johnson et al, 1993). Ordering plans with respect to quality leads to greater choice of high quality plans, even when the number of choices (5 plans) is not that large (Vaiana and McGlynn, 2002). These findings suggest that consumers do not fully process the information even in relatively small choice sets.
10
Expectation #2: More Diverse Offerings Enhance Consumers’ Well-being In reality: Diversifying insurance choice by allowing narrowly defined policies increases subjective
satisfaction, but reduces consumer welfare. When offered two hospitalization policies (one of which covered diseases, the other accidents), subjects were willing to pay a combined monthly premium of $80 for the two policies (Johnson et al, 1993). Those offered a single policy covering both sets of risks (or a single policy covering an even more inclusive set of risks) were willing to pay about $40 dollars a month in premiums.
Roughly two-thirds consumers purchasing health insurance prefer policies that offer rebates
for low levels of claims to those that incorporate deductibles, even if the latter have a higher actuarial value (Johnson et al, 1993).
Consumers appear to overvalue the coverage of high probability, low-cost expenditures in
making choices among health insurance plans (Marquis and Holmer, 1996; Ellis, 1989; Slovic et al., 1977).
Anomalies Associated with the Provision of Additional Information: Expectation #3: Giving Consumers More Information Will Lead to More Informed Choices In reality: People will actively avoid information related to their own health needs. Studies indicate that
40-60% of patients do not want information about genetic susceptibility to various cancers (Lerman et al, 1999; 1996). If treatment is not effective, information avoidance increases to as high as 85% (Caplin and Leahy, 2003). 20% avoid testing information, even if problem is serious and treatable (Dawson et al, 2006).
Information avoidance is most pronounced among individuals who are most at personal risk.
Patients who have a family history of particular health problems are less likely to seek testing for that problem; those with more serious symptoms delay longer in seeking a diagnosis from a physician (Kőszegi. 2003; Vick and Scott, 1998)).
Although tailored information (explaining risks for the relevant class of patients, such as
women aged 45-50) increases screening rates, relative to standard clinical practice, more detailed personalized information about risks is associated with lower take-up rates for screening exams. These suppressing effects of detailed risk data are most pronounced among groups at highest risk (Edwards et al, 2005)
Providing consumers with information about health plan or hospital characteristics thought to
be connected to quality of care had a mixed effect: those who better understood these characteristics initially paid them greater attention after the information. Those who understood them less initially paid less attention after getting the additional information (Schlesinger et al, 2004).
11
Roughly 90 percent of new Medicare beneficiaries living in communities with multiple plan choices express confidence in their ability to choose among health plans (Gold et al, 2001), though few actually had sufficient information to make knowledgeable choices (Hibbard et al, 2001). Reported confidence is highest among 85+ than 65-84, although evidence suggests that they are least informed.
Roughly two-thirds of all Americans favor anecdotal accounts from family and friends over
statistical evidence and expert ratings in selecting doctors, hospitals, and health plans, to the extent that they will choose providers that are lower rated in performance rankings and report cards (Kaiser Family Foundation, 2000).
Expectation #4: Helping Patients Identify Problems Will Induce Search for Better Providers In reality: Published statistics on mortality rates for cardiac care did not alter patients’ choices among
hospitals, even among those people who saw this information (Frank, 2004). Although exposure to CAHPS ratings increased knowledge of differences in plan performance, it did not affect enrollee choices (Hibbard et al, 2002a)
Performance ratings of health plans had a higher value to consumers when they identified
bad performance than when they identified comparable improvements in performance, but bad ratings induced relatively little switching of plans (3%) and almost no change in the market share of health plans (<1%) (Chernew et al, 2001). Recall of information about poor performance tended to degrade more rapidly over time than did recall of good performance (Hibbard et al. 2005).
Problematic experiences with general practitioners lead patients to lower their evaluation of
service quality, but do not induce them to switch physicians (Ro and Wetzel, 1998). Patients who are anxious about the quality of their own medical care become worse decision-
makers. Efforts to educate people about health-related problems are much less effective (20% of subjects vs. 80%) among those who are most anxious about those problems (Lerman and Coyle, 1995). The more fearful people are about the future, the more they tend to inflate risks associated with uncertain events (Lerner and Keltner, 2001), making them act in ways that appear more risk averse (Sunstein, 2002)
Anomalies Associated with Efforts to Promote Consumer Engagement: Expectation #5: Paying Out of Pocket Will Lead Consumers to Make Wiser Purchases In reality: People are willing to pay less for the reduction of risks to which they are frequently exposed
in person than for risks that involve less frequent personal exposure (Sunstein, 2002).
12
Roughly two-thirds of purchasers of health insurance prefer policies that offer rebates for low levels of claims to those that incorporate deductibles, even if the latter have a higher actuarial value (Johnson et al, 1993).
Although increasing the out-of-pocket costs for medical care decreases medical care use, the
elasticity of demand is not consistently related to whether there are close clinical substitutes (Frank, 2004) or medical care is efficacious (Newhouse et al, 1990).
Consumers choose health insurance policies based on copayment rates, rather than the full
expected cost (expense of the procedure times the frequency of its occurrence times the copayment rate) of the treatment (Rizzo and Zeckhauser, 2004)
People are willing to pay approximately the same insurance premium to insure against low-
probability events that differ a hundred-fold in their prevalence (Kunreuther et al, 2001). Expectation #6: Assigning to Consumers the Right to Choose Will Engage Them More
Effectively in Health Care Decisions In reality: Although 65 percent of people surveyed say that if they were to get cancer, they would want
to choose their own treatment, in fact, among people who do get cancer, only 12 percent actually want to do so (Schwartz, 2004). Evaluations of the Patient Self-Determination Act have identified substantial changes in patient awareness about treatment, but few measurable changes in treatment itself.
Patients who explicitly agree to undergo unpleasant medical procedures have a tendency to
overstate their subjective benefits after the treatment, relative to measures of improved clinical functioning and patients who played a less active role in choice (Redelmeier et al, 2001b).
Patients’ choices, even about value-laden issues of considerable moral salience, are highly
sensitive to external influences In experimental contexts, when subjects are told that the law requires that they explicitly agree to organ donations, 42 percent do so. When they are told that the law assumes implied consent unless they indicate otherwise, 82 percent agree to donate (Johnson and Goldstein, 2003).
Patient choices are also powerfully affected simple differences in the depiction of their
options and likely outcomes. When outcomes of treatment are described in terms of mortality rates rather than survival rates (though the two are obviously equivalent in a statistical sense) yields a substantial (40-50% of patients) shifts in preferred choices (McNeil et al, 1982). Comparably large shifts are evident for choices involving health promoting behaviors and use of disease screening, again based on whether the outcomes are described in terms of gains (years of life added by inaction) or losses (life shortened from inaction) (Rothman and Salovey, 1997).
13
Bounded Rationality and Decision Heuristics as Explanations for These Anomalies All told, the accumulated evidence from the anomalies cited above come from about 40 different
studies, investigating a wide range of different choices related to health and health care. To be
sure, the findings from any one study are open to multiple interpretations. Not all these findings
necessarily impugn the value of consumer choice in medical settings; neither do they suggest that
informed choices are beyond the capacity of all health care consumers. But considered in
aggregate, they do raise serious questions about conventional notions of how consumer choice
can be improved by market-oriented health policies.
In describing the findings above as anomalies, we emphasize their inconsistency with the
assumptions that undergird conventional health policy. That does not mean that they are
inexplicable. Quite the contrary – those who are familiar with the literature on bounded
rationality and decision heuristics will immediately recognize in these findings a variety of
familiar patterns. These can be linked to more general propositions about how people make
choices involving complex alternative and uncertain prospects, without having the time or
substantive expertise to accurately anticipate future outcomes.
Although it is beyond the scope of this paper to fully identify all of these conceptual linkages, it
may be useful to highlight a handful of the key psychological underpinnings that can help to
explain some of the anomalous findings cited above. We briefly describe here three such
foundational influences and trace some of their likely implications for health(care)-related
choices.9
Emotional Homeostasis and Filtered Perceptions: Faced with an uncertain and potentially
stressful reality, people develop perceptual filters that enhance their ability to cope with their
own bounded rationality. The literature suggests that these filters serve three distinct purposes.
The first involves the construction of causal relationships that reduce the apparent randomness of
life. When life seems filled with inexplicable events, people feel a need to maintain a constant
state of heightened alertness. As McFadden wrote, some 30 years ago, “Chance jolts the
harmony of conscious beliefs; relief from this dissonance is gained by imposing an order over
chance, a fabric of cause and effect, out of the jumbled coincidences of random events.” (1974,
14
p.127). Given this motivation, people may be too quick to see regularities and purpose in
randomness. This may lead them attribute undue influence to other actors, to understate
regression to the mean, and embrace what Kahneman and Tversky labeled the “law of small
numbers” – the belief that a small sample accurately represents the probability of an uncertain
event (Rabin, 2002).10
A second form of filtering inflates perceptions of self-efficacy, by ignoring evidence suggesting
that one’s past decisions were flawed (Soll, 1996; Taylor and Brown, 1988). This becomes
evident in: (1) a bias toward optimism about one’s own life course, discounting risks that are
known to apply to the general population (Armor and Taylor, 2002; Weinstein and Klein, 1995),
(2) selectively revising one’s prior expectations, so that they are retrospectively made more
consistent with realized outcomes (Hawkins and Hastie, 1990; Conway and Ross, 1984), (3)
interpreting ambiguous evidence to make it more consistent with one’s prior expectations (a.k.a.
“hypothesis-driven filtering”) (Rabin and Schrag, 1999), and (4) avoiding difficult choices that
one might later regret (Frank, 2004; Nicol-Mauveyraud, 2003; Anderson, 2003).
A third form of filtering, famed as the refuge perfected by the ostrich, involves reducing
awareness of external risks by simply ignoring frightening information (Witte and Allen, 2000),
often termed “information aversion.” It is a frequent response to uncertain situations,
particularly for those who lack the time, resources or self-efficacy to directly manage the risks
that they face. Research dating back to the 1950s documents anxiety-induced avoidance of
information of various types (Loewenstein et al, 2001).
These forms of perceptual filtering have different degrees of relevance to the anomalies
documented above. An inflated sense of causal order does not play a central role in these
anomalies, though is relevant to a phenomenon that proves important when we later consider
how policymakers will respond to evidence of bounded rationality: the tendency of most
Americans to overestimate the skills of their own health care providers. Since most people visit
the doctor only when they feel sick, odds are that they’re likely to feel better the next day or two,
even if the visit yielded no clinical benefits. Patients who underestimate regression to mean will
over-attribute these improvements to their clinician’s skills or healing touch
15
By contrast, coping mechanisms that enhance consumers’ ability to deal with uncertainty appear
culpable in a number of health (care) anomalies. (1) Studies suggest, for example, that
assessments of health threats undervalue those risks that are viewed as ”controllable.” This may
reflect an over-inflated belief in one’s personal ability to control uncertain events. (2) Selective
editing of prior expectations may lead consumers to be overconfident in their ability to choose
physicians, hospitals, and health plans that fit their personal needs and preferences. Discounting
or simply “forgetting” preferences that are ill-served by current providers is a surefire way to feel
confident about one’s ability to pick the best provider in the first place. (3) Regret avoidance
may account for people abstaining on choices among health plans, providers, or treatment
options, even if they have adequate information to make those decisions. This may lead to high
rates of default assignments, even for people who have well-defined preferences among their
choice options. And studies suggest that avoidance will be even more pronounced if people
anticipate subsequent feedback from report cards that might remind them of their ill-conceived
choices (Larrick and Boles, 1995).
Finally, information avoidance to reduce anxiety accounts for a number of the perverse patterns
of behavior documented above. Because information about risk (health risks, risks of medical
error from having chosen a bad doctor or hospital) induces more anxiety for the people who face
the highest risks, anxiety reduction explains why those who most need to make informed choices
are also the most ready to ignore information relevant to those choices (Hale and Dillard, 1995).
By similar logic, the more accurate is the information in clarifying the probability of a bad
outcome, the more aversive it may become to those likely to be affected by that outcome (Caplin
and Eliaz, 2003; Lieberman and Chaiken, 1992). Vague information about risks, because it
induces less anxiety, may actually prove more informative than accurate information, tailored to
an individual consumer’s risk profile.
Limited Capacity for Information Processing: The world is a complex place, relative to our
ability to make sense of it. Faced with this complexity, people develop a set of decision-
heuristics that allow them to make choices based on incomplete, but workably inclusive,
information sets (Conlisk, 1996). These can take a variety of different forms:
16
The first involves partitioning the choice sets into narrower categories, so that decision-makers
are comparing among fewer alternatives, a process labeled “mental accounting” (Thaler, 1999).
This partitioning of trade-offs can lead to inconsistent behavior: For example, for most people
the rate of discounting embedded in short-term decisions is far larger than that evident in choices
made about longer-term considerations. Second, decision-makers may substitute affect driven
heuristics for cognitive assessments, such as selecting service providers because they feel a sense
of personal connection with them (Shelley, 2001). Voters often choose candidates because they
seem “likable” (Kahneman and Frederick, 2002), they avoid choice settings that induce high
levels of foreboding or anxiety (Slovic et al, 2002).
Under some circumstances, these coping mechanisms are insufficient, so that people prefer to
defer choice, even if the status quo is seemingly less acceptable than any of the choice options
(Dhar and Nowlis, 2004; Dhar, 1997; Tversky and Shafir, 1992; Akerlof, 1991). Experimental
findings suggest that putting off choice is a function of the number of choices (Iyengar and
Lepper, 2000), the potential for regret in the aftermath of a poor choice (Anderson, 2003), the
risk of blame for making a poor choice (Anderson, 2003) and the relative attractiveness of those
choices to one another (close calls induce choice deferral) (Dhar et al, 1999; Dhar, 1997).
Each of these responses to cognitive constraints has potentially important implications for health
care choices. Partitioning may affect patients interpret their health care experiences. Problematic
outcomes may have a variety of causes (Rosenthal and Schlesinger, 2002). Patients appear to
simplify these attributions by partitioning problems into distinctive categories: health plans are
held responsible for problems related to payment and paperwork, physicians for failures of
treatment (Goold and Klipp, 2002; Schlesinger et al, 2002). This partitioning may account for the
limited impact that information on quality has on shifting patients away from low-quality health
plans – they simply don’t see quality of medical care as falling into the plan’s “account”.
Affect plays an important role in shaping patient’s assessment of health care professionals.
Studies suggest that patient rely heavily on a sense of emotional connection with their doctors to
determine whether physicians are trustworthy (Mechanic and Meyer, 2000; Goold, 1998). In this
way, affective heuristics serve as proxies for other aspects of medical professionalism that
17
patients recognize as important for medical care, but which they cannot assess without a
burdensome investment in information collection and deliberation (Slovic et al, 2002).
Choice deferral may explain why increasing the number of health plans in a community does not
induce more consumers to enroll or greater satisfaction with the plans they select. All else equal,
having more plans available must allow better matching for some consumers – a plan that’s
located closer to home, or one that includes the physician with whom they or their family has a
long-standing relationship. But the average satisfaction with health plans in communities with
many plans is no higher (and some studies suggest is lower) than those with few plans. There
must therefore be an offsetting reduction in satisfaction experienced by some consumers when
confronted with large sets of alternatives. Perhaps some become so overwhelmed by their
options that they fail to switch out of health plans that are performing poorly (Schlesinger et al,
1999). Others may defer choice, because the additional plans induce greater potential for regret
at having made a poor choice.
The Evaluation of Uncertain Prospects: First articulated by Kahneman and Tversky in the 1970s
(Tversky and Kahneman, 1974; 1979), the implications of Prospect Theory have been extended
and refined over the subsequent 25 years (Camerer and Lowenstein, 2004; Kahneman and
Frederick, 2002; McFadden, 1999). But its central premises remain unchanged: (a) that expected
utility is assessed relative to some reference point (typically the individual’s current
endowment), (b) that gains and losses, relative to this reference point, have asymmetric affects,
with losses having larger consequences for expected utility than do equal- sized gains, and (c)
that people are risk averse in gains, but risk loving in losses.11 Two sets of implications are most
relevant to health care related choices.
The first involves the framing of choices involving uncertain prospects. Because people’s risk
preferences differ for losses and gains, choices among courses of action can be altered by
shifting their characterization from a “loss frame” to a “gain frame”, or vice-versa: more people
prefer risky over certain options when the outcomes are framed as losses, but certain over risky
outcomes when frames as gains.12 Framing effects are evident (a) in individual behavior, as well
as choices among states of the world, (b) for frames in which the alternative options are only
18
partially described, and (c) for frames that subjects held prior to the experiment, as well as those
imposed in the experiment itself (Rothman and Salovey, 1997).
The asymmetry of the expected utility function has a second important implication: people will
pay more to avoid losing something than they would to acquire it. This has been termed the
“endowment effect” and has been documented in a wide range of choice settings (McFadden,
1999). This same asymmetry induces a sort of “status quo bias,” which can effectively lock
decision-makers into their current circumstances (Zeckhauser and Samuelson. 1988).
Framing effects can account for some of the reasons that patient decisions appear so sensitive to
the context in which they make their choices. Framing has been shown to affect take-up rates for
mammography, AIDS testing, and various protective health practices (Apanovich et al, 2003;
Finney and Iannoti, 2002). The magnitude of the framing effect is typically mediated by several
additional factors, including the perceived efficacy of the intervention (framing effects appear
substantial only if people cognitively engage with the health promoting message, see Block and
Keller, 1995) and the prevailing frame for the behavior itself, that has been defined by the culture
and media representations (Rothman and Salovey, 1997).
Status quo bias may be evident in patient responses to unsatisfactory experiences with their
medical care. 13 Although problematic events are quite frequent, relatively few people respond by
either switching doctors or health plans (Schlesinger et al, 2002). Of course, some of this inertia
reflects a rational calculation of the costs and benefits of exit. Yet it is striking that when
members of the public are given a choice among health plans, only a third of those who had
experienced severe problems with their plan actually switched during the next open-enrollment
period (Schlesinger et al, 1999). Since the costs of switching in these settings were low, a
devaluing of the alternative plans may better explain why switching was so inhibited.
19
PART II SOME CHALLENGES IN APPLYING BOUNDED RATIONALITY TO HEALTH-RELATED
CHOICES AND POLICIES
We have identified anomalies in the ways in which people learn, evaluate, and make choices
about health-related matters, compared to the assumptions of conventional health policy. We also
identified heuristics and biases from the literature on bounded rationality that could account for
these seemingly perverse outcomes. This might seem a sufficient rationale for policy change. But
evidence consistent with these concerns is already evident from the early implementation of from
Medicare Part D, yet seems to have induced little reassessment of the conventional market-based
paradigms of health care reform. More choice, more information, and more consumer
responsibility continue to be seen as prerequisites for improving health system performance.
The Challenge of Inducing Paradigm Shifts Among Policy Makers
Students of the policymaking process can understand why this might be the case (Lindblom,
1990). Policymakers typically have little time to dwell on the nuances of policy deliberation.
Though health policy is typically one of the most important concerns on the domestic policy
agenda, it is also one of the most complex (Heinz et al, 1993). Faced with this complexity and
the constant emergence of new health-related concerns, most policymakers can devote relatively
little time to assessing reform strategies (Jones, 1994). They must apply their own set of
heuristics for simplifying the choices that confront them, often relying on historical precedent to
provide them with guidance (Houghton, 1998; Khong, 1992).
Under these circumstances, policymaker’s conceptual frameworks don’t have the same dynamic
of paradigm shifts that Kuhn (1962) identified for scholarly endeavors. Policymakers generally
don’t aspire to the sort of predictive reliability that motivates academic endeavors. Consequently,
an accumulation of anomalous bits of evidence is unlikely in itself to motivate policy change,
since the operative goal of policymaking is not descriptive accuracy, but marginal improvement
in health system performance. From this perspective, it’s easy for policymakers to argue that
even if most Americans are unable to make sense of report cards or take advantage of increased
health-related choices, some will benefit and that is a sufficient justification. Particularly if those
who care enough to pay close attention to their choices are, for that very reason, seen as most
deserving of better health care outcomes (Tomes, 2006; Schlesinger, 2002).
20
Consequently, a higher standard of evidence is required to induce a shift in policymakers’
conceptual framework – one that suggests that the current paradigm will yield outcomes that are
overtly counterproductive, thus leaving policymakers vulnerable to being seen as incompetent or
hypocritical (promising one outcome, but promoting another quite different). The neglect of
bounded rationality can arguably produce exactly these sorts of counterproductive outcomes, by
distorting policymakers’ sense of what their constituents want, or by inducing policy outcomes
that are the opposite of those intended when the policy was enacted.
Misreading Public Preferences
Precisely because health care matters have considerable political salience, policymakers try to
monitor constituents’ preferences in this arena (Jacobs and Shapiro, 2000). But a neglect of
bounded rationality could distort the ways in which policymakers assess Americans’ concerns
about health care.
Choice and Satisfaction: It has become a truism in the health policy literature that some
choice of physician (Kalda et al., 2003; Krupat et al., 2002) or health plan (Schone and
Cooper, 2001; Gawande et al, 1998) yields better health system performance. This is based
on statistical results that show a positive association between having had such a choice and
subsequent patient satisfaction. But the literature on bounded rationality suggests that this
relationship may be entirely artifactual – that giving people choice causes them to assess
their physician or health plan in a more positive manner, whatever their actual performance,
simply because having made a choice creates a conceptual filter that favors the perception
and remembering of positive outcomes that are consistent with that choice. Put differently,
giving constituents the choice that they want may in turn distort their subsequent perceptions
of the care they receive, undermining the utility of consumer voice as a marker for health
care quality and an inducement for improved performance.
21
Revealed Preferences and the Importance of Quality: There is extensive evidence that
choices among health plans are not particularly responsive to measures of quality on report
cards. This is often interpreted as demonstrating that quality simply doesn’t matter much to
Americans. But this is inconsistent with the ways in which consumers describe their
preferences, which places quality at the forefront of their considerations. This apparent
paradox can be explained by what we termed earlier the “primacy of the concrete”. Quality
may not appear to matter, because statistical measures of quality miss the emotional content
– the “caring” in health care -- that consumers use as a heuristic for assessing the quality of
medical care. Policymakers may thereby reach an entirely inaccurate impression of consumer
preferences – the factor that matters most to people appears by revealed preferences not to
matter much at all.
Potentially Counterproductive Policy Design
The findings above also suggest a number of ways in which public policies, sensibly designed
from the standpoint of conventional intuition, can actually produce counterproductive outcomes.
More Information, But Less-Informed Consumers: In a health system predicated on effective
consumer participation and choice, providing the public with more accurate and more
complete information would seem an obvious step forward. But the research cited above
suggests that when this information is threatening, or when it exceeds the rather limited
capacity of most consumers to process multi-dimensional attributes, information campaigns
may actually leave consumers less informed. In the first instance, they may actively avoid
possibly aversive knowledge to reduce their own fears. In the second case, cognitive
overload may lead consumers to give up trying to become better informed. Studies of
information aversion suggests that these perverse effects can be quite substantial: even for
vital health problems, with remediable consequences, 40-60% of the public simply prefers
not to know.
More Choices, But Less Choosing: When consumer choice is a centerpiece of health policy,
offering more choices would seem to allow for better matching to consumers’ particular
needs and preferences. Yet evidence from the drug discount cards in Medicare Part D, health
plan options under Medicare Part C suggests that introducing more options may overwhelm
22
consumers’ ability to differentiate among plans. Consequently, enrollees in health plans
would become less willing to leave plans that were performing poorly, undermining the
central rationale for introducing greater choice in the first place.
More Coherent Policies, But A Less Coherent System: Consumers are willing to pay more
for insurance that covers narrow and well-defined risks over policies that cover broader sets
of risks (Johnson et al, 1993). Although the former offer less protection, narrower outcomes
are easier to recall, comprehend and value.14 In private insurance markets, this suggests that
multiple narrow policies will be preferred to single broad policies. This was evident in the
market for Medicare supplementation (Medi-gap) policies; beneficiaries would often
purchase multiple policies with overlapping coverage, because the incremental benefits
offered by each policy seemed so salient and valuable. Under Medicare Part D, this same
bias may encourage consumers to select free-standing drug plans over more comprehensive
Medicare Advantage (Part C) plans that include prescription drugs in their coverage, even if
the latter offer better coverage or better coordination between prescriptions and other types of
care. It is striking in this regard that 90 percent of all the new enrollment in private plans
once Part D became operational involved free-standing drug plans (Kaiser Family
Foundation, 2006a).
Greater Consumer Responsibility, Less Consumer Empowerment: Some reformers favor
increasing the financial incentives affecting consumer’s use of medical care. But studies
suggest that most consumers cannot form accurate predictions about expected costs, make
systematic errors in their assessments of the value of insurance coverage (Johnson et al,
1993), and, if faced with the potential for selecting options that leave them exposed to large
financial risks, will avoid making choices they might later regret, effectively locking them
into their existing plans (Anderson, 2003).
23
The Promise of Improved Performance
By illustrating the crucial role of framing, emotional filters, information avoidance and errors in
risk perceptions, the litany of anomalies reported above highlights opportunities for interventions
that can improve health outcomes and health choices. In the second part of this paper we identify
particular heuristics that link to each of these anomalies. To foreshadow, small scale experiments
and demonstrations suggest that these effects may be quite dramatic, including:
More Effective Health Promotion: Small-scale experiment suggest that effectively matching
the framing of interventions to the types of health-behaviors they are attempting to alter
(loss-framed messages for disease detection; gain-framed messages for health promotion)
can enhance the performance of health promoting interventions by about 50%.
Reduced Information Aversion: Other experiments suggest that increasing the perception that
health problems can be successfully treated can reduce the extent of information aversion by
between 20 and 50 percentage points, depending on the source of the information.
More Effective Choice: Other studies suggest that minor manipulations in the presentation of
information (e.g. ordering plans/providers from high to low performance), the dimensions in
which performance is arrayed, the characterization of choices set, or practice with making
decisions can enhance the consistency between choices and preferences as much as 30% and
increase the take-up of new options by more than 50%.
Truly Patient-Centered Care: With greater attention to effective communication about risks
between physicians and patients, plan designs that can be understood by their enrollees, and
anxiety-inducing information, can allow patients to more effectively engage in health-related
choices. Evidence from studies of clinical settings suggests that choice deferral is on the
order of 40-50% ; how much of this could be offset by revising policies remains to be
established.
24
Taken together, these observations suggest ways in which one could build the case that
policymakers no longer have the luxury of ignoring bounded rationality in health care settings To
ignore the ways in which people evaluate uncertain prospects and assess risks can induce
misreading of public preferences, create counterproductive outcomes or produce policies far less
effective in their performance. But recasting health policy to take bounded rationality into
account requires taking on two additional concerns: American’s strong distrust of a paternalistic
role for government in health care, and policymakers strong incentive to favor policies that
diffuse political blame for poor health system performance.
Addressing Concerns About Government Paternalism
In contemporary political terms, the purported virtues of more market-like medical care are an
easy sell, since they seem compatible with both prevailing elite ideologies and deep-rooted
American values favoring individual choice and responsibility (Schlesinger et al, 2002). By
contrast, notions of bounded rationality suggest that health care consumers may be so
incompetent that they must be protected from their own bad choices. This is a scary prospect to
most Americans.
Precisely because of the psychological factors that we identified at the end of Part I of the paper
(most powerfully, the need to perceive oneself as efficacious and thus to ignore evidence to the
contrary), Americans hold peculiarly inconsistent perceptions about medical care. Most
Americans have a distinctively negative perception of the medical profession’s altrusim,
competence, communication skills, or commitment to their patients’ interests. But when it comes
to the personal physician whom they have selected, their perceptions are quite positive in all
these aspects of care (Jacobs and Shapiro, 1994). The same inconsistency is evident in the
perceived risk of quality problems for hospitals (Hays and Ware, 1986). As a result, when
Americans contemplate the impact of government intervention in the health care arena, they feel
as if they have a lot personally to lose – given a great doctor and hospital now, things can only
go downhill if government starts tinkering with their medical care.
25
So politicians concerned with health policy seem caught between a rock and a hard place. They
can continue to extol the virtues of consumer empowerment and individual choice, earning
symbolic kudos but knowing that the choices that emerge will neither protect the welfare of most
consumers nor advance broader societal aspirations for health system performance.
Alternatively, they can acknowledge bounded rationality and raise questions about consumer
choice, thereby seeming to threaten their constituents’ beloved medical arrangements.
Recent scholarship, however, suggests that there may be a feasible middle ground between these
two pessimistic extremes. Under the rubric of “libertarian paternalism” or “asymmetric
paternalism”, this work identifies interventions designed to respond to bounded rationality
without unduly constraining the options available to better-informed decision-makers. We first
describe these approaches in generic terms, then apply them to health (care)-related policies.
The Building Blocks of Libertarian (Asymmetric) Paternalism
Scholars have identified a half-dozen strategies under this rubric. Each involves a family of
interventions; the relevance of particular approaches will vary depending on whether policy is
designed to promote healthy lifestyles or disease screening, encourage selecting the best clinical
therapies, or improve choices among health care providers or health plans. We list the strategies
in order of their potential impact on individuals’ autonomy – those toward the end of the list
remain most susceptible to charges that they are more paternalistic than libertarian.
(1) Inoculating Against Heuristic Biases: For the most part, warning people about heuristic
biases does not appear to diminish their impact (Weinstein and Klein, 1995). Even when
explicit errors of judgment are identified, most people refuse to alter their initial choices.
There are, however, a few findings that suggest ways in which biases might be reduced. First,
the impact of gain-loss framing appears to be smaller when uncertain events are described in
terms of their natural rates rather than as probabilities (Gigerenzer and Edwards, 2003).
Second, framing effects are smaller for individuals who exhibit a more patient style of
decision-making (Frederick, 2005). Consequently, interventions that encourage people to
take more time in their evaluation of uncertain prospects appear to reduce biases in perceived
probabilities (El-Gamal and Grether, 1995; Grether, 1992). Third, presenting choices
described in terms of both losses (lives lost) and gains (lives saved) yields preference
orderings that fall in between those associated with either gain or loss framing alone.
26
Whether these mixed frame choices are more consistent with “true” consumer preferences
has yet to be determined.
(2) Altering Order: A second strategy changes the order in which people are presented with
information about uncertain prospects (Jolls et al, 1998) or the choices themselves (Thaler
and Sunstein, 2003). For example, information that tends to be undervalued (involving risks
that are seen as within an individual’s control, for example) might be placed at the beginning
of the description of choices. Cafeteria lines might be rearranged so that people pass by
choices of fruit before they get to the dessert options. In either case, the choices of well-
informed decision-makers (who read descriptions in full or look over the food options before
getting into line) are unlikely to be affected by these changes in order, whereas those who
make choices more impulsively or can process less information will arguably be influenced
in favor of choices that are socially beneficial.
(3) Changing Default Conditions: Given the many heuristics that contribute to status quo bias,
setting default conditions in favor of action rather than inaction seem likely to enhance social
welfare (Camerer et al, 2003; Thaler and Sunstein, 2003). For example, rather than requiring
workers to indicate that they want to have a portion of their earnings placed into a retirement
account, these deductions could be made unless the worker indicates that he or she prefers
otherwise. Shifts in defaults have been shown to significantly increase both participation in
retirement accounts and overall savings (Beshears et al, 2006; Madrian and Shea, 2001).
With this strategy, costs are clearly imposed on those whose preferences differ from the
default condition. Costs to override the default can be minimized.
(4) Cooling-Off Periods: Because strong emotions can alter perceptions of risk, but often to fade
over time, a fourth set of strategies mandates delays before decisions can be made (e.g.
waiting periods between issuing a marriage license and the ceremony itself) or before the
decision becomes “final” (e.g. trial periods during which a product can be returned, free of
charge) (Camerer et al, 2003). Delays of this sort impose some costs by reducing the joy of
impulsive action, but arguably short delays (a few days or a week) are a relatively small cost
for more consequential decisions that may affect people’s lives for years to come.
27
(5) Disaggregating Decisions: A fifth strategy focuses on collective decisions that have
important societal consequences – such as deliberations in legislatures or juries (Jolls et al.,
1998). Given evidence that once decisions are made, they induce perceptual filters that
distort subsequent evaluations of the merits of the choice in question, it can be argued that
different decision-makers should be brought into the evaluation process. From this
perspective, for example, juries that determine guilt should not decide on sentences; juries
that determine negligence should not decide on the monetary value of the compensatory
payment. Although typically justified in other ways, term-limits could be embraced for
similar reasons, on the grounds that the legislators who enacted a program cannot offer
unbiased judgments about whether such a program should be re-authorized at a later date.
(6) Limiting Choices: The final strategy produces the most skepticism because it impinges most
on individuals’ autonomy. Proposals of this sort include setting deadlines for decisions to
participate in public programs (Camerer et al. 2003).15 These approaches take away the
option of watchful waiting and prolonged learning, but otherwise does not restrict valued
alternatives. A second version involves the elimination of dominated options – that is, those
that perform less well than some other member of the choice set on all measurable
dimensions of performance, The costs are higher in this case -- although economic theory
suggests that these dominated options ought to be irrelevant to choice, not only do some
people select these alternatives, but they are also used by others to help them decide which
option to select (Dahr, 1997). Most controversial are policies that would foreclose options
that have low probability pay-offs, but high emotional valence (Sunstein, 2002). Examples
might include playing the lottery or betting on the Cubs to win the World Series.
28
Applications to Health (Care) Policy
One can readily envision examples of each of these strategies in the health policy domain. Some
can already be found in practice, though others remain only hypothetical.
Inoculating Against Heuristic Biases: Patients who are presented information about clinical
options in pictorial form or natural rates are less affected by gain-loss frames than those
whose treatment options are described in terms of probabilities (Gigerenzer and Edwards,
2003). But the large emotional valences that inhere in health-related choices make framing
effects and other emotion-driven heuristics more powerful than in other policy domains and
probably make it infeasible to largely ameliorate decision biases in this manner.
Ordering Effects: As noted earlier in this paper, minor alterations in the presentation of
information have real consequences for the ability of consumers to choose among insurance
options in a coherent manner. Ordering plans with respect to quality (as opposed to, say,
alphabetically) leads to greater choice of high quality plans, even when the number of
choices (5 plans) is not that large (Vaiana and McGlynn, 2002). Changing the order in which
the attributes of health insurance policies are presented leads to significant shifts in the
valuation of those policies (Johnson et al, 1993). Ordering of options within each dimension
of performance is straightforward. The challenge in this case involves the complexity of
many health care related choices. When options differ in a number of different dimensions,
the appropriate ordering of information becomes harder to establish. And since ordering will
affect the relative market share of the alternatives, the economic stakes may be so high that it
becomes impossible to implement the most effective ordering strategies.
Changing Default Conditions: Organ donation offers a compelling example. European
countries that require explicit consent for donation average around a 15 percent effective
consent rate; those that rely on presumed consent average above 90 percent (Johnson and
Goldstein, 2003). Of course, these differences may also reflect cultural values that make
certain consent practices seem more legitimate – though the cultural distinctions between the
Netherlands and Belgium (consents rates of 28 and 96 percent respectively), Germany and
Austria (12 vs. 99.9 percent) or Denmark and Sweden (4 vs. 86 percent) are not obvious.
Experiments in the U.S. have found that 42 percent of Americans will agree to become
29
donors if required to explicitly opt-in, compared to 82 percent if required to opt-out.
(Preferences elicited in the absence of default conditions indicate that 79 percent wanted to
be donors). Default conditions might be similarly altered for participation in employer-
sponsored insurance, public programs (Medicare Part D has assumed consent for low-income
beneficiaries, but not for others), follow-up medical exams or health screening initiatives
(imagine a postcard in the mail stating: “your colonoscopy has been scheduled for …..”).
Cooling Off Periods: Because many health-related choices can be so emotionally charged,
one can imagine protocols that build-in time buffers for making choices related to treatment.
In this case, however, its not clear that the emotional valences diminish over time. Consider,
for instance, patients diagnosed with prostate cancer. Based on the logic of this strategy, one
could delay asking patients about their preferences for treatment. Arguably, however, the
delays in this interim period could as likely exacerbate patients’ anxiety, so that they would
be making delayed decisions in a state of heightened rather than diminished anxiety.
A variant of this approach has been found to be more successful. For medical procedures that
generate a certain level of discomfort (colonoscopies, for example) patients report lower
levels of pain if the procedure is deliberately prolonged so that peak levels of discomfort
occur in the middle of the episode rather than the end (Gilbert et al. 2002). Similar effects
might follow if consumers were required to stay with their doctor or health plan for a certain
period of time following a problematic episode, rather than encouraged to switch to a
preferred alternative.
Disaggregating Decisions: Because choices in health care entail such complex contingencies
(the health plan one selects establishes the choices of primary physicians, since not all
doctors in the community are affiliated with every health plan; the primary care doctor one
selects determines the specialists to which one will be referred; the specialists to which one is
referred determine the hospital in which severe illnesses are treated), a disaggregation
strategy does not seem to have great relevance in health care settings
30
Limiting Choice Sets: A number of health policies already incorporate decision deadlines.
For example, Medicare Part D penalizes beneficiaries who do not select a plan during the
initial enrollment period. Limited periods for open-enrollment have been argued to reduce
biases associated with procrastination, since people must make their choices during the
enrollment “window” (a deadline) or wait a full year before they can do so again (Camerer et
al, 2003). Its not clear that there is great potential for additional reform along these lines.
The opportunities for limiting choice sets depends upon the nature of the health services
involved. For standardized products like prescription drugs, its easy to imagine a variety of
choice constraints being applied (one generic drug vs. many; a small number of mail-order
pharmacies). Medicare Part D, in fact, seems a prime target for such simplification, precisely
because the quality of the drugs themselves or the accuracy of a patients’ prescription are
unrelated to the health insurer that covers the costs. For other aspects of medical care,
however, there are so many dimensions in which outcomes may be evaluated in highly
idiosyncratic ways (how close are the nearest health plan offices to my home? does the
physician’s interaction style remind me of my kindly grandmother?) that it would be difficult
to single out dominated alternatives.
The Challenge Posed By Blame Avoidance
Whatever the promise of libertarian paternalism as a general response to bounded rationality, its
scope in health care settings seems sufficiently limited that interventions intended to be
responsive by low-information decision heuristics will remain contested and controversial. Take
a case in point from the examples we just considered. There is a strong case for setting a default
of implied consent for organ donation, in the sense that (a) the availability of organs has
important externalities for the quality of American medicine, (b) the supply of organs appears to
be substantially increased by implied consent laws, and (c) the resulting proportion of donors is
closely congruent with hypothetical intentions to donate, among Americans who are not primed
by knowledge of what the legal defaults currently are.
31
Nonetheless, one can imagine the potential repercussions the first (or tenth, or hundredth) time
that organs are harvested from someone recently deceased over their family’s intense objections.
Because these objections often stem from moral or religious precepts, past research suggests that
the public is likely to see the outcomes as particularly blameworthy for the politicians who
enacted the law (McGraw, 1991). Of course, these same studies suggest that if the politicians can
invoke notions of the greater public good, they can diffuse much of this blame. But whatever its
impact on this particular policy choice, the general lesson seems clear. The extent to which
policy responses to bounded rationality are feasible will depend on the political consequences to
the policymakers who endorse them.
The tensions inherent in these prospects are perhaps most pronounced for policies that purport to
improve the welfare of health care consumers, but at the expense of their self-assessed well-
being. Interventions of this sort might include:
prohibiting health insurance policies that cover a narrow range of diseases of treatment,
despite the high salience of these risks as perceived by consumers and the high value that
they therefore place on this coverage,
presenting Americans with information that can more accurately inform them about the risks
they face from particularly treatments or incompetent health care providers, despite the fact
that this moderately increases their anxiety by undermining their faith in medical care and
their own doctors,
restricting the forms of other information that are available by prohibiting or limiting direct-
to-consumer advertising by health care providers, despite the fact that this sort of marketing
does make consumers better informed about their treatment alternatives,
limiting the number of health plans that can operate in each community,16 despite the fact
that this would foreclose some options that would otherwise have allowed some consumers
to better match their health needs with the particular strengths of a given health insurer.
Under conditions of bounded rationality in medical markets, each of these interventions could be
plausibly argued to leave Americans better off and improve health system performance.
Nonetheless, it is difficult to imagine circumstances under which contemporary politicians would
willingly endure their constituents’ antagonism that would surely result from these policies. If
32
accurate, this calculus of blame avoidance suggests that reforms addressing the implications of
bounded rationality may require institutional changes that could create a political buffer between
elected officials and the difficult choices required to improve American medicine. One could
imagine an independent agency charge with monitoring and improving health system
performance, given the same mandate and general scope of authority as the Federal Reserve has
for the national economy. The governance structure could take a similar form as well, ensuring
some political accountable while maintaining considerable independence for the leaders of this
new agency.
DISCUSSION AND CONCLUSIONS
The limitations of conventional analysis are evident in current policies that attempt to empower
medical consumers, expand the scope of their choices, or encourage them to make better-
informed decisions. The gap between policy aspiration and realization appears large and the risks
of policies producing perverse outcomes quite substantial. Many of these shortcoming and
anomalies can be traced to the reliance of policymakers on neoclassical models of consumer
behavior. We have suggested here that a more promising alternative (or, more accurately, a
complement) exists. Although our current understanding of bounded rationality remains
incomplete and its conceptual underpinnings less coherent than neoclassical economic theory, it
nonetheless holds the promise of more efficacious policy design and implementation.
To realize this promise, it will be necessary to augment and extend existing research in several
ways. Three sorts of intellectual investments will be required. In this concluding section, we
sketch out some broad parameters for each, leaving more detailed development and priority
setting for a later time.
Enriched Understanding of the Intersection Between Cognitive and Emotional Engagement
The first involves some theoretical integration of several streams of scholarship that remain
relatively independent in the heuristics and biases literature: the intersection of emotion and
cognition in low-information decision heuristics. There have been some important recent steps
forward in the form of various flavors of “dual-processing” models that combine affect and
cognition (Loewenstein, 2004; Sloman, 2002; Slovic et al, 2002; Lowenstein et al, 2001). To
33
date, however, there has been more proliferation of model types than testing of their comparative
explanatory power.
Health care decisions are, by their nature, choices in which the stakes are high, the potential for
regret substantial, and the emotional overlays pervasive (Kőszegi, 2003; Hale and Dillard, 1995).
These powerful affective connotations explain in part why framing effects have been shown to
have the most substantial effects in health-related decisions (Khhberger, 1998). It is not
surprising that most people favor less certain prospects when the outcomes are framed in terms
of life and death – it is hard to even imagine choosing options that create a certainty of death,
even if there are fewer lives lost than the worst-case scenario in an uncertain prospect.
For similar reasons, better understanding the intersection of affective and cognitive heuristics
seems essential for sorting through how people make choices in health care settings. Emotional
considerations like regret avoidance that may play a small role in other decisions may be crucial
when it comes to the selection of medical treatments and sites for care. And given the high
ambient levels of anxiety that attend any health related choice, it becomes all the more important
to understand whether emotional factors enhance or impede cognitive engagement and heuristic
reasoning, much as they have been shown in prior research to affect information processing.
Extending Experimental Research on Bounded Rationality to Health Care Choices
There is both need and potential for additional research that extends studies of heuristic decision-
making to a wider array of health choices. To be sure, there is already a critical mass of
experimental studies documenting the impact of gain versus loss frames for health promotion;
another impressive body of scholarship demonstrates how endowment and anchoring effects
shape citizens’ valuation of public goods and programs, particularly those involving
environmental health consequences. Some types of health care (OTC prescriptions, optometric
services) are sufficiently similar to other consumer products that one can sensibly extrapolate the
findings of marketing research on choice heuristics and deferral to these health settings.
But important gaps remain. First, we have little sense of the nature or relative importance of
heuristics that come into play for high-salience decisions, such as choosing a health plan by the
frail elderly (who are almost certain to need to medical care in any given year), or selecting a
34
hospital or medical specialist for treating a particular health problem. Second, there are a
substantial number of heuristics that have never been studied in health settings, either for policy
choices or selection of health plans/providers. One might expect, for example, that affective
heuristics would be particularly important for selection of a primary care physician, but this
whole family of heuristics has been largely overlooked in health-related research. Third, we have
little sense of how heuristics influence the acquisition of and aversion to information mediated
through social networks. Finally, to date virtually every application of research on bounded
rationality in health settings has been limited to prospective decisions – e.g. choosing a health
plan at a new job, choosing a physician before being sick. Yet as we have noted above, many
important choices in medical settings are based on the evaluation of experience – and we have
little understanding of how heuristics affect choices and behavior under these conditions.
A New Flavor of Translational Research:
Although academic disciplines have only gradually accepted the implications of bounded
rationality (and some more slowly than others), even the most laggardly groups of scholars have
paid far more attention to these implications than have policy-makers and policy analysts. There
is clearly a need for some investment in scholarship that would focus on how this diffusion might
be further encouraged. This is in some ways analogous to recent initiatives at the National
Institutes of Health that fund research on the “translation” of new findings from basic medical
research into clinical settings. Indeed, some of the insights from clinical settings may be directly
applicable to policy-making.
But there are also important differences, as we discussed above. Although we identified a
number of ways in which interventions derived from the principles of libertarian paternalism
held promise for improving health care choices, many of the choices that have the greatest
external consequences – decisions about organ donation that extend the benefits of new advances
in transplantation, choices about medical providers that provide incentives for them to improve
their performance, attentiveness to information that might encourage the selection of more
appropriate treatments or health-related behaviors – all come with potentially problematic
ramifications if policymakers set default conditions in particular ways or limit choice sets to
exclude some treatment options. How citizens attribute blame for these outcomes will go a long
way toward determining the feasibility of policy changes. And how researchers understand that
35
process of blame attribution will help policymakers attend to the risks that accompany this
alternative policy paradigm.
The Bottom Line
Even sketched in this very general manner, it’s apparent that the agenda for translating insights
from bounded rationality to the health policy arena is an ambitious one. The theoretical
extensions will require careful thought, the extensions of empirical research considerable
investment of resources. The translational efforts may encounter resistance from policy analysts,
advocates, and decision-makers already set in their conventional ways of thinking about choice.
Ultimately, to be effective, they may require structural innovation in the ways in which we
monitor and govern the health care system. Such innovation would be challenging, given the
evident lack of governance that exists today. But the current absence of established institutional
authority also creates the opportunity for change that might otherwise not be feasible.
Yet it is precisely because these challenges exist that there is a need for some concerted effort in
this area. Diffusion and adoption of insights based on decision heuristics are unlikely to occur in
a timely or balanced manner without such a concerted initiative. The potential benefits from
reconstructing aspects of health policy from the perspective of bounded rationality seem well
worth the effort.
36
References
Akerlof, G., 1991.“Procrastination and Obedience” American Economic Review 81(2): 1-19 Amir, On, Dan Ariely, Alan Cooke, David Dunning, Botond Koszegi, Donald Lichtenstein, Nina
Mazar, Sendhil Mullainathan, Drazen Prelec, Eldar Shafir, Jose Silva, 2006. Behavioral Economics, Psychology, and Public Policy Working Paper.
Anderson, Christopher, 2003. “The Psychology of Doing Nothing: Forms of Decision Avoidance
Result from Reason and Emotion” Psychological Bulletin 129(1): 139-66. Apanovitch, Anne Marie, Danielle McCarthy and Peter Salovey, 2001. “Message Framing to
Motivate HIV Testing Among Low-Income, Ethnic Minority Women” Health Psychology 22(1): 60-67.
Armor, David and Shelley Taylor, 2002. “When Predictions Fail: The Dilemma of Unrealistic
Optimism” in Heuristics and Biases: The Psychology of Intuitive Judgment Eds. Thomas Gilovich, Dale Griffin, and Daniel Kahneman (New York: Cambridge University Press) 334-47.
Banks, Sara, Peter Salovey, Susan Greener, Alexander Rothman, Anne Moyer, John Beauvais,
and Elissa Epel, 1995. “The Effects of Message Framing on Mammography Use” Health Psychology 14(2): 179-84.
Baumann, A. O., R. B. Deber, and G. G. Thompson, 1991 ``Overconfidence among Physicians
and Nurses: The `Micro-Certainty, Macro-Uncertainty,' Phenomenon,'' Social Science and Medicine, 32(__): 167-174.
Beaulieu,N.D. 2002. Quality Information And Consumer Health Plan Choices. Journal of Health
Economics 21(__): 43–63. Bernheim B. Douglas and Antonio Rangel, 2005. “Behavioral Public Economics: Welfare and
Policy Analysis with Non-Standard Decision Makers” NBER Working Paper #11518 (Cambridge, MA: National Bureau of Economic Research, July)
Beshears, John, James Choi, David Laibson and Brigitte Madrian, 2006. “The Importance of
Default Options for Retirement Savings Outcomes: Evidence from the United States” NBER Working Paper #12009 (Cambridge, MA: National Bureau of Economic Research, July)
Birbaum, Michael and Teresa Martin, 2003. “Generalization Across People, Procedures and
Predictions: Violations of Stochastic Dominance and Coalescing” in Emerging Perspectives on Judgment and Decision Research Eds. Sandra Schneider and James Shanteau, (New York: Cambridge University Press): 84-107.
37
Block, Lauren and Punam Keller, 1995. “When to Accentuate the Negative: The Effects of Perceived Efficacy and Message Framing on Intentions to Perform a Health-Related Behavior” Journal of Marketing Research 32 (May): 192-203
Camerer, Colin and Howard Kunreuther, 1989. “Decision Processes for Low Probability Events:
Policy Implications” Journal of Policy Analysis and Management 8(_): 565-__. Camerer, Colin, Samuel Issacharoff, George Loewenstein, Ted O’Donoghue and Matthew
Rabin, 2003. “Regulation for Conservatives: Behavioral Economics and the Case for ‘Asymmetric Paternalism’” University of Pennsylvania Law Review 151: 1211-54.
Camerer, Colin and George Lowenstein, 2004. “Behavioral Economics: Past, Present and
Future” in Advances in Behavioral Economics Eds. C Camerer, G. Lowenstein and M Rabin, (New York and Princeton: Russell Sage Foundation and Princeton University Press): 3-51.
Caplin, Andrew, and K. Eliaz, 2003. “AIDS and Psychology: A Mechanism-Design Approach”,
RAND Journal of Economics, 34(4): 631-646. Caplin, Andrew and John Leahy, 2003. “Behavioral Policy.” In I. Brocas and J.D. Carrillo, eds.,
The Psychology of Economic Decisions: Vol. 1. New York: Oxford University Press, 2003. Chaulk, C. Patrick 1994. "Preventive Health Care in Six Countries: Models for Reform?" Health
Care Financing Review 15(4): 7-21. Chernew, Michael. Gautham Gowrisankaran, Dennis Scanlon, 2001. “Learning and the Value of
Information: The Case of Health Plan Report Cards” NBER Working Paper #8589 (Cambridge MA: National Bureau of Economic Research)
Conlisk, J. (1996), “Why Bounded Rationality?” Journal of Economic Literature, 34(2): 669-700 Conway, Michael and Michael Ross, 1984. “Getting What You Want by Revising What You
Had” Journal of Personality and Social Psychology 47(4): 738-48. Delli Carpini, Michael, and Scott Keeter. 1996. What Americans Know About Politics and Why It
Matters. New Haven: Yale University Press. De Ruyter, Ko and Martin Wetzels, 1998. “On the Complex Nature of Patient Evaluation of
General Practitioner Service” Journal of Economic Psychology 19: 565-90. DeSteno, David, Richard Petty, Duane Regener, Derek Rucker, 2000. “Beyond Valence in the
Perception of Likelihood: The Role of Emotion Specificity” Journal of Personality and Social Psychology 78(3): 397-416
Detweiler, Jerusha, Brian Bedell, Peter Salovey, Emily Pronin, Alexander Rothman, 1999,
“Message Framing and Sunscreen Use: Gain-Framed Messages Motivate Beachgoers” Health Psychology 18(2): 189-96.
38
Dhar, Ravi and Stephen Nowlis, 2004. “To Buy or Not to Buy: Response Mode Effects on Consumer Choice” Journal of Marketing Research 41(4): 423-32.
Dhar, Ravi, Stephen Nowlis and Steven Sherman, 1999. “Comparison Effects on Preference
Construction” Journal of Consumer Research 26(3): 293-306. Dhar, Ravi, 1997. “Consumer Preference for a No-Choice Option” Journal of Consumer
Research 24(2): 215-31. Dougherty, Michael, Scott Gronlund and Charles Gettys, 2003. “Memory as a Fundamental
Heuristic for Decision Making” in Emerging Perspectives on Judgment and Decision Research Eds. Sandra Schneider and James Shanteau, (New York: Cambridge University Press): 125-64.
Druckman, James, 2001. “Evaluating Framing Effects” Journal of Economic Psychology 22: 91-
101 EBRI/Commonwealth Fund, 2005. Consumerism in Health Care Survey (New York:
Commonwealth Fund). Edwards, A, S. Unigwe, G Elwyn and K Hood, 2005. “Personalized Risk Communication for
Informed Decision-Making About Entering Screening Programs” Cochrane Database of Systematic Reviews
El-Gamal, Mahmoud and David Grether, 1995. “Are People Bayesian? Uncovering Behavioral
Strategies” Journal of the American Statistical Association 90(432): 1137-45. Ellis, Randall, 1989. “Employee Choice of Health Insurance” Review of Economics and
Statistics 71(2): 215-31. Finney, Lila and Ronald Iannotti, 2002. “Message Framing and Mammography Screening: A
Theory-Driven Intervention” Behavioral Medicine 28(Spring): 5-14. Frank, Richard, 2004. “Behavioral Economics and Health Economics” NBER Working Paper
#10881 (National Bureau of Economic Research, Cambridge, MA) Frederick, Shane, 2005. “Cognitive Reflection and Decision Making” Journal of Economic
Perspectives 19(4): 25-42. Garbarino, Ellen and Julie Edell, 1997. “Cognitive Effort, Affect and Choice” Journal of
Consumer Research 24(3): 147-58. Gamson, William, 1992. Talking Politics (New York: Cambridge University Press) Gawande AA, Blendon RJ, Brodie M, et al, 1998. “Does Dissatisfaction With Health Plans Stem
From Having No Choices” Health Affairs 17 (5): 184-194 Gigerenzer, Gerd and Adrian Edwards, 2003. “Simple Tools for Understanding Risk: From
Innumeracy to Insight” British Medical Journal 327: 741-4.
39
Gilbert, Daniel, Elizabeth Pinel, Timothy Wilson, Stephen Blumberg and Thalia Wheatley, 2002. “Durability Bias in Affective Forecasting” in Heuristics and Biases: The Psychology of Intuitive Judgment Eds. Thomas Gilovich, Dale Griffin, and Daniel Kahneman (New York: Cambridge University Press) 292-312.
Gilovich, Thomas, Dale Griffin, and Daniel Kahneman, 2002 Heuristics and Biases: The
Psychology of Intuitive Judgment Eds. (New York: Cambridge University Press) Gilovich, Thomas and Dale Griffin, 2002. “Introduction – Heuristics and Biases: Then and Now”
in Heuristics and Biases: The Psychology of Intuitive Judgment Eds. Thomas Gilovich, Dale Griffin, and Daniel Kahneman (New York: Cambridge University Press): 1-18.
Glaeser, E.L, 2003. “Psychology and the Market”, NBER Working Paper #10203, Cambridge,
MA. December Gold, Marsha, Michael Sinclair, Mia Cahill, Natalie Justh and Jessica Mittler, 2001. Medicare
Beneficiaries and Health Plan Choice (Washington DC: Mathematica Policy Research). Gold Marsha, Laurie Achman and Randall Brown, 2003 “The Salience Of Choice For Medicare
Beneficiaries” Managed Care Quarterly 2003;11; 24-33. Gold Marsha, 2006. The Growth of Private Plan in Medicare, 2006. (Kaiser Family Foundation,
Menlo Park, CA).
Goold S.D., 1998. “Money and Trust: Relationship Between Patients, Physicians and Health Plans” Journal of Health Politics, Policy and Law 23(4): 687-95.
Goold, Susan Dorr and Glenn Klipp, 2002. “Managed Care Members Talk About Trust” Social
Science and Medicine 54: 879-88. Gordon, Colin, 2003. Dead on Arrival: The Politics of Health Care in Twentieth-Century
America (Princeton, Princeton University Press). Green D, KE Jacowitz, D Kahneman and D McFadden, 1998. “Referendum Contingent
Valuation, Anchoring, and Willingness to Pay for Public Goods” Resource and Energy Economics 20(_): 85-116.
Grether, David, 1992. “Testing Bayes Rule and the Representativeness Heuristic: Some
Experimental Evidence” Journal of Economic Behavior and Organization 17: 31-57. Haines, Beth and Colleen Moore, 2003. “Integrating Themes from Cognitive and Social
Cognitive Development into the Study of Judgment and Decision Making” in Emerging Perspectives on Judgment and Decision Research Eds. Sandra Schneider and James Shanteau, (New York: Cambridge University Press): 246-83.
Hale JL and JP Dillard, 1995. “Fear Appeals in Health Campaigns: Too Much, Too Little, or Just
Right?” Designing Health Messages Eds. E Maibach and R Parrott (Thousand Oaks: Sage Publications): 65-80.
40
Hanoch, Yaniv and Thomas Rice, 2006. “Can Limiting Choice Increase Social Welfare? The Elderly and Health Insurance” Milbank Quarterly 84; 37-73
Hays, R and J Ware, 1986. “My Medical Care is Better Than Yours: Social Desirability and
Patient Satisfaction Ratings” Medical Care 24(6): 519-24. Hawkins, Scott and Reid Hastie, 1990. “Hindsight: Biased Judgments of Past Events After the
Outcomes Are Known” Psychological Bulletin 107(3): 311-27. Heath, C and J. Soll, 1996. “Mental Budgeting and Consumer Decisions” Journal of Consumer
Research 23(1): 40-52. Hibbard, Judith, Nancy Berkman, Lauren McCormack, Elizabeth Jael, 2002a. “The Impact of a
CAHPS Report on Employee Knowledge, Beliefs, and Decision” Medical Care Research and Review 59(1): 104-116.
Hibbard JH, Slovic P, Peters EM, Finucane M, 2002b. Strategies for Reporting Health Plan
Performance Information to Consumers: Evidence from Controlled Studies. Health Services Research. April 2002 37(2): 291-313
Hibbard, Judith H., Jean Stockard and Martin Tussler, 2005. “Hospital Performance Reports:
Impact on Quality, Market Share, and Reputation” Health Affairs 24(4): 1150-60. Houghton, David, 1998. Historical Analogies and the Cognitive Determinants of Domestic
Policy Making” Political Psychology 19(2): 279-303. Hsee CK, and EU Weber, 1997. “A Fundamental Prediction Error: Self-Other Discrepancies in
Risk Preference” Journal of Experimental Psychology: General 126: 45-53. Hsu J, Schmittdiel J, Krupat E, et al., 2003. Patient Choice - A Randomized Controlled Trial Of
Provider Selection Journal Of General Internal Medicine 18 (5): 319-325 Isen, Alice and Aparna Labroo, 2003. “Some Ways in Which Positive Affect Facilitates Decision
Making and Judgment” in Emerging Perspectives on Judgment and Decision Research Eds. Sandra Schneider and James Shanteau, (New York: Cambridge University Press): 365-93.
Iyengar, Shanto, 1991. Is Anyone Responsible? How Television Frames Political Issues
(Chicago, University of Chicago Press). Iyengar, Sheena and Mark Lepper, 2000. “When Choice is Demotivating: Can One Desire Too
Much of a Good Thing?” Journal of Personality and Social Psychology 79(6): 995-1006. Jacobs, Lawrence R. and Robert Y Shapiro, 1994. AQuestioning the Conventional Wisdom on
Public Opinion Toward Health Reform@ PS: Political Science and Politics 27(2): 208-14.
41
Jacobs, L.R., R.Y. Shapiro, 1999. “The American Public’s Pragmatic Liberalism Meets Its Philosophical Conservatism” Journal of Health Politics, Policy and Law 24(5): 1021-31.
Jennings, M Kent, 1999. “Political Responses to Pain and Loss” American Political Science
Review 93(1): 1-13. Johnson, Eric, John Hershey, Jacqueline Meszaros, and Howard Kunreuther, 1993. “Framing,
Probability Distortions and Insurance Decisions” Journal of Risk and Uncertainty 7(_): 35-51.
Jolls, Christine, Cass Sunstein and Richard Thaler, 1998. “A Behavioral Approach to Law and
Economics” Stanford Law Review 50(5): 1471-1550. Jones, Bryan, 1994. Reconceiving Decision-Making in Democratic Politics: Attention, Choice
and Public Policy (Chicago: University of Chicago Press). Kahneman, Daniel, Ilana Ritov and David Schkade, 1999. “Economic Preferences or Attitude
Expressions?” Journal of Risk and Uncertainty 19(1-3): 203-35. Kahneman, Daniel and Shane Frederick, 2002. “Representativeness Revisited: Attribute
Substitution in Intuitive Judgment” in Heuristics and Biases: The Psychology of Intuitive Judgment Eds. Thomas Gilovich, Dale Griffin, and Daniel Kahneman (New York: Cambridge University Press) 49-81.
Kaiser Family Foundation, 2000. National Survey on Americans As Health Care Consumers: An
Update on the Role of Quality Information (Menlo Park, CA: Henry J Kaiser Family Foundation).
Kaiser Family Foundation, 2001. Understanding the Effects of Direct-to-Consumer Prescription
Drug Advertising (Menlo Park, CA: Henry J Kaiser Family Foundation). Kaiser Family Foundation, 2006a. Medciare: Prescription Drug Coverage Under Medicare
Beneficiaries Publication No. 7453 (Menlo Park, CA: Kaiser Family Foundation) Kaiser Family Foundation, 2006b. Senior’s Early Experience with the Medicare Prescription
Drug Benefit – April, 2006 Publication No. 7501 (Menlo Park, CA: Kaiser Family Foundation)
Kaiser Family Foundation, 2006c. Kaiser Family Foundation National Survey of Physicians:
Findings on Medicare Part D Publication No 7554 (Menlo Park, CA: Kaiser Family Foundation)
Kaiser Family Foundation, 2006d. Kaiser Family Foundation National Survey of Pharmacists:
Findings on Medicare Part D Publication No 7555 (Menlo Park, CA: Kaiser Family Foundation)
Kalda R, Polluste K, Lember M, 2003 Patient Satisfaction With Care Is Associated With
Personal Choice Of Physician Health Policy 64 (1): 55-62
42
Korobkin, Russell and Thomas Ulen, 2000. “Law and Behavioral Science: Removing the Rationality Assumption from Law and Economics” California Law Review 88(4): 1051-1144.
Kőszegi, B, 2003, “Health, Anxiety and Patient Behavior”, Journal of Health Economics, 22(6):
1073-1084. Krupat E, Stein T, Selby JV, et al., 2002 Choice Of A Primary Care Physician And It's
Relationship To Adherence Among Patients With Diabetes American Journal Of Managed Care 8 (9): 777-784
Khhberger, Anton, 1998. “The Influence of Framing on Risky Decisions: A Meta-Analysis”
Organizational Behavior and Human Decision Processes 75(1): 23-55. Kunreuther, Howard et al, 2001. “Making Low Probabilities Useful” Journal of Risk and
Uncertainty 23: 103-__ Kuhn, Thomas S, 1962. The Structure of Scientific Revolutions. (Chicago: Chicago University
Press. Kuran, Timur and Cass Sunstein, 1999. “Availability Cascades and Risk Regulation” Stanford
Law Review 51(4): 683-68. Lambert, DA and TG McGuire, 1990. “Political and Economic Determinants of Insurance
Regulation in Mental Health” Journal of Health Politics, Policy and Law 15(__): 169-89. Larrick, Richard and Terry Boles, 1995. “Avoiding Regret in Decisions with Feedback: A
Negotiation Example” Organizational Behavior and Human Decision Processes 63(1): 85-97.
Laschober, Mary, 2005 “Estimating Medicare Advantage Lock-In Provisions Impact on
Vulnerable Medicare Beneficiaries” Health Care Financing Review 26(3): 63-79 Lau, Richard R. and David P. Redlawsk. 2001. "Advantages and Disadvantages of Cognitive
Heuristics in Political Decision Making." American Journal of Political Science, 45(October): 951 - 971.
Leiberman A and S Chaiken, 1992. “Defensive Processing of Personally Relevant Health
Messages” Personality and Social Psychology Bulletin 18(_): 669-79. Lerman, C. and R. Croyle, 1996. “Genetic Testing for Cancer Predisposition” Journal of the
National Cancer Institute Monograph 17 (Bethesda, MD). Lerman, C., Hughes, C., Trock, B.J., Myers, R.E., Main, D., Bonney, A., Abbaszadegan, M.R.,
Harty, A.E., Franklin, B.A., Lynch, J.F., Lynch, H.T., 1999. Genetic testing in families with hereditary nonpolyposis colon cancer. Journal of the American Medical Association 281 (17), 1618–1622.
43
Lerman, C., Narod, S., Schulman, K., Hughes, C., Gomez-caminero, A., Bonney, G., Gold, K., Trock, B., Main, D.,Lynch, J., Fulmore, C., Snyder, C., Lemon, S.J., Conway, T., Tonin, P., Lenoir, G., Lynch, H., 1996b. BRCA1 testing in families with hereditary breast-ovarian cancer. A prospective study of patient decision making and outcomes. Journal of the American Medical Association 275 (24), 1885–1892.
Lerner, Jennifer and Dacher Keltner, 2000. “Beyond Valence: Toward a Model of Emotion-
Specific Influences on Judgment and Choice” Cognition and Emotion 14(4): 473-93. Lerner, Jennifer and Dacher Keltner, 2001. “Fear, Anger and Risk” Journal of Personality and
Social Psychology 81(1): 146-59 Lerner, Jennifer, Roxana Gonzalez, Deborah Small and Baruch Fischhoff, 2003. “Effects of Fear
and Anger on Perceived Risks of Terrorism: A Natural Experiment” Psychological Science 14(2): 144-50.
Lewin SA, ZC Skea, V Entwistle, M Zwarenstein, J. Dick, 2005. “Interventions for Providers to
Promote a Patient-Centered Approach In Clinical Consultation” Cochrane Database of Systematic Reviews
Lowenstein, G, 2000. “Costs and Benefits of Health and Retirement-Related Choice” in S.
Burke, E Kingson and U Reinhardt (eds), Social Security and Medicare: Individual vs Collective Risk and Responsibility; Washington: Brookings Institution Press
Loewenstein, George, Elke Weber, Christopher Hsee, Ned Welch, 2001. “Risk as Feelings”
Psychological Bulletin 127(2): 267-86 Loewenstein, George, 2004. “Out of Control: Visceral Influences on Behavior” in Advances in
Behavioral Economics Eds. C Camerer, G. Lowenstein and M Rabin, (New York and Princeton: Russell Sage Foundation and Princeton University Press): 689-723.
McFadden, Daniel, 1974. “On Some Facets of Betting” in Essays in Economic Behavior Under
Uncertainty Ed. M Balch et al (Amsterdam: North Holland): 126-31. McFadden, Daniel, 1999. “Rationality for Economists?” Journal of Risk and Uncertainty 19 (1-
3): 73-105. McGraw, Kathleen, 1991. “Managing Blame: An Experimental Test of the Effects of Political
Accounts” American Political Science Review 85(4): 1133-57. Madrian, Brigitte and Dennis Shea, 2001. “The Power of Suggestion: Inertia in 401(k)
Participation and Savings Behavior” Quarterly Journal of Economics 116(_): 1149- __ Maheswaran, Durairaj and Joan Meyers-Levy, 1990. “The Influence of Message Framing and
Issue Involvement” Journal of Marketing Research 27 (August): 361-7. Marquis, Susan and Martin Holmer, 1996. “Alternative Models of Choice Under Uncertainty and
Demand for Health Insurance” Review of Economics and Statistics 78(3): 421-27.
44
Mechanic, David, 1997. “Managed Care as a Target for Distrust” Journal of the American Medical Association 277(22): 1810-12.
Mechanic, David and S. Meyer, 2000. “Concepts of Trust Among Patients with Serious Illness”
Social Science and Medicine 5195): 657-68. Mello Michelle, Sally Stearns and Edward Norton, 2002. “Do Medicare HMOs Still Reduce
Health Services Use After Controlling For Selection Bias?” Health Economics 11; 323-340.
Milstein, Arnold and Nancy Adler, 2003. “Out of Sight, Out of Mind: Why Doesn’t Widespread
Clinical Quality Failure Command Our Attention?” Health Affairs 22(2): 119-27. Mondak, Jeffrey, 1994. "Cognitive Heuristics, Heuristic Processing and Efficiency in Political
Decisionmaking" in Research in Micropolitics: New Directions in Political Psychology Eds. Michael Delli Carpini, Leonie Huddy, and Robert Shapiro, (Greenwich, CT: JAI Press): 117-42.
Nicol-Mauveyraud, J. 2003. “Mind Over Money” Psychology 16(5): 240-42. Noble, Alice and Troyen Brennan, 1999. AThe Stages of Managed Care Regulation: Developing
Better Rules@ Journal of Health Politics, Policy and Law 24(6): 1275-1935 Noll, Roger and James Krier, 1990. “Some Implications of Cognitive Psychology for Risk
Regulation” Journal of Legal Studies 19: 747-__ Oberlander, Jonathan, 2000. “Is Premium Support the Right Medicine For Medicare?” Health
Affairs 19(5): 84-99. O’Brien, Bernie, Kirsten Gertsen, Andrew Willan and Lisa Faulkner, 2002. “Is There a Kink in
Consumers’ Threshold Value for Cost-Effectiveness in Health Care?” Health Economics 11: 175-80.
Patenaud, Andrea, 2005. Genetic Testing for Cancer: Psychological Approaches for Helping
Patients and Families (Washington DC: American Psychological Association). Pear, Robert, 2006. “Medicare Insurers Plan New Drug Options” New York Times (October 1):
p.25 Peterson, Mark, Shaila Miranda, Peter Smith and Valerie Haskell, 2003. “The Sociocultural
Contexts of Decision Making in Organizations” in Emerging Perspectives on Judgment and Decision Research Eds. Sandra Schneider and James Shanteau, (New York: Cambridge University Press): 512-58.
Pliske, Rebecca and Gary Klein, 2003. “The Naturalistic Decision-making Perspective” in
Emerging Perspectives on Judgment and Decision Research Eds. Sandra Schneider and James Shanteau, (New York: Cambridge University Press): 559-85.
45
Rabin, M, 1998, “Psychology and Economics” Journal of Economic Literature 36 (1):11-46. Rabin, M. 2002, “Inference by Believers in the Law of Small Numbers”, Quarterly Journal of
Economics 117(3): 775-816. Rabin M. and J. Schrag, 1999, “First Impressions Matter: A Model of Confirmatory Bias”
Quarterly Journal of Economics 114(1) 37-82. Rabin, Matthew and Richard Thaler, 2001. “Anomalies: Risk Aversion” Journal of Economic
Perspectives 15(1): 219-32. Redelmeier, Donald, Derek Koehler, Varda Liberman, Amos Tversky, 1995. “Probability
Judgment in Medicine: Discounting Unspecified Alternatives” Medical Decision Making 15(_): 227-31.
Redelmeier Donald and Daniel Kahneman, 1996. “Patients’ Memories of Painful Medical
Treatments: Real-time and Retrospective Evaluations of Two Minimally Invasive Procedures” Pain 66: 3-8.
Redelmeier, Donald, Michael Schull, Janet Hux, Jack Tu, Lorraine Ferris, 2001a. “Problems for
Clinical Judgment: Eliciting an Insightful History of Present Illness” Canadian Medical Association Journal 164(5): 647-51.
Redelmeier, Donald, Jack Tu, Michael Schull, Lorraine Ferris, Janet Hux, 2001b. “Problems for
Clinical Judgment: Obtaining a Reliable Past Medical History” Canadian Medical Association Journal 164(6): 809-13.
Rettinger, David and Reid Hastie, 2003. “Comprehension and Decision Making” in Emerging
Perspectives on Judgment and Decision Research Eds. Sandra Schneider and James Shanteau, (New York: Cambridge University Press): 165-200.
Rosenthal Marsha, and Mark Schlesinger, 2002. “Not Afraid To Blame: The Neglected Role Of
Blame Attribution In Medical Consumerism And Some Implications For Health Policy” Milbank Quarterly 2002; 80(1): 41-95.
Rosenthal, Meredith and Arnold Milstein, 2004. “Awakening Consumer Stewardship Over
Health Benefits: Prevalence and Differentiation of New Health Plan Models” Health Services Research 39(4, Part II): 1055-70.
Rostain, Tanina, 2000. “Educating Homo Economicus: Cautionary Notes on the New Behavioral
Law and Economics Movement” Law and Society Review 34(4): 973-1006. Rothman, Alexander and Peter Salovey, 1997. “Shaping Perceptions to Motivate Healthy
Behavior: The Role of Message Framing” Psychological Bulletin 121(1): 3-19. Rothschild, Jeffrey and Lucian Leape, 2000. The Nature and Extent of Medical Injuries in Older
Patients (Washington DC: AARP Policy Institute)
46
Sage, William, 1999. “Regulating Through Information: Disclosure Laws and American Health Care” Columbia Law Review 99(7): 1701-1829.
Scanlon, D. M. Chernew, C. McLaughlin and G. Solon , 2002. The Impact Of Health Plan
Report Cards On Managed Care Enrollment. Journal of Health Economics, 21(__): 19–41 Scanlon Dennis, S. Swaminathan, Michael Chernew, J. Bost and J. Shevock, 2005. “Competition
And Health Plan Performance: Evidence From Health Maintenance Organization Insurance Markets” Medical Care 43; 338-346.
Schlesinger Mark, Benjamin Druss, Tracey Thomas, 1999. “No Exit? The Effect Of Health
Status On Dissatisfaction And Disenrollment From Health Plans” Health Services Research 34(2): 547-76.
Schlesinger Mark, Richard Lau, 2000. The Meaning And Measure Of Policy Metaphors.
American Political Science Review 94(3): 611-26. Schlesinger Mark, 2002. “On Values And Democratic Policymaking: The Fragile Consensus
Around Market-Oriented Medical Care” Journal of Health Politics, Policy and Law 27(6): 889-926
Schlesinger Mark, Shannon Mitchell, Brian Elbel, 2002. “Voices Unheard: Barriers To The
Expression Of Dissatisfaction With Health Plans” Milbank Quarterly 2002; 80(4): 709-755 Schlesinger Mark, Shannon Mitchell, BradfordGray, 2004. “Restoring Public Legitimacy To The
Nonprofit Sector: A Survey Experiment Using Descriptions Of Nonprofit Ownership” Nonprofit and Voluntary Sector Quarterly 33(4): 673-710.
Schmittdiel J, Selby JV, Grumbach K, et al., 1997. Choice Of A Personal Physician And Patient
Satisfaction In A Health Maintenance Organization Journal Of The American Medical Association 278 (19): 1596-1599
Schone BS and P.F.Cooper, 2001. Assessing The Impact Of Health Plan Choice Health Affairs
20 (1): 267-275 Schwartz, Barry, 2004. The Paradox of Choice: Why More Is Less (New York: Harper Collins). Schwartz, Norbert, 2002. “Feelings as Information: Moods Influence Judgments and Processing
Strategies” in Heuristics and Biases: The Psychology of Intuitive Judgment Eds. Thomas Gilovich, Dale Griffin, and Daniel Kahneman (New York: Cambridge University Press): 534-47.
Shaller D., S. Sofaer, S.D. Findlay et al., 2003. “Consumers and Quality-Driven Health Care: A
Call to Action” Health Affairs 22(2):95-101. Shaller, Daniel, 2005. Consumers in Health Care: The Burden of Choice (Oakland, CA:
California Healthcare Foundation).
47
Sherman, Steven, Robet Cialdini, Donna Schwartzman and Kim Reynolds, 1985. “Imagining Can Heighten or Lower the Perceived Likelihood of Contracting a Disease: The Mediating Effect of Ease of Imagery” Personality and Social Psychology Bulletin, 11(_): 118-27.
Simonson I and A Tversky, 1992. “Choice in Context: Tradeoff Contrast and Extremeness
Aversion” Journal of Marketing Research 29(_): 281-95. Skocpol, Theda, 1996. Boomerang: Health Care Reform and the Turn Against Government
(New York: W.W. Norton). Sloan, Frank, 2001. “Arrow’s Concept of the Health Care Consumer: A Forty-Year
Retrospective” Journal of Health Politics, Policy and Law 26(50: 899-912. Sloman, Steven, 2002. “Two Systems of Reasoning” in Heuristics and Biases: The Psychology
of Intuitive Judgment Eds. Thomas Gilovich, Dale Griffin, and Daniel Kahneman (New York: Cambridge University Press): 379-96.
Slovic, Paul, Baruch Fischhoff, Sarah Lichtenstein, Bernard Corrigan and Barbara Combs, 1977.
“Preference for Insuring Against Probable Small Losses: Insurance Implications” The Journal of Risk and Insurance 44(2): 237-58.
Slovic, Paul, Melissa Finucane, Ellen Peters, and Donald MacGregor, 2002. “The Affect
Heuristic” in Heuristics and Biases: The Psychology of Intuitive Judgment Eds. Thomas Gilovich, Dale Griffin, and Daniel Kahneman (New York: Cambridge University Press): 397-20.
Sniderman, Paul Richard Brody, and Philip Tetlock, with Henry Brady, 1991. Reasoning and
Choice: Explorations in Political Psychology (New York: Cambridge University Press) Sofaer S, Gruman J. 2003. Consumers Of Health Information And Health Care: Challenging
Assumptions And Defining Alternatives. American Journal of Health Promotion 18(2): 151-156.
Soll, J, 1996. “Determinants of Overconfidence and Miscalibration: The Roles of Random Error
and Ecological Structure” Organizational Behavior and Human Decision Processes 65(_): 117-37.
Stone, Deborah, 1989. "Causal Stories and the Formation of Policy Agendas" Political Science
Quarterly 104(2): 281-300. Suen W, 1991. “The Value of Product Diversity” Oxford Economic Papers 43; 217-223. Sunstein, Cass, 2002. “Probability Neglect: Emotions, Worst Cases and the Law” Yale Law
Journal 112(1): 61-107 Taylor, Shelley and Jonathan Brown, 1988. “Illusion and Well-Being: A Social Psychological
Perspective on Mental Health” Psychological Bulletin 103(2): 193-210.
48
Tetlock, Philip, 2002. “Intuitive Politicians, Theologians and Prosecutors: Exploring the Empirical Implications of Deviant Functionalist Metaphors” in Heuristics and Biases: The Psychology of Intuitive Judgment Eds. Thomas Gilovich, Dale Griffin, and Daniel Kahneman (New York: Cambridge University Press): 582-99.
Thaler, R, 1999. “Mental Accounting Matters” Journal of Behavioral Decision-Making 12(2):
183-206. Thaler, Richard and Cass Sunstein, 2003. “Libertarian Paternalism” American Economic Review
93(2): 175-9. Tiedens, Larissa and Susan Linton, 2001. “Judgment Under Emotional Certainty and
Uncertainty: The Effects of Specific Emotions on Information Processing” Journal of Personality and Social Psychology 81(6): 973-88.
Tomes, Nancy, 2006. “Patients or Health-Care Consumers? Why the History of Contested Terms
Matters” History and Health Policy in the United States Eds. Rosemary Stevens, Charles Rosenberg and Lawton Burns (New Brunswick, Rutgers University Press): 83-110.
Tversky, Amos and Daniel Kahneman, 1974. “Judgment Under Uncertainty: Heuristics and
Biases” Science 185(Sept 27): 1124-31. Tversky, Amos and Daniel Kahneman, 1984. “Extensional versus Intuitive Reasoning: The
Conjunction Fallacy in Probability Judgment” Psychological Review 91: 293-315. Tversky, Amos and Eldar Shafir, 1992. “Choice Under Conflict: The Dynamics of Deferred
Decision” Psychological Science 6(November): 358-61. Tverky, Amos and Derek Koehler, 1994. “Support Theory: A Nonextensional Representation of
Subjective Probability” Psychological Review 101(_): 547-67. Uhrig JD, Short PF 2002-03. Testing The Effect Of Quality Reports On The Health Plan
Choices Of Medicare Beneficiaries. Inquiry. 39(4):355-71. Vaiana, Mary and Elizabeth McGlynn, 2002. “What Cognitive Science Tells Us About the
Design of Reports for Consumers” Medical Care Research and Review 59(1): 3-35. Viscusi, W. Kip, Wesley A Magat and Joel Huber, 1987. “An Investigation of the Rationality of
Consumer Valuation of Multiple Health Risks” RAND Journal of Economics 18(4): 465-79.
Wagner, Todd, Teh-wei Hu, Grace Dueñas, Celia Kaplan, Bang Nguyen, Rena Pasick, 2001
“Does Willingness to Pay Vary By Race/Ethnicity? An Analysis Using Mammography Among Low-Income Women” Health Policy 58: 275-88
49
Wathieu, Luc, Lyle Brenner, Ziv Carmon, Amitava Chattopadhyay, Klaus Wertenbroch. Aimee Drolet, John Gourville, A. V. Mukhukrrishnan, Nathan Novemsky, Rebecca Ratner, George Wu, 2002. “Consumer Control and Empowerment: A Primer” Marketing Letters 13(3): 297-305
Weaver, R. Kent, 1986. “The Politics of Blame Avoidance” Journal of Public Policy 6: 371-98. Weinstein, Neil and William Klein, 1995. “Resistance of Personal Risk Perceptions to Debiasing
Interventions” Health Psychology 14(_): 132-40. Werner, Rachel and David Asch, 2005. “The Unintended Consequences of Publicly Reporting
Quality Information” Journal of the American Medical Association 293(10): 1239-44. Wilkes, Michael, Robert Bell and Richard Kravitz, 2000. “Direct-to-Consumer Prescription Drug
Advertising: Trends, Impact and Implications Health Affairs 19: 110-28. Witte, Kim and Mike Allen, 2000. “A Meta-Analysis of Fear Appeals: Implications for Effective
Public Health Campaigns” Health Education and Behavior 27(5): 591-615.
50
Endnotes 1 Health care professionals were even more critical of Part D’s implementation than were the program’s
beneficiaries. Eighty-four percent of physicians and 90 percent of pharmacists gave the new program a grade of C or lower.
2 In the economics literature, models of bounded rationality have been most frequently applied to other
decision-makers who are arguably impaired in cognition, such as those addicted to various substances. “The methods of behavioral economics might be expected to be prevalent in modifying traditional models to take account of these features that appear to conflict with simple notions of rationality in economic behavior. Yet the application of behavioral economics to issues in health economics have been largely confined to understanding addictive behavior around cigarettes, drugs and alcohol.” (Frank, 2004, p.3)
3 For example, researchers have documented that as genetic testing has made it possible to identify
predispositions to develop particular illnesses, many people are not only averse to the tests themselves but will avoid health care settings in which they might be inadvertently exposed to that information. (Caplin and, Leahy, 2003).
4 The literature on choice deferral is reasonably robust, but little applied to health care choices. See, for
example (Dhar, 1997). 5 Hibbard and colleagues have shown in a series of experiments that an increase in choices is associated
with more frequent errors in selecting among plans (See, for example, Hibbard et al., 2002a). 6 Increasing awareness of disparities should induce minority patients to be more alert for possible forms
of mistreatment. But the literature on cognition suggests that their assessment of their own recent health care experiences will also cause them to revise their expectations for health care generally – in ways that may discourage them from taking action, because they fear that all doctors or health plans will treat minority patients poorly. See, for example (Rabin, 2002); these theoretical predictions have not yet been tested in health settings.
7 Many people prefer multiple insurance policies that cover narrow-defined risks, even if these provide
less coverage and are more expensive than more comprehensive policies. People also favor policies with implicit or explicit rebates for healthy behavior over those with deductibles, even if the expected value of the latter is considerably higher than the former (Johnson et al., 1993),
8 Willingness to pay for public programs affecting health has been shown to be powerfully affected by
whether these choices care framed as a willingness to pay for increased health or a compensation for increased health care threats. Because the social perceptions of various health concerns will be inconsistently framed (some as losses, others as health gains), so too will public preferences for the associated policy interventions (O’Brien et al, 2002)
9 This categorization leave out a fourth set of factors that are important in the bounded rationality
literature: those that affect the accessibility of particular sorts of information from memory. Although these are clearly relevant to some of the decision biases that we identified in our list of anomalies, they tend to be most relevant in affecting decisions regarding willingness to pay for collective investment or risk perceptions, neither of which are a central focus of our current analysis.
51
10 It’s not clear whether this aspect of pattern-finding is motivated by emotional considerations or simply
a flawed understanding of probabilities. Past studies suggest that belief in the law of small numbers holds for situations in which the personal stakes are small, inducing little emotional pressure to reduce randomness in that instance. Nonetheless, the drive to find order in one’s personal experiences may be so pervasive that it carries over to other settings, including hypothetical judgments. We are not aware of any studies that test whether the tendency to impose predictability is greater in circumstances that induce greater personal anxiety.
11 These asymmetric risk preferences can be reversed for low probability events, which under prospect
theory are assigned a greater weight than one would expect solely based on their objective probability. 12 In choices that involve future states of the world, framing is entirely determined by the characterization
of the outcome. For settings that involve individual choosing to pursue some action, however, framing is more complex (Rothman and Salovey, 1997). Gain frames can involve someone acting to achieve a desirable outcome, or avoiding an undesirable one. Loss frames, conversely, can involve exposure to an undesirable outcome, or failure to attain a desired one.
13 It often difficult to separate out the sort of status quo bias that might emerge from the evaluation of
uncertain prospects from that produced by perceptual filters or unrealistically optimistic self-assessments of consumers’ capabilities for effective choice.
14 This paradoxical assessment of insurance may explain the anomaly reported above involving cross-
national studies of citizens’ sense of security with respect to medical spending. If comprehensive national health insurance plans protect against many financial risks, but leave those risks ill-defined, people may nonetheless feel insecure because they cannot concretely anticipate the costs that they are being protected against.
15 Deadlines are typically proposed to deal with concerns about pathologically high rates of time
discounting, but are sometimes justified in terms of other heuristic–based biases. 16 This could be done through franchise licenses; regular rebidding for the license could maintain
competitive pressures while limiting the number of plan choices in each market.
52