Some terms come redolent with positive associations. This is no less true in the policy lexicon of democracy (freedom, responsiveness, leadership and voting) than in everyday English (hope, friend, kindness and chocolate).Footnote 1 ‘Choice’ makes both lists. Its connotations reach to the heart and soul of every liberal polity, evoking the virtues of autonomy, self-determination, commitment and individual responsibility. Its emotional overtones are equally affirmative: to be ‘choice’ is not simply to be good, but to be the very best – the cream of the crop.
It is therefore not surprising that ‘choice’ has become a watchword for contemporary health care reformersFootnote 2 and remains a centerpiece of the health policy agenda whichever party controls the political agenda (Bevan and van de Ven, Reference Bevan and van de Ven2010, this issue; Paulus et al., Reference Paulus2003; Klein, Reference Klein2007). Choice among providers and health plans also seems consistent with the aspirations to patient autonomy that have dominated bioethics over the past 25 years (Schneider, Reference Schneider2006; Jennings, Reference Jennings2007). Choice in medical settings constructs a policy bridge to other socially valued domains (education, pensions and housing) in which the choice motif has become equally central to neo-liberal policies.
Promoting ‘choice’ also seems an attractive way for policymakers to tackle constructively the complex challenges of improving health system performance. Though assigned different labels – failures of ‘care coordination’ in France and Germany (Or et al., Reference Or, Cases, Lisac, Vrangbæk and Ulrika Winblad Spångberg2010, this issue), patients who do not attend their hospital appointments in the UK (Dixon et al., Reference Dixon, Robertson and Dixon2010, this issue) – patients’ disconnection from the health care system represents a persisting and pervasive problem in every country (Schoen et al., Reference Schoen, Osborn, Huynhm, Doty, Zapert, Peugh and Davis2005, Reference Schoen, Osborn, Doty, Bishop, Peugh and Murukutla2007). Enhanced choice seemingly tackles disconnectedness from both ends: first, by encouraging patients to become more actively engaged in the process of obtaining medical care, and second, by inducing clinicians to be more responsive to their patients’ preferences in these regards.
Perhaps most crucially, choice enhancing policies have – from policymakers’ perspective – the paramount virtue of simplicity: they seem a straightforward policy that benefits everyone. Just like taking a daily vitamin, going to yoga or ingesting a baby aspirin to reduce cardiovascular risks. After all, choice comes naturally in a consumerist society. Choice-promoting policies thus ‘normalize’ health care, making it seem more like other commodities that are familiar purchases (Bevan et al., Reference Bevan, Helderman and Wilsford2010, this issue).
Of course, choice in health care is not that simplistically virtuous. The authors and editors of this special issue recognize this, exploring some key complexities: (1) the importance of legal entitlements and prohibitions in shaping consumers’ choice sets (Bevan et al., Reference Bevan, Helderman and Wilsford2010, this issue; Flood and Haugan, Reference Flood and Haugan2010, this issue), (2) the differences between choosing among providers or insurers, and the very different policy approaches required to enhance each (Bevan and van de Ven, Reference Bevan and van de Ven2010, this issue; Dixon et al., Reference Dixon, Robertson and Dixon2010, this issue), (3) the ways in which institutional contexts and policy legacies alter expectations regarding choice and the practices through which it can be encouraged (Or et al., Reference Or, Cases, Lisac, Vrangbæk and Ulrika Winblad Spångberg2010, this issue; Flood and Haugan, Reference Flood and Haugan2010, this issue) and (4) the ways in which payment incentives and practice arrangements mediate (and often impede) the capacity of clinicians to support patient choice (Dixon et al., Reference Dixon, Robertson and Dixon2010, this issue).
These are all important insights. In this commentary, however, I propose to redirect our attention from the system to the psyche, by exploring how health care choices are experienced by individual patients. Though choice-promoting policies may simplify demands on public officials, choice itself is never simple from the patient’s perspective. The stakes are too high, making the choices fraught with anxiety. The consequences are impossible to fully anticipate, rendering judgments at best semi-informed guesses. And the circumstances are never really amenable: either the patients are healthy – and therefore can not be bothered to dwell much on medical matters – or they are sick, which often limits their capacity to process a lot of complicated information.
Yet difficult as choice may be in health care settings, its challenges cannot be circumvented. Even if policymakers did not strive to transform patients into consumers by empowering them to make more proactive choices, patients would still be confronted by decisions each time they encounter the health care system – discerning how best to respond to their experiences, how best to guard against future problems and how best to induce practitioners to respond to their needs. These choices inhere in every episode of medical care; though they take on a somewhat different character in light of policies that are designed to encourage more active patient selection among physicians, hospitals or health plans.
Social and behavioral scientists have a term for choices that are made under difficult circumstances such as those found in health care settings. Following Herbert Simon they refer to them as ‘boundedly rational’ (Simon, Reference Simon1957, Reference Simon1982). ‘Rational’ in the sense that these choices are made in a purposeful manner; ‘bounded’ in the sense that the complexity of choice makes it impossible for decision-makers to collect or make sense of all the information that they would need to make fully informed judgements. Over the past quarter century, scholars have learned a lot about how choices are made by boundedly rational decision-makers. The cognitive constraints and decision heuristics that shape their decisions do so in predictable ways that ought to be taken into account by policymakers. Yet health policies in every country instead presume that patients will be magically endowed with information, capable of using it to make sense of their past health care and fully anticipate their future medical needs. These assumptions are horribly flawed, making policies predicated upon them very likely to go awry.
To explore these implications, I first lay out some aspects of bounded rationality that are most relevant to choices in health care contexts, then identify:
First, some crucial distinctions among:
• various pathways though which choices might leverage improved performance, and
• the forms of support that are offered to encourage patients’ engagement with choice.
And second, some ways in which choice-promoting policies inevitably interact with:
• other efforts to promote patient empowerment, and
• individuals’ efforts to coherently interpret their experiences with medical care.
My goal is to parse policymakers’ promotion of health care choices in light of the nature and limitations of individual decision-making. To be clear, my primary purpose here is not to assess whether choice-enhancing or choice-constraining policies make patients (collectively or individually) better off. Like the other papers in this special issue, this commentary is foremost focused on how choice affects health system performance overall. But arguably these broader implications can be fully anticipated only if policymakers pay attention to the ways in which patients actually engage in choices.
The literature on bounded rationality documents an array of with choice pathologies.Footnote 3 Yet ironically in health care – a domain where the circumstances of choice render these pathologies most frequent – policymakers and policy analysts persist in treating choice as if it were simply a matter of sufficient time, information and attention. In this review of choice-promoting policies, I will identify those most vulnerable to choice pathologies and consider how policy initiatives might be adapted to minimize either these risks or ameliorate the problematic consequences.
In other words, we are looking here for the choicest cuts of choice-based policies, seeking strategies that avoid what Flood and Haugan termed the ‘bitter’ parts while promoting those less fraught with psychological pitfalls (Flood and Haugan, Reference Flood and Haugan2010, this issue). To be sure, none of these refinements can so enhance consumer choice that it should ever be viewed as either the sole strategy – nor even the primarily mainstay – of efforts to improve health system performance. But they can increase the chances that choices – including those that inhere in every medical encounter – will be exercised in ways that more often benefit individual patients and guard against some of the dangers of simplistic market-oriented medical care.
In exploring these matters, I will draw heavily on programmatic experience and research findings from the US. Americans have the advantage – if it can rightfully be called that – of having started down the primrose path toward patient empowerment about 10–15 years ahead of Europeans. As a result, the US has invested considerably more for considerably longer in support of patient choice; it offers more complete evidence for assessing the prospects for choice-promoting policies, particularly the implications that emerge gradually over time, as patients learn and providers adapt to the opportunities and aspirations for choice-making.
These sorts of cross-national comparisons must be deployed judiciously. The potential and pathologies of choice in medical settings are powerfully shaped by the circumstances under which the choices are made. Americans’ health care choices are made in contexts that are more challenging than are the circumstances facing patients in Europe.Footnote 4 For example, there is a lot more turnover in insurance in the U.S. Compared to citizens in the Netherlands, Germany or Switzerland (all countries in which citizens choose among insurers), Americans are 4–7 times as likely to have recently changed their insurer.Footnote 5 Partly as a result, Americans also change clinicians more frequently. They less frequently identify a single clinician as a regular source of care than do Europeans. And three times as many Americans who do have a regular source of care have had that relationship for less than five years.Footnote 6 Whether this frequent switching renders Americans better or worse at choosing remains unclear – but one must guard against drawing overly facile lessons from cross-country comparisons and be mindful of these contextual differences.
Foundational perspectives: bounded rationality and health care choices
As Bevan et al. (Reference Bevan, Helderman and Wilsford2010, this issue) note, most choice-enhancing policies are predicated on a simple, but intuitive, model of decision-making. This presumes that (a) consumers have well defined expectations (i.e. they know what they want from their medical care or health insurance), (b) they can coherently evaluate their current health care experiences in light of these expectations and (c) if dissatisfied, will learn about alternatives to their current circumstances. But this turns out to be not all that accurate a description of how choices are actually made by patients.
The primary reason is that the world is a complex place, relative to our ability to make sense of it. In the lexicon of the decision sciences, we are all ‘boundedly rational’ – that is, we try to make the most sensible choices that we can, but are aware of our limited capacity to make them in a thoroughly informed way. Faced with these circumstances, people develop a set of decision-heuristics (rules of thumb) that allow them to make choices without devoting excessive time, attention or thought (Conlisk, Reference Conlisk1996).
The impact of heuristic reasoning has been amply documented for a variety of health-related choices, including the selection of clinicians and health insurers (Sage, Reference Sage2000; Hibbard et al., Reference Hibbard, Slovic, Peters and Finucane2002; Milstein and Adler, Reference Milstein and Adler2003; Hanoch and Rice, Reference Hanoch and Rice2006; Schneider, Reference Schneider2006), risk perceptions (Noll and Krier, Reference Noll and Krier1990), utilization of screening tests and exams (Rothman and Salovey, Reference Rothman and Salovey1997). How consumers cope with limited decision-capacity proves to be as consequential as the fact that they are boundedly rational. Several of these coping mechanisms are most policy relevant.
Cognitive complexity and mental accounts
People reduce the cognitive burdens of choice in part by categorizing choices into a limited number of ‘mental accounts’, each of which highlights particular aspects of the choices in question (Thaler, Reference Thaler1999). For example, if primary care clinicians are viewed essentially as personal caregivers, they will be chosen based on their compassion, empathy and interpersonal rapport. Conversely, if primary care doctors are viewed predominantly as coordinators of secondary and tertiary care, they will be evaluated more in terms of their technical skills and knowledge (Fung et al., Reference Fung, Elliot, Hays, Kahn, Kanouse, Glynn, Spranca and Shekelle2005).
The key insight from the study of bounded rationality is that the crucial stage in consumers’ choice of doctor really occurs one step prior: when they initially determine whether primary care is more appropriately seen as care-giving or service coordination. Once this happens, it simplifies the cognitive challenges of physician assessment by bringing to the forefront a subset of attributes that make for a ‘good’ clinician, while suppressing others. Most consumers choose among doctors on the basis of either humaneness or technical skills, not both (Fung et al., Reference Fung, Elliot, Hays, Kahn, Kanouse, Glynn, Spranca and Shekelle2005).
These findings are not simply academic curiosities. Policy initiatives can ‘prime’ (i.e. trigger into more active use) certain frames for thinking about health care choices. They may do so even when directed toward other ends. For example, policies such as the ‘soft gatekeeping’ described by Or et al. (Reference Or, Cases, Lisac, Vrangbæk and Ulrika Winblad Spångberg2010, this issue) are typically intended to influence patients’ choice among secondary providers, by engaging their primary care clinician to assist with these choices. Whether or not these policies work in their intended way, they also indirectly affect how patients think about primary care physicians, and thus affect what those patients look for in a primary care provider in the first place and whether they can make these assessments well.
Emotional heuristics and filters
When cognition is burdensome, people adapt by making choices less by thinking and more by feeling. Of course, these are neither mutually exclusive nor entirely separable. Psychologists have posited (and partially validated through experimental research) a variety of ‘dual processing’ models: decision-making processes in which individuals’ deploy some combination of cognitive assessment and emotional response to discern where they should focus their attention and how they should choose among complicated choice sets. My intent here is simply to highlight some of the ways in which emotions can shape health care choices. And to consider what policy implications follow from this.
The role of emotions is certainly relevant in medical settings, where the prospects of potentially serious illnesses and (at least in the US) potentially catastrophic medical expenses give many choices a potent emotional charge. Although this raises the spectre of rash and ill-considered judgments (Sunstein, Reference Sunstein2002; Loewenstein, Reference Loewenstein2004), research suggests that in fact people can deploy emotional heuristics in relatively sophisticated and nuanced ways, producing decisions among uncertain prospects that can be as congruent with their preferences as are more ‘thoughtful’ choices (Schwartz, Reference Schwartz2002; Slovic et al., Reference Slovic, Finucane, Peters and MacGregor2002). Nonetheless, decisions in which emotions are at the fore can embody certain ‘biases’ or predispositions relevant for choice-promoting health policies.
The first set stems from the emotional heuristics that influence health-related choices. Quite different emotions play a central role in directing people’s attention, compared to those that influence how they make choices. In a nutshell, attention is driven most powerfully by negative emotions: fear, anger and anxiety (Tiedens and Linton, Reference Tiedens and Linton2001). This means that there are few better ways of capturing the public’s attention than to have them feel somewhat threatened by future prospects. Fear appeals have become a central motif in many health promotion campaigns (since peoples’ lives are too busy to pay much attention to healthy lifestyles, unless they are scared into so doing) (Hale and Dillard, Reference Hale and Dillard1995; Witte and Allen, Reference Witte and Allen2000). Comparable fear-based appeals have yet to be deployed to call attention to choices among insurers or clinicians, but some public education initiatives – such as those warning against ‘medical errors’ – may indirectly induce heightened anxiety.
Conversely, how choices are made is influenced more by positive emotions (Isen and Labroo, Reference Isen and Labroo2003). In the political sphere, for example, voters often base their support for particular politicians or policies on the perceived ‘likability’ of those officials or the groups that would benefit from the policies in question (Sniderman et al., Reference Sniderman, Brody and Tetlock1991; Kahneman and Frederick, Reference Kahneman and Frederick2002). Whether similar emotional heuristics would enhance or detract from patients’ choices among clinicians depends in part on what they most value in physician–patient relationships. Emotional heuristics that help choosing among primary care physicians (e.g. signing up with the first doctor with whom one feels a sense of affinity) might prove less useful – and possibly even counterproductive – when selecting a surgeon. Moreover, seeking out a likable primary care clinician looks less sensible if their primary role is to be helpful selecting a surgeon: for that one wants considered judgment more than warm empathy.
In emotionally charged settings, boundedly rational consumers will also filter the information and experiences to which they pay attention. They do this to manage their anxiety about choices, particularly in situations that make them feel insecure. Several forms of filtering are likely to prove most consequential in health care settings.
The first involves selectively remembering past events or evaluating present circumstances in ways that inflate perceived self-efficacy in dealing with uncertain situations. Risks induce the greatest fear when they are seen as outside one’s personal control. The more efficacious one feels, the less anxiety is induced (Loewenstein et al., Reference Loewenstein, Weber, Hsee and Welch2001). The problem is that actually gaining control over health-related risks is typically quite difficult.
Fortunately for most patients’ emotional equilibrium (although not necessarily the clarity of their subsequent choices), it is far easier to create the illusion of self-efficacy than it is to actually make sound choices. Illusions can be maintained through some combination of (1) inflated optimism about one’s own life course, discounting risks for oneself that are known to apply to the general population (Armor and Taylor, Reference Armor and Taylor2002; Weinstein and Klein, Reference Weinstein and Klein1995), (2) selectively revising one’s remembered expectations about past choices, so that they are retrospectively made more consistent with realized outcomes (Conway and Ross, Reference Conway and Ross1984; Hawkins and Hastie, Reference Hawkins and Hastie1990; Van Dijk et al., Reference Van Dijk, Zeelenberg and van der Pligt2003) and (3) interpreting ambiguous evidence to make it more consistent with one’s prior expectations (Rabin and Schrag, Reference Rabin and Schrag1999). Exactly this sort of self-justifying filtering may explain why patients always have far more positive assessments of their own physicians than of doctors in general, since they feel responsible for having chosen the former (Jacobs and Shapiro, Reference Jacobs and Shapiro1994).
A second way of moderating anxiety is to ignore information that makes risks more salient. Many people turn out quite adept at this: not only will they not seek out information that heightens awareness of risk; they actively avoid that information, even if it is freely available to them. For example, studies indicate that 40%–60% of patients do not want information about genetic susceptibility to various cancers (Lerman et al., Reference Lerman, Narod, Schulman, Hughes, Gomez-caminero, Bonney, Gold, Trock, Main, Lynch, Fulmore, Snyder, Lemon, Conway, Tonin, Lenoir and Lynch1996, Reference Lerman, Hughes, Trock, Myers, Main, Bonney, Abbaszadegan, Harty, Franklin, Lynch and Lynch1999).Footnote 7 This emotional logic yields some ironic consequences: information avoidance is most pronounced among individuals who would benefit most from knowing it, because they face the greatest personal risks (and therefore have the most acute anxiety). Patients with a family history of particular health problems are less likely to seek testing for those problems (Kőszegi, Reference Köszegi2003).
Choice-oriented policies ought to take into account affective influences on boundedly rational decision-making. Optimal patient engagement may require framing choices to induce mild anxiety – enough to call attention to the choice, without reducing patients to a state of frenzied stasis. Realistically, promoting this sort of emotional equilibrium requires a sort of infrastructure in which patients might consider emotionally fraught choices in the company of some trusted third-party who might help them cope with the anxieties in a constructive manner. I explore more fully below how this might be facilitated.
The limits of decision heuristics: cognitive overload and processing constraints
Cognitive and emotional heuristics make choices easier. Emotional filters make it possible to contemplate uncertain prospects without becoming paralysed by anxiety. But under some circumstances (choice sets that are large and complex; choices that are particularly frightening) and for some individuals (those with limited cognitive capacity, self-efficacy or ‘emotional intelligence’), these coping mechanisms prove inadequate. When this happens, a sort of ‘cognitive overload’ occurs: consumers defer choice, even if the status quo is seemingly less acceptable than any of the choice options (Dhar and Nowlis, Reference Dhar and Nowlis2004). Experimental findings suggest that putting off choice happens more frequently when consumers have more alternatives to choose among (Iyengar and Lepper, Reference Iyengar and Lepper2000), when there is heightened potential for regret in the aftermath of a poor choice (Anderson, Reference Anderson2003), and when the alternatives do not appear that different in key attributes (Dhar et al., Reference Dhar, Nowlis and Sherman1999).
In less extreme cases of cognitive overload, decision makers take into account some options or aspects of performance, but neglect others. This truncated process is often simplistically mechanical: rather than triaging options and focusing on those most promising, consumers appear to just take-in information until the limits of their processing capacity are reached, at which point they stop. The order in which options are presented can thus have a large impact, since many consumers never get around to considering options lower on the list (Johnson et al., Reference Johnson, Hershey, Meszaros and Kunreuther1993; Vaiana and McGlynn, Reference Vaiana and McGlynn2002).
Policymakers interested in facilitating choice might therefore decide to limit choice options, to support patients’ choices in ways that reduce the risk of cognitive overload or to help consumers deal with the consequences of bad choices, reducing the potential for post-choice regret. Because not all patients have the same capacity to take on complex choices, policy initiatives could target additional support to patients most likely to be overwhelmed by their options and calm those for whom choices induce more intense anxiety. Though this may initially sound like an unrealistic aspiration for health policy, there exist a variety of recent initiatives with demonstrated capacity to help patients negotiate the health care system with reduced anxiety.
I explore the potential for doing this more fully in the remainder of the article, considering first some useful distinctions that can make choice-oriented policies themselves more responsive and adaptive. I then make the case that choice-promoting policies must be considered in the context of patients’ personal histories as well as in terms of how their impact intersects other empowerment policies.
Elucidating key distinctions relevant to health care related choices
Bounded rationality holds important lessons for (a) the various ways in which policymakers might expect expanded choice to benefit patients and society, and (b) the different forms of support that might be provided to patients as they make these choices.
Four ways in which the benefits of enhanced choice might be realized
Bevan et al. draw a useful distinction between favoring choice as a virtuous actFootnote 8 (the ‘evaluative reason’) and supporting choice as instrumental to other ends (the ‘effectiveness reason’). Bounded rationality – and American health policy experience coping with its implications – argues for further differentiating the evaluative and effectiveness motives.
Parsing the evaluative reason(s)
Apart from the promise of choice enhancing personal autonomy (Schneider, Reference Schneider2006; Jennings, Reference Jennings2007), policymakers have come to view choice as a sort of contagious virtue. Here’s their logic. As patients make choices among health care providers and insurers, they generate information that other consumers can then use to make their own choices easier. This is an intrinsic rather than instrumental purposes because it is justified in terms of its benefits in promoting choice, rather than (or in addition to) its implications for improving health system performance. Data on disenrollment from health plans and patient switching among doctors has been widely deployed in the US to inform subsequent consumer choice (Schlesinger et al., Reference Schlesinger, Druss and Thomas1999; Spranca et al., Reference Spranca, Eliot, Shaw and Kanouse2007). But neither intrinsic nor contagious virtues prove entirely compelling.
It is true that in some cultures, choice in medical settings is seen as fundamental to patients’ autonomy. Ninety-eight percent of Germans, Poles and Slovenians favor ‘free choice’ of primary care doctors – about as close to unanimity as public surveys ever get (Coulter and Jenkinson, Reference Coulter and Jenkinson2005: 359). It is also true that when people are given choice – even if they have not personally exercised it – they report greater satisfaction with insurers and providers, both in the US and in other countries (Davis et al., Reference Davis, Collins, Schoen and Morris1995; Kalda et al., Reference Kalda, Polluste and Lember2003). It is the choice that leads to the greater satisfaction – as demonstrated through randomized trials of choice (Hsu et al., Reference Hsu, Schmittdiel, Krupat, Stein, Thom, Fireman and Selby2003).
Yet neither bit of evidence is enough to treat choice as inherently virtuous in medical settings. Even in countries where choice is most embraced for primary care, enthusiasm is lower for choosing among specialists or hospitals (Coulter and Jenkinson, Reference Coulter and Jenkinson2005: 359). And support for choice is not always universal for primary care: in some countries, 15%–20% of the public sees little value in choice. The combination of cross-national and cross-service differences leads to a nadir among the Swedes, less than a third of whom favor choice among medical specialists (in other countries, support for choice among specialists is mixed, but generally embraced by 45%–60% of the public). With this checkered support, choice can hardly be necessary for personal autonomy in medical settings.
Even more consequentially, bounded rationality raises serious questions about whether choice actually makes people better off. Giving consumers a choice implicitly holds them responsible if that choice were poorly made or subsequent health care outcomes were bad (Weiner, Reference Weiner1995). Once consumers have made such a choice, their perceived need to feel self-efficacious encourages them to view subsequent medical experiences through rose-colored lenses, drawing overly positive assessments of these experiences. In other words, opportunities for choice are also opportunities for anxiety. To manage these fears requires the pretense of having chosen wisely. That pretense may in turn then distort future choices, so that even empowered consumers are not fully capable of protecting their own interests, rightly understood.
If patients’ perceived satisfaction is in part a distortion, rather than a reflection of their true experience, using these choices to guide other consumers also seems less attractive. If consumers have overly rosy assessments of past choices, and are overly deterred by the complexity of future choices, health care markets and quasi-markets will be characterized by substantial status quo biases – too few people who have been poorly treated will actually switch insurers or providers (Schlesinger et al., Reference Schlesinger, Druss and Thomas1999; Elbel and Schlesinger, Reference Elbel and Schlesinger2009). It is not even clear that the worst providers would generate the most switching. If the prospect of having made a poor choice evokes sufficient anxiety, it may discourage patients from seeking out the very information that would allow them to evaluate those choices, causing them to be less likely to seek out a new source of care or coverage.
Complicating matters further, patients who have the most extensive experiences in medical settings because they have complex, chronic conditions are least likely to switch either health plans or doctors, because they value of continuity of care most highly. Paradoxically, it is the relatively healthy who make most of the moves (Schlesinger et al., Reference Schlesinger, Druss and Thomas1999). Consequently, disenrollment rates from health plans and switching among physicians will disproportionately reflect the choices of patients who know least about whether their doctors or insurers really help those who are sick.
Parsing the effectiveness reason(s)
The instrumental rationale for choice also comes in two flavors. First, choice is expected to reshape the attitudes and behaviors of health care providers as they strive to attract and retain clientele. I’ll refer to this as ‘empowering’ patients. Second, choice is expected to induce patients to pay closer attention to medical matters, as they try to sort through their treatment options and clinicians. I’ll refer to this as ‘engaging’ patients.
Most of the promise and debate related to consumer choice in Europe is focused on the objectives of introducing new incentives by empowering consumers to make choices between hospitals in England and Scandinavia (Bevan and van de Ven, Reference Bevan and van de Ven2010, this issue; Or et al., Reference Or, Cases, Lisac, Vrangbæk and Ulrika Winblad Spångberg2010, this issue); between specialists in England, although this has proved problematic (Dixon et al., Reference Dixon, Robertson and Dixon2010, this issue); and between insurers in the Netherlands, Belgium, Germany and Switzerland (Bevan and van de Ven, Reference Bevan and van de Ven2010, this issue; van de Ven et al., Reference van de Ven, Beck, Van de Voorde, Wasem and Zmora2007). As the articles in this issue suggest, the track record in Europe has been more checkered than proponents of choice expected (Corrigan and Maynard, Reference Corrigan and Maynard2005; Kroneman et al., Reference Kroneman, Maarse and van der Zee2006; Grytten and Sørensen, Reference Grytten and Sørensen2009). Nor is the evidence much rosier in the US (Nichols et al., Reference Nichols2004; Tu and May, Reference Tu and May2007).
This modest impact on provider behavior is not entirely surprising, because relatively few patients ever switch providers or health plans, even if they experienced a plethora of problems (Elbel and Schlesinger, 2009). This reflects in part the status quo biases induced by bounded rationality, in part a desire to maintain continuity of care among patients with chronic health problems (Schlesinger et al., Reference Schlesinger, Druss and Thomas1999). Health plans with the most dissatisfied enrollees experienced little switching (3% of enrollees) and almost no change in the market share (under 1%) (Chernew et al., Reference Chernew, Gowrisankaran and Scanlon2001). Problematic experiences with general practitioners lead patients to lower their evaluation of quality, but do not generally increase their intensions to switch physicians (de Ruyter and Wetzels, Reference de Ruyter and Wetzels1998).Footnote 9
As efforts to empower medical consumerism have faltered, American policymakers have placed greater emphasis on consumer engagement.Footnote 10 To this end, they have made substantial investments collecting information on provider performance and patient experience (Darby et al., Reference Darby, Crofton and Clancy2006) and augmented this with financial incentives to encourage its use (to avoid out-of-pocket medical expenses). This has been most evident in the design of so-called ‘consumer-driven’ health plans,Footnote 11 which incorporate large deductibles as financial inducements for greater consumer engagement, and medical savings accounts, designed to make consumers more sensitive to the costs of generous coverage (Milstein and Adler, 2003; Robinson and Ginsburg, Reference Robinson and Ginsburg2009). In both cases, the notion is that if consumers pay for a visit to the doctor or an additional covered benefit out of their own pocket, then they will think more carefully about whether the additional expenditure is worth the cost. That additional scrutiny is thought to induce consumers to be more engaged in all their health care choices.
But here too, the impact has been disappointing. Enrollees in consumer-directed plans are no more likely than other Americans to seek out information about providers or to choose on the basis of that information (Frostin, Reference Frostin2008; Hibbard et al., Reference Hibbard, Greene and Tusler2008), even though these consumers were indeed more cost conscious, reducing their use of certain medical services (Frostin, Reference Frostin2008). Consumer engagement has proven surprisingly difficult to foster or incentivize in medical settings.
Three strategies for ‘Supporting’ health care choices
This American experience with market-oriented health care made clear the need for policies to support consumer choice. These supportive policies have taken three forms: (1) providing standardized information on the performance of clinicians and insurers, (2) engaging primary care physicians to assist their patients making choices among hospitals and specialists and (3) funding third-parties (e.g. patient navigators, help-lines) to help patients find relevant information and negotiate their way through the health care system.Footnote 12
All three strategies have been pursued in Europe as well as the US, though the papers in this issue focus primarily on the first and second.Footnote 13 This emphasis is not surprising; the third option has been given scant attention in the academic literature on health care choice. However, this omission is problematic, since arguably bounded rationality suggests that it is actually the third strategy that has the greatest potential for enhancing patient choice while minimizing its pathological effects.
Neither providing information nor engaging clinicians seems adequate to the task. Despite quite substantial investments in collecting and publishing data comparing health plans, doctors and hospitals in the US, consumers remain largely uninterested. Between 1996 and 2008, the percentage of Americans using comparative data on physicians, for example, grew from 4% to 6% (Kaiser Family Foundation, 2008). Although information on health plans and hospitals was seen by a larger share of the public, use of this information actually declined over this same period. Even those seriously contemplating change were deterred by the complexity of the information available (Kaiser Family Foundation, 2004).
To be sure, most people ignore this information because they would not even consider switching doctors or insurers; few are inclined toward this as recreational reading. Boundedly rational individuals simply do not have the time and attention, have an overly positive assessment of their current circumstances and have little desire to learn that their past choices were flawed. This lack of motivation is in itself part of the problem. The core paradox of medical consumerism is that it is hard for most patients to determine if their health care experiences are really problematic, even if the outcomes were disappointing. If they cannot discern this, they have little motive to seek out comparative performance data, even though this is precisely the information that might convince them that their experiences were not up to par (assuming that they would find these metrics coherent and salient).
Lack of motivation might be overcome if primary care clinicians reliably informed patients when their own care was inadequate. The American experience suggests that competing demands on physicians’ time, inadequate compensation for patient counseling and concerns about legal liability render this sort of physician assistance uneven at best, even in settings designed to facilitate choice (Bodenheimer, Reference Bodenheimer2008). But even if these logistical and motivational limitations could be overcome, it is not clear that physicians are appropriate for helping most patients make health care choices, particularly those involving high-stakes outcomes. Most patients will deploy an array of emotional heuristics and filters to ease choice. Clinicians are trained to be objectively dispassionate – possibly making them the least appropriate partners, since they are committed to a style of decision-making so at odds with how many patients approach their choices.
Recognizing these limitations, policymakers have engaged third-party ‘advocates’ (nurses, social workers and lay people with certificate-level training) to help patients make sense of information, choices, and health-care experiences.Footnote 14 These publicly funded infrastructures first emerged in the US to facilitate choices among insurers (McCormack et al., Reference McCormack, Schnaier, Lee and Garfinkel1996) but have in the past decade expanded to include ‘patient navigator’ programs easing patients’ transitions along the care continuum for chronic illnesses (Dohan and Schrag, Reference Dohan and Schrag2005). A comparable and in some ways more extensive infrastructure has emerged in the UK over the past ten years, in the form of the National Health Service (NHS)-Direct help-line and the Patient Advising and Liaison Service, though access to these supports remains checkered and their impact modest (Evans et al., Reference Evans, Powell and Cross2008).
The lack of academic attention to these third-party advisors is unfortunate since evidence is emerging that they improve patient choices and reduce confusion (O’Cathain et al., Reference O’Cathain, Goode, Luff, Strangleman, Hanlon and Greatbatch2005; Bowles et al., Reference Bowles, Tuzzio, Wiese, Kirlin, Greene, Clauser and Wagner2008). Most crucially, from the standpoint of bounded rationality, they have also provided the sort of emotional support required to reduce anxieties that may induce information avoidance and choice deferral (O’Cathain et al., Reference O’Cathain, Goode, Luff, Strangleman, Hanlon and Greatbatch2005; Ferrante et al., Reference Ferrante, Chen and Kim2008). However, despite this potentially crucial role for third-party advisors and advocates, we still know relatively little about the factors that shape their success.
Choice in context: mediating factors in the application of choice-promoting policies
The American experience suggests that the consequences of promoting choice will also depend on how choice-enhancing policies interact with other empowerment initiatives and with patients’ ability to learn from their medical experiences.
Two complementary strategies for empowering patients
Some choices in health care are made prospectively – for example, seeking a specialist for a newly emergent condition. But many are made in response to previous experiences, including those evoking patients’ ire or concern. The authors and editors of this issue recognize that when patients are so aggrieved, they might respond through exit or voice (Bevan et al., Reference Bevan, Helderman and Wilsford2010, this issue; Dixon et al., Reference Dixon, Robertson and Dixon2010, this issue). Consequently, the impact of policies promoting choice depends crucially on the prevailing level of patient voice and the policies designed to enhance it. But this interaction can be complex and subtle.
Recent evidence suggests that it is voice, rather than exit, that provides a more reliable marker for problematic patient experiences (Elbel and Schlesinger, Reference Elbel and Schlesinger2009). Repeated and severe problems with health care are more likely to induce complaints than they are to stimulate exit. Conditions expected to empower medical consumers are associated with increased voice in response to serious problems, but not increased exit. This suggests that if policymakers are truly concerned about making the health care system more responsive to patients’ concerns, they will emphasize arrangements that enhance voice. But what might one expect of policies that promote both exit and voice?
Health policy discourse typically portrays exit and voice as substitutes – flight vs fight (Annas, Reference Annas1997). The limited evidence from the US actually documents a positive relationship between voicing grievances and switching health plans (Schlesinger et al., Reference Schlesinger, Mitchell and Elbel2002), suggesting that perhaps the potential for exit can encourage voice. This might be the case if vocal patients risk alienating their doctor or insurer (or fear doing so), making it useful to have a handy escape route to avoid any subsequent retaliation.
Although the potential for exit might encourage voice, sometimes patients must choose between switching and complaining: at least which to do first. Bounded rationality also constrains patients’ ability to the choice between exit and voice, yet we know little about how it does so. Speculatively, one might suspect that the patients who were the most anxious (due to personality, frailty, limited personal resources or life-threatening conditions) would lean toward exit, since avoidance behavior is powerfully fear-driven. It may also be easier for disadvantaged groups to switch than to voice; the latter requires a better understanding of the presenting problem and greater capacity to articulate one’s concerns and expectations for remediation (Schlesinger et al., Reference Schlesinger, Mitchell and Elbel2002).
If this portrayal is accurate, it suggests the need for support systems that can enhance the capacity for voice among patients who are feeling most vulnerable. These parties might help patients cope with their anxieties to lessen the urge to flight. Or they might serve as advocates to reduce the personal burdens of voice. Either way, these concrete interventions would need to be guided by a thoughtful vision of the right balance between exit and voice, for health care generally as well as for particular services.
One crucial missing link: experience, learning and decision-making
Most policy analyses of health care choices implicitly treat them as one-off events. Certainly some of them are: few patients experience multiple polio vaccinations, tonsillectomies, hysterectomies or entries into hospice. Nonetheless, much health care involves repeat business. As people get older, they accumulate experiences with physicians and other caregivers; those who develop chronic conditions may have many encounters with the health care system each year. The capacity to make effective choices in these latter circumstances (which constitute the bulk of medical encounters in the US and European Union) depends crucially on consumers’ ability to learn from prior experience and draw appropriate inferences about their future needs and prospects.
Curiously, given all the effort and resources invested in an infrastructure to make informed choice possible in medical settings, policymakers have paid little attention to patients’ learning. The lessons of bounded rationality make this seem a troubling omission. To bolster their sense of self-confidence and self-efficacy in a domain of life fraught with uncertainty, people recall their past medical experiences that they have had a hand in choosing with an optimistic filter that renders their memories of treatment more positive than the actual experiences, which in turn makes patients too sanguine about the risks in future medical encounters (Armor and Taylor, Reference Armor and Taylor2002). Ironically, the more fraught with anxiety are the medical circumstances, the more distorting this optimistic bias is likely to be.
A second bias can sometimes cut in the opposite direction – it comes into play when people recall an extended series of past events, such as a sequence of treatments for a chronic condition. Their assessment of these experiences tends to be dominated by ‘peak’ episodes – discrete events that were particularly positive or negative (Gilbert et al., Reference Gilbert, Pinel, Wilson, Blumberg and Wheatley2002). Consequently, patients judge clinicians with whom they’ve had one memorably bad visit (even midst a long-run of positive interactions) more harshly than a clinician who is consistently ineffectual – even if the overall care provided by the latter is actually far worse.
Supporting effective consumer choice requires addressing these biases. Once again, patients’ primary physician could take on this role. But here again, clinicians may not be the most promising candidates. For therapeutic purposes, doctors are motivated to encourage optimistic biases, not dispel them. They may be even more reluctant to do so when past events involve problems with a patient’s medical care, which the physicians themselves may be loathe to acknowledge – either out of professional pride or fear of future legal ramifications.
Once more, this makes a clear case for a third-party infrastructure to facilitate choice (though this support system could also be based in clinicians’ offices, as in some conceptions of an expanded ‘medical home’). A focus on patients’ learning from their own medical experiences alters the expectations for advocacy and advice programs. Currently, these are designed primarily to answer discrete queries or respond to short-term crises; they lack the capacity to maintain continuity of relationships with patients for extended periods. Yet this is precisely what would be required to overcome biased learning. Some of the patient navigator experiments in the US do provide continuity, but only over the course of treatment for a specific medical condition (Dohan and Schrag, Reference Dohan and Schrag2005).
Conclusion
Our brief tour of the literature on bounded rationality suggests that health care choices ‘don’t come easy’. It is not easy to induce patients to seek information or engage in choice. It is not easy for those who are choosing to do so well – that is, in a manner that helps them to both understand and pursue their own best interests. Nor is it easy for policymakers, having embarked on choice-oriented strategies, to discern how well they are working or how best they can be refined.
All these challenges are exacerbated by the recent vintage of choice-enhancing health policies. But even in the US, with its more extensive track record, policymakers remain largely clueless about these initiatives. This was evident in the implementation of Part D of the Medicare program, covering prescription drugs for older and disabled Americans. Designed from the start to rely on consumer choice among private insurers, the number of choices proliferated (with government encouragement) so that each Medicare beneficiary was choosing among scores (in some cases hundreds) of insurers, making choice overload certain and enhancing the risks of status quo bias suppressing the very choice that these numerous insurers were intended to promote (Hanoch and Rice, Reference Hanoch and Rice2006; Thaler and Sunstein, Reference Thaler and Sunstein2008).
Similar shortcomings can emerge in subtler ways. For example, the patient choice policy in the NHS initially (2005) offered a choice of four or five secondary providers (Dixon et al., Reference Dixon, Robertson and Dixon2010, this issue). A choice set this small can be assessed by most consumers. And yet within three years, it had expanded to ‘any provider in England who met national standards and billed at the same rate as NHS facilities’. Surely, this seemed like a benign move to its advocates, at least to those enamored with the purported benefits of enhanced patient empowerment. Yet bounded rationality suggests otherwise: expanding the choice set likely undermined the capacity for coherent choice among at least some patients.
If policymakers understand how people actually make choices among uncertain prospects, they can encourage better choices and enhance patient security. Of course no set of choice-oriented policies, however, sensibly designed or thoughtfully implemented, can ever fully expunge choice pathologies. They cannot eliminate the disparate impact of these biases on patients with more limited cognitive capacities, personal resources, perceived self-efficacy or predisposition to anxiety. As a result, not all patients will be sufficiently supported to make choices consistent with their needs or preferences. Choice does not equate with either freedom or responsiveness. Nor to greater equity, if the most vulnerable patients are least able to choose in ways that can actually promote their own well being.
Policy aspirations need to be more circumspect and realistic. Choice is no panacea, but it can make a difference in terms of empowerment (particularly through its positive synergies with voice) for at least some patients under certain circumstances. For that reason, though it would be easy to become disillusioned when the impact of choice falls short of its proponents’ promises (as it surely will), it would be unfortunate if choice-oriented policies are too quickly abandoned. Policymakers instead should take seriously the challenges of supporting patients in their choices (including those that might encourage voice more than exit), a task that will require learning over time about how best to implement this infrastructure.
Providing an infrastructure that can advise patients’ in their choices, and potentially act as advocate on their behalf when they have difficulty putting choice into practice, would more reliably protect patients’ interests and render the health care system more responsive to their preferences. To be sure, patient advocate, navigators and help-lines have, to date, demonstrated only moderate benefits. They cannot be seen as panaceas to documented pathologies of choice, nor can choice-based policies be made the sole foundation of policies to enhance health system performance, even when augmented by this scaffolding, Yet pursued in this manner, choice-oriented policies can become a useful and constructive element of a portfolio of interventions through which government can enhance health system performance.
%20-%20https://sketch.com%20--%3e%3ctitle%3e3ECBBF50-FD08-412A-943D-2CFC5E025A16%3c/title%3e%3cdesc%3eCreated%20with%20sketchtool.%3c/desc%3e%3cg%20id='Stylesheet'%20stroke='none'%20stroke-width='1'%20fill='none'%20fill-rule='evenodd'%3e%3cg%20transform='translate(-926.000000,%20-668.000000)'%20id='BUTTONS'%3e%3cg%20transform='translate(100.000000,%20597.000000)'%3e%3cg%20id='BTN-/-RIGHTS'%20transform='translate(816.000000,%2065.000000)'%3e%3cg%20id='ICON-/-Rights'%20transform='translate(10.000000,%206.000000)'%3e%3cpath%20d='M13.5752841,0%20L1.34943182,0%20C0.59375,0%200,0.586666667%200,1.33333333%20L0,22.6666667%20C0,23.4133333%200.620738636,24%201.34943182,24%20L17.6505682,24%20C18.40625,24%2019,23.3866667%2019,22.6666667%20L19,5.36%20L13.5752841,0%20Z%20M13.5752841,1.89333333%20L17.0838068,5.36%20L13.5752841,5.36%20L13.5752841,1.89333333%20Z%20M1.34943182,22.6666667%20L1.34943182,1.33333333%20L12.1988636,1.33333333%20L12.1988636,5.36%20C12.1988636,6.10666667%2012.7926136,6.69333333%2013.5482955,6.69333333%20L17.6235795,6.69333333%20L17.6505682,22.6666667%20L1.34943182,22.6666667%20Z'%20id='Shape'%20fill='%23595959'%20fill-rule='nonzero'%3e%3c/path%3e%3ccircle%20id='Oval'%20stroke='%23595959'%20stroke-width='1.5'%20fill='%23595959'%20cx='9.5'%20cy='13.5355339'%20r='5.5'%3e%3c/circle%3e%3cpolyline%20id='Path-2'%20stroke='%23FFFFFF'%20stroke-width='1.5'%20transform='translate(9.500000,%2013.035534)%20rotate(45.000000)%20translate(-9.500000,%20-13.035534)%20'%20points='8%2016.0355339%2011%2016.0355339%2011%2010.0355339'%3e%3c/polyline%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
Some terms come redolent with positive associations. This is no less true in the policy lexicon of democracy (freedom, responsiveness, leadership and voting) than in everyday English (hope, friend, kindness and chocolate).Footnote 1 ‘Choice’ makes both lists. Its connotations reach to the heart and soul of every liberal polity, evoking the virtues of autonomy, self-determination, commitment and individual responsibility. Its emotional overtones are equally affirmative: to be ‘choice’ is not simply to be good, but to be the very best – the cream of the crop.
It is therefore not surprising that ‘choice’ has become a watchword for contemporary health care reformersFootnote 2 and remains a centerpiece of the health policy agenda whichever party controls the political agenda (Bevan and van de Ven, Reference Bevan and van de Ven2010, this issue; Paulus et al., Reference Paulus2003; Klein, Reference Klein2007). Choice among providers and health plans also seems consistent with the aspirations to patient autonomy that have dominated bioethics over the past 25 years (Schneider, Reference Schneider2006; Jennings, Reference Jennings2007). Choice in medical settings constructs a policy bridge to other socially valued domains (education, pensions and housing) in which the choice motif has become equally central to neo-liberal policies.
Promoting ‘choice’ also seems an attractive way for policymakers to tackle constructively the complex challenges of improving health system performance. Though assigned different labels – failures of ‘care coordination’ in France and Germany (Or et al., Reference Or, Cases, Lisac, Vrangbæk and Ulrika Winblad Spångberg2010, this issue), patients who do not attend their hospital appointments in the UK (Dixon et al., Reference Dixon, Robertson and Dixon2010, this issue) – patients’ disconnection from the health care system represents a persisting and pervasive problem in every country (Schoen et al., Reference Schoen, Osborn, Huynhm, Doty, Zapert, Peugh and Davis2005, Reference Schoen, Osborn, Doty, Bishop, Peugh and Murukutla2007). Enhanced choice seemingly tackles disconnectedness from both ends: first, by encouraging patients to become more actively engaged in the process of obtaining medical care, and second, by inducing clinicians to be more responsive to their patients’ preferences in these regards.
Perhaps most crucially, choice enhancing policies have – from policymakers’ perspective – the paramount virtue of simplicity: they seem a straightforward policy that benefits everyone. Just like taking a daily vitamin, going to yoga or ingesting a baby aspirin to reduce cardiovascular risks. After all, choice comes naturally in a consumerist society. Choice-promoting policies thus ‘normalize’ health care, making it seem more like other commodities that are familiar purchases (Bevan et al., Reference Bevan, Helderman and Wilsford2010, this issue).
Of course, choice in health care is not that simplistically virtuous. The authors and editors of this special issue recognize this, exploring some key complexities: (1) the importance of legal entitlements and prohibitions in shaping consumers’ choice sets (Bevan et al., Reference Bevan, Helderman and Wilsford2010, this issue; Flood and Haugan, Reference Flood and Haugan2010, this issue), (2) the differences between choosing among providers or insurers, and the very different policy approaches required to enhance each (Bevan and van de Ven, Reference Bevan and van de Ven2010, this issue; Dixon et al., Reference Dixon, Robertson and Dixon2010, this issue), (3) the ways in which institutional contexts and policy legacies alter expectations regarding choice and the practices through which it can be encouraged (Or et al., Reference Or, Cases, Lisac, Vrangbæk and Ulrika Winblad Spångberg2010, this issue; Flood and Haugan, Reference Flood and Haugan2010, this issue) and (4) the ways in which payment incentives and practice arrangements mediate (and often impede) the capacity of clinicians to support patient choice (Dixon et al., Reference Dixon, Robertson and Dixon2010, this issue).
These are all important insights. In this commentary, however, I propose to redirect our attention from the system to the psyche, by exploring how health care choices are experienced by individual patients. Though choice-promoting policies may simplify demands on public officials, choice itself is never simple from the patient’s perspective. The stakes are too high, making the choices fraught with anxiety. The consequences are impossible to fully anticipate, rendering judgments at best semi-informed guesses. And the circumstances are never really amenable: either the patients are healthy – and therefore can not be bothered to dwell much on medical matters – or they are sick, which often limits their capacity to process a lot of complicated information.
Yet difficult as choice may be in health care settings, its challenges cannot be circumvented. Even if policymakers did not strive to transform patients into consumers by empowering them to make more proactive choices, patients would still be confronted by decisions each time they encounter the health care system – discerning how best to respond to their experiences, how best to guard against future problems and how best to induce practitioners to respond to their needs. These choices inhere in every episode of medical care; though they take on a somewhat different character in light of policies that are designed to encourage more active patient selection among physicians, hospitals or health plans.
Social and behavioral scientists have a term for choices that are made under difficult circumstances such as those found in health care settings. Following Herbert Simon they refer to them as ‘boundedly rational’ (Simon, Reference Simon1957, Reference Simon1982). ‘Rational’ in the sense that these choices are made in a purposeful manner; ‘bounded’ in the sense that the complexity of choice makes it impossible for decision-makers to collect or make sense of all the information that they would need to make fully informed judgements. Over the past quarter century, scholars have learned a lot about how choices are made by boundedly rational decision-makers. The cognitive constraints and decision heuristics that shape their decisions do so in predictable ways that ought to be taken into account by policymakers. Yet health policies in every country instead presume that patients will be magically endowed with information, capable of using it to make sense of their past health care and fully anticipate their future medical needs. These assumptions are horribly flawed, making policies predicated upon them very likely to go awry.
To explore these implications, I first lay out some aspects of bounded rationality that are most relevant to choices in health care contexts, then identify:
First, some crucial distinctions among:
• various pathways though which choices might leverage improved performance, and
• the forms of support that are offered to encourage patients’ engagement with choice.
And second, some ways in which choice-promoting policies inevitably interact with:
• other efforts to promote patient empowerment, and
• individuals’ efforts to coherently interpret their experiences with medical care.
My goal is to parse policymakers’ promotion of health care choices in light of the nature and limitations of individual decision-making. To be clear, my primary purpose here is not to assess whether choice-enhancing or choice-constraining policies make patients (collectively or individually) better off. Like the other papers in this special issue, this commentary is foremost focused on how choice affects health system performance overall. But arguably these broader implications can be fully anticipated only if policymakers pay attention to the ways in which patients actually engage in choices.
The literature on bounded rationality documents an array of with choice pathologies.Footnote 3 Yet ironically in health care – a domain where the circumstances of choice render these pathologies most frequent – policymakers and policy analysts persist in treating choice as if it were simply a matter of sufficient time, information and attention. In this review of choice-promoting policies, I will identify those most vulnerable to choice pathologies and consider how policy initiatives might be adapted to minimize either these risks or ameliorate the problematic consequences.
In other words, we are looking here for the choicest cuts of choice-based policies, seeking strategies that avoid what Flood and Haugan termed the ‘bitter’ parts while promoting those less fraught with psychological pitfalls (Flood and Haugan, Reference Flood and Haugan2010, this issue). To be sure, none of these refinements can so enhance consumer choice that it should ever be viewed as either the sole strategy – nor even the primarily mainstay – of efforts to improve health system performance. But they can increase the chances that choices – including those that inhere in every medical encounter – will be exercised in ways that more often benefit individual patients and guard against some of the dangers of simplistic market-oriented medical care.
In exploring these matters, I will draw heavily on programmatic experience and research findings from the US. Americans have the advantage – if it can rightfully be called that – of having started down the primrose path toward patient empowerment about 10–15 years ahead of Europeans. As a result, the US has invested considerably more for considerably longer in support of patient choice; it offers more complete evidence for assessing the prospects for choice-promoting policies, particularly the implications that emerge gradually over time, as patients learn and providers adapt to the opportunities and aspirations for choice-making.
These sorts of cross-national comparisons must be deployed judiciously. The potential and pathologies of choice in medical settings are powerfully shaped by the circumstances under which the choices are made. Americans’ health care choices are made in contexts that are more challenging than are the circumstances facing patients in Europe.Footnote 4 For example, there is a lot more turnover in insurance in the U.S. Compared to citizens in the Netherlands, Germany or Switzerland (all countries in which citizens choose among insurers), Americans are 4–7 times as likely to have recently changed their insurer.Footnote 5 Partly as a result, Americans also change clinicians more frequently. They less frequently identify a single clinician as a regular source of care than do Europeans. And three times as many Americans who do have a regular source of care have had that relationship for less than five years.Footnote 6 Whether this frequent switching renders Americans better or worse at choosing remains unclear – but one must guard against drawing overly facile lessons from cross-country comparisons and be mindful of these contextual differences.
Foundational perspectives: bounded rationality and health care choices
As Bevan et al. (Reference Bevan, Helderman and Wilsford2010, this issue) note, most choice-enhancing policies are predicated on a simple, but intuitive, model of decision-making. This presumes that (a) consumers have well defined expectations (i.e. they know what they want from their medical care or health insurance), (b) they can coherently evaluate their current health care experiences in light of these expectations and (c) if dissatisfied, will learn about alternatives to their current circumstances. But this turns out to be not all that accurate a description of how choices are actually made by patients.
The primary reason is that the world is a complex place, relative to our ability to make sense of it. In the lexicon of the decision sciences, we are all ‘boundedly rational’ – that is, we try to make the most sensible choices that we can, but are aware of our limited capacity to make them in a thoroughly informed way. Faced with these circumstances, people develop a set of decision-heuristics (rules of thumb) that allow them to make choices without devoting excessive time, attention or thought (Conlisk, Reference Conlisk1996).
The impact of heuristic reasoning has been amply documented for a variety of health-related choices, including the selection of clinicians and health insurers (Sage, Reference Sage2000; Hibbard et al., Reference Hibbard, Slovic, Peters and Finucane2002; Milstein and Adler, Reference Milstein and Adler2003; Hanoch and Rice, Reference Hanoch and Rice2006; Schneider, Reference Schneider2006), risk perceptions (Noll and Krier, Reference Noll and Krier1990), utilization of screening tests and exams (Rothman and Salovey, Reference Rothman and Salovey1997). How consumers cope with limited decision-capacity proves to be as consequential as the fact that they are boundedly rational. Several of these coping mechanisms are most policy relevant.
Cognitive complexity and mental accounts
People reduce the cognitive burdens of choice in part by categorizing choices into a limited number of ‘mental accounts’, each of which highlights particular aspects of the choices in question (Thaler, Reference Thaler1999). For example, if primary care clinicians are viewed essentially as personal caregivers, they will be chosen based on their compassion, empathy and interpersonal rapport. Conversely, if primary care doctors are viewed predominantly as coordinators of secondary and tertiary care, they will be evaluated more in terms of their technical skills and knowledge (Fung et al., Reference Fung, Elliot, Hays, Kahn, Kanouse, Glynn, Spranca and Shekelle2005).
The key insight from the study of bounded rationality is that the crucial stage in consumers’ choice of doctor really occurs one step prior: when they initially determine whether primary care is more appropriately seen as care-giving or service coordination. Once this happens, it simplifies the cognitive challenges of physician assessment by bringing to the forefront a subset of attributes that make for a ‘good’ clinician, while suppressing others. Most consumers choose among doctors on the basis of either humaneness or technical skills, not both (Fung et al., Reference Fung, Elliot, Hays, Kahn, Kanouse, Glynn, Spranca and Shekelle2005).
These findings are not simply academic curiosities. Policy initiatives can ‘prime’ (i.e. trigger into more active use) certain frames for thinking about health care choices. They may do so even when directed toward other ends. For example, policies such as the ‘soft gatekeeping’ described by Or et al. (Reference Or, Cases, Lisac, Vrangbæk and Ulrika Winblad Spångberg2010, this issue) are typically intended to influence patients’ choice among secondary providers, by engaging their primary care clinician to assist with these choices. Whether or not these policies work in their intended way, they also indirectly affect how patients think about primary care physicians, and thus affect what those patients look for in a primary care provider in the first place and whether they can make these assessments well.
Emotional heuristics and filters
When cognition is burdensome, people adapt by making choices less by thinking and more by feeling. Of course, these are neither mutually exclusive nor entirely separable. Psychologists have posited (and partially validated through experimental research) a variety of ‘dual processing’ models: decision-making processes in which individuals’ deploy some combination of cognitive assessment and emotional response to discern where they should focus their attention and how they should choose among complicated choice sets. My intent here is simply to highlight some of the ways in which emotions can shape health care choices. And to consider what policy implications follow from this.
The role of emotions is certainly relevant in medical settings, where the prospects of potentially serious illnesses and (at least in the US) potentially catastrophic medical expenses give many choices a potent emotional charge. Although this raises the spectre of rash and ill-considered judgments (Sunstein, Reference Sunstein2002; Loewenstein, Reference Loewenstein2004), research suggests that in fact people can deploy emotional heuristics in relatively sophisticated and nuanced ways, producing decisions among uncertain prospects that can be as congruent with their preferences as are more ‘thoughtful’ choices (Schwartz, Reference Schwartz2002; Slovic et al., Reference Slovic, Finucane, Peters and MacGregor2002). Nonetheless, decisions in which emotions are at the fore can embody certain ‘biases’ or predispositions relevant for choice-promoting health policies.
The first set stems from the emotional heuristics that influence health-related choices. Quite different emotions play a central role in directing people’s attention, compared to those that influence how they make choices. In a nutshell, attention is driven most powerfully by negative emotions: fear, anger and anxiety (Tiedens and Linton, Reference Tiedens and Linton2001). This means that there are few better ways of capturing the public’s attention than to have them feel somewhat threatened by future prospects. Fear appeals have become a central motif in many health promotion campaigns (since peoples’ lives are too busy to pay much attention to healthy lifestyles, unless they are scared into so doing) (Hale and Dillard, Reference Hale and Dillard1995; Witte and Allen, Reference Witte and Allen2000). Comparable fear-based appeals have yet to be deployed to call attention to choices among insurers or clinicians, but some public education initiatives – such as those warning against ‘medical errors’ – may indirectly induce heightened anxiety.
Conversely, how choices are made is influenced more by positive emotions (Isen and Labroo, Reference Isen and Labroo2003). In the political sphere, for example, voters often base their support for particular politicians or policies on the perceived ‘likability’ of those officials or the groups that would benefit from the policies in question (Sniderman et al., Reference Sniderman, Brody and Tetlock1991; Kahneman and Frederick, Reference Kahneman and Frederick2002). Whether similar emotional heuristics would enhance or detract from patients’ choices among clinicians depends in part on what they most value in physician–patient relationships. Emotional heuristics that help choosing among primary care physicians (e.g. signing up with the first doctor with whom one feels a sense of affinity) might prove less useful – and possibly even counterproductive – when selecting a surgeon. Moreover, seeking out a likable primary care clinician looks less sensible if their primary role is to be helpful selecting a surgeon: for that one wants considered judgment more than warm empathy.
In emotionally charged settings, boundedly rational consumers will also filter the information and experiences to which they pay attention. They do this to manage their anxiety about choices, particularly in situations that make them feel insecure. Several forms of filtering are likely to prove most consequential in health care settings.
The first involves selectively remembering past events or evaluating present circumstances in ways that inflate perceived self-efficacy in dealing with uncertain situations. Risks induce the greatest fear when they are seen as outside one’s personal control. The more efficacious one feels, the less anxiety is induced (Loewenstein et al., Reference Loewenstein, Weber, Hsee and Welch2001). The problem is that actually gaining control over health-related risks is typically quite difficult.
Fortunately for most patients’ emotional equilibrium (although not necessarily the clarity of their subsequent choices), it is far easier to create the illusion of self-efficacy than it is to actually make sound choices. Illusions can be maintained through some combination of (1) inflated optimism about one’s own life course, discounting risks for oneself that are known to apply to the general population (Armor and Taylor, Reference Armor and Taylor2002; Weinstein and Klein, Reference Weinstein and Klein1995), (2) selectively revising one’s remembered expectations about past choices, so that they are retrospectively made more consistent with realized outcomes (Conway and Ross, Reference Conway and Ross1984; Hawkins and Hastie, Reference Hawkins and Hastie1990; Van Dijk et al., Reference Van Dijk, Zeelenberg and van der Pligt2003) and (3) interpreting ambiguous evidence to make it more consistent with one’s prior expectations (Rabin and Schrag, Reference Rabin and Schrag1999). Exactly this sort of self-justifying filtering may explain why patients always have far more positive assessments of their own physicians than of doctors in general, since they feel responsible for having chosen the former (Jacobs and Shapiro, Reference Jacobs and Shapiro1994).
A second way of moderating anxiety is to ignore information that makes risks more salient. Many people turn out quite adept at this: not only will they not seek out information that heightens awareness of risk; they actively avoid that information, even if it is freely available to them. For example, studies indicate that 40%–60% of patients do not want information about genetic susceptibility to various cancers (Lerman et al., Reference Lerman, Narod, Schulman, Hughes, Gomez-caminero, Bonney, Gold, Trock, Main, Lynch, Fulmore, Snyder, Lemon, Conway, Tonin, Lenoir and Lynch1996, Reference Lerman, Hughes, Trock, Myers, Main, Bonney, Abbaszadegan, Harty, Franklin, Lynch and Lynch1999).Footnote 7 This emotional logic yields some ironic consequences: information avoidance is most pronounced among individuals who would benefit most from knowing it, because they face the greatest personal risks (and therefore have the most acute anxiety). Patients with a family history of particular health problems are less likely to seek testing for those problems (Kőszegi, Reference Köszegi2003).
Choice-oriented policies ought to take into account affective influences on boundedly rational decision-making. Optimal patient engagement may require framing choices to induce mild anxiety – enough to call attention to the choice, without reducing patients to a state of frenzied stasis. Realistically, promoting this sort of emotional equilibrium requires a sort of infrastructure in which patients might consider emotionally fraught choices in the company of some trusted third-party who might help them cope with the anxieties in a constructive manner. I explore more fully below how this might be facilitated.
The limits of decision heuristics: cognitive overload and processing constraints
Cognitive and emotional heuristics make choices easier. Emotional filters make it possible to contemplate uncertain prospects without becoming paralysed by anxiety. But under some circumstances (choice sets that are large and complex; choices that are particularly frightening) and for some individuals (those with limited cognitive capacity, self-efficacy or ‘emotional intelligence’), these coping mechanisms prove inadequate. When this happens, a sort of ‘cognitive overload’ occurs: consumers defer choice, even if the status quo is seemingly less acceptable than any of the choice options (Dhar and Nowlis, Reference Dhar and Nowlis2004). Experimental findings suggest that putting off choice happens more frequently when consumers have more alternatives to choose among (Iyengar and Lepper, Reference Iyengar and Lepper2000), when there is heightened potential for regret in the aftermath of a poor choice (Anderson, Reference Anderson2003), and when the alternatives do not appear that different in key attributes (Dhar et al., Reference Dhar, Nowlis and Sherman1999).
In less extreme cases of cognitive overload, decision makers take into account some options or aspects of performance, but neglect others. This truncated process is often simplistically mechanical: rather than triaging options and focusing on those most promising, consumers appear to just take-in information until the limits of their processing capacity are reached, at which point they stop. The order in which options are presented can thus have a large impact, since many consumers never get around to considering options lower on the list (Johnson et al., Reference Johnson, Hershey, Meszaros and Kunreuther1993; Vaiana and McGlynn, Reference Vaiana and McGlynn2002).
Policymakers interested in facilitating choice might therefore decide to limit choice options, to support patients’ choices in ways that reduce the risk of cognitive overload or to help consumers deal with the consequences of bad choices, reducing the potential for post-choice regret. Because not all patients have the same capacity to take on complex choices, policy initiatives could target additional support to patients most likely to be overwhelmed by their options and calm those for whom choices induce more intense anxiety. Though this may initially sound like an unrealistic aspiration for health policy, there exist a variety of recent initiatives with demonstrated capacity to help patients negotiate the health care system with reduced anxiety.
I explore the potential for doing this more fully in the remainder of the article, considering first some useful distinctions that can make choice-oriented policies themselves more responsive and adaptive. I then make the case that choice-promoting policies must be considered in the context of patients’ personal histories as well as in terms of how their impact intersects other empowerment policies.
Elucidating key distinctions relevant to health care related choices
Bounded rationality holds important lessons for (a) the various ways in which policymakers might expect expanded choice to benefit patients and society, and (b) the different forms of support that might be provided to patients as they make these choices.
Four ways in which the benefits of enhanced choice might be realized
Bevan et al. draw a useful distinction between favoring choice as a virtuous actFootnote 8 (the ‘evaluative reason’) and supporting choice as instrumental to other ends (the ‘effectiveness reason’). Bounded rationality – and American health policy experience coping with its implications – argues for further differentiating the evaluative and effectiveness motives.
Parsing the evaluative reason(s)
Apart from the promise of choice enhancing personal autonomy (Schneider, Reference Schneider2006; Jennings, Reference Jennings2007), policymakers have come to view choice as a sort of contagious virtue. Here’s their logic. As patients make choices among health care providers and insurers, they generate information that other consumers can then use to make their own choices easier. This is an intrinsic rather than instrumental purposes because it is justified in terms of its benefits in promoting choice, rather than (or in addition to) its implications for improving health system performance. Data on disenrollment from health plans and patient switching among doctors has been widely deployed in the US to inform subsequent consumer choice (Schlesinger et al., Reference Schlesinger, Druss and Thomas1999; Spranca et al., Reference Spranca, Eliot, Shaw and Kanouse2007). But neither intrinsic nor contagious virtues prove entirely compelling.
It is true that in some cultures, choice in medical settings is seen as fundamental to patients’ autonomy. Ninety-eight percent of Germans, Poles and Slovenians favor ‘free choice’ of primary care doctors – about as close to unanimity as public surveys ever get (Coulter and Jenkinson, Reference Coulter and Jenkinson2005: 359). It is also true that when people are given choice – even if they have not personally exercised it – they report greater satisfaction with insurers and providers, both in the US and in other countries (Davis et al., Reference Davis, Collins, Schoen and Morris1995; Kalda et al., Reference Kalda, Polluste and Lember2003). It is the choice that leads to the greater satisfaction – as demonstrated through randomized trials of choice (Hsu et al., Reference Hsu, Schmittdiel, Krupat, Stein, Thom, Fireman and Selby2003).
Yet neither bit of evidence is enough to treat choice as inherently virtuous in medical settings. Even in countries where choice is most embraced for primary care, enthusiasm is lower for choosing among specialists or hospitals (Coulter and Jenkinson, Reference Coulter and Jenkinson2005: 359). And support for choice is not always universal for primary care: in some countries, 15%–20% of the public sees little value in choice. The combination of cross-national and cross-service differences leads to a nadir among the Swedes, less than a third of whom favor choice among medical specialists (in other countries, support for choice among specialists is mixed, but generally embraced by 45%–60% of the public). With this checkered support, choice can hardly be necessary for personal autonomy in medical settings.
Even more consequentially, bounded rationality raises serious questions about whether choice actually makes people better off. Giving consumers a choice implicitly holds them responsible if that choice were poorly made or subsequent health care outcomes were bad (Weiner, Reference Weiner1995). Once consumers have made such a choice, their perceived need to feel self-efficacious encourages them to view subsequent medical experiences through rose-colored lenses, drawing overly positive assessments of these experiences. In other words, opportunities for choice are also opportunities for anxiety. To manage these fears requires the pretense of having chosen wisely. That pretense may in turn then distort future choices, so that even empowered consumers are not fully capable of protecting their own interests, rightly understood.
If patients’ perceived satisfaction is in part a distortion, rather than a reflection of their true experience, using these choices to guide other consumers also seems less attractive. If consumers have overly rosy assessments of past choices, and are overly deterred by the complexity of future choices, health care markets and quasi-markets will be characterized by substantial status quo biases – too few people who have been poorly treated will actually switch insurers or providers (Schlesinger et al., Reference Schlesinger, Druss and Thomas1999; Elbel and Schlesinger, Reference Elbel and Schlesinger2009). It is not even clear that the worst providers would generate the most switching. If the prospect of having made a poor choice evokes sufficient anxiety, it may discourage patients from seeking out the very information that would allow them to evaluate those choices, causing them to be less likely to seek out a new source of care or coverage.
Complicating matters further, patients who have the most extensive experiences in medical settings because they have complex, chronic conditions are least likely to switch either health plans or doctors, because they value of continuity of care most highly. Paradoxically, it is the relatively healthy who make most of the moves (Schlesinger et al., Reference Schlesinger, Druss and Thomas1999). Consequently, disenrollment rates from health plans and switching among physicians will disproportionately reflect the choices of patients who know least about whether their doctors or insurers really help those who are sick.
Parsing the effectiveness reason(s)
The instrumental rationale for choice also comes in two flavors. First, choice is expected to reshape the attitudes and behaviors of health care providers as they strive to attract and retain clientele. I’ll refer to this as ‘empowering’ patients. Second, choice is expected to induce patients to pay closer attention to medical matters, as they try to sort through their treatment options and clinicians. I’ll refer to this as ‘engaging’ patients.
Most of the promise and debate related to consumer choice in Europe is focused on the objectives of introducing new incentives by empowering consumers to make choices between hospitals in England and Scandinavia (Bevan and van de Ven, Reference Bevan and van de Ven2010, this issue; Or et al., Reference Or, Cases, Lisac, Vrangbæk and Ulrika Winblad Spångberg2010, this issue); between specialists in England, although this has proved problematic (Dixon et al., Reference Dixon, Robertson and Dixon2010, this issue); and between insurers in the Netherlands, Belgium, Germany and Switzerland (Bevan and van de Ven, Reference Bevan and van de Ven2010, this issue; van de Ven et al., Reference van de Ven, Beck, Van de Voorde, Wasem and Zmora2007). As the articles in this issue suggest, the track record in Europe has been more checkered than proponents of choice expected (Corrigan and Maynard, Reference Corrigan and Maynard2005; Kroneman et al., Reference Kroneman, Maarse and van der Zee2006; Grytten and Sørensen, Reference Grytten and Sørensen2009). Nor is the evidence much rosier in the US (Nichols et al., Reference Nichols2004; Tu and May, Reference Tu and May2007).
This modest impact on provider behavior is not entirely surprising, because relatively few patients ever switch providers or health plans, even if they experienced a plethora of problems (Elbel and Schlesinger, 2009). This reflects in part the status quo biases induced by bounded rationality, in part a desire to maintain continuity of care among patients with chronic health problems (Schlesinger et al., Reference Schlesinger, Druss and Thomas1999). Health plans with the most dissatisfied enrollees experienced little switching (3% of enrollees) and almost no change in the market share (under 1%) (Chernew et al., Reference Chernew, Gowrisankaran and Scanlon2001). Problematic experiences with general practitioners lead patients to lower their evaluation of quality, but do not generally increase their intensions to switch physicians (de Ruyter and Wetzels, Reference de Ruyter and Wetzels1998).Footnote 9
As efforts to empower medical consumerism have faltered, American policymakers have placed greater emphasis on consumer engagement.Footnote 10 To this end, they have made substantial investments collecting information on provider performance and patient experience (Darby et al., Reference Darby, Crofton and Clancy2006) and augmented this with financial incentives to encourage its use (to avoid out-of-pocket medical expenses). This has been most evident in the design of so-called ‘consumer-driven’ health plans,Footnote 11 which incorporate large deductibles as financial inducements for greater consumer engagement, and medical savings accounts, designed to make consumers more sensitive to the costs of generous coverage (Milstein and Adler, 2003; Robinson and Ginsburg, Reference Robinson and Ginsburg2009). In both cases, the notion is that if consumers pay for a visit to the doctor or an additional covered benefit out of their own pocket, then they will think more carefully about whether the additional expenditure is worth the cost. That additional scrutiny is thought to induce consumers to be more engaged in all their health care choices.
But here too, the impact has been disappointing. Enrollees in consumer-directed plans are no more likely than other Americans to seek out information about providers or to choose on the basis of that information (Frostin, Reference Frostin2008; Hibbard et al., Reference Hibbard, Greene and Tusler2008), even though these consumers were indeed more cost conscious, reducing their use of certain medical services (Frostin, Reference Frostin2008). Consumer engagement has proven surprisingly difficult to foster or incentivize in medical settings.
Three strategies for ‘Supporting’ health care choices
This American experience with market-oriented health care made clear the need for policies to support consumer choice. These supportive policies have taken three forms: (1) providing standardized information on the performance of clinicians and insurers, (2) engaging primary care physicians to assist their patients making choices among hospitals and specialists and (3) funding third-parties (e.g. patient navigators, help-lines) to help patients find relevant information and negotiate their way through the health care system.Footnote 12
All three strategies have been pursued in Europe as well as the US, though the papers in this issue focus primarily on the first and second.Footnote 13 This emphasis is not surprising; the third option has been given scant attention in the academic literature on health care choice. However, this omission is problematic, since arguably bounded rationality suggests that it is actually the third strategy that has the greatest potential for enhancing patient choice while minimizing its pathological effects.
Neither providing information nor engaging clinicians seems adequate to the task. Despite quite substantial investments in collecting and publishing data comparing health plans, doctors and hospitals in the US, consumers remain largely uninterested. Between 1996 and 2008, the percentage of Americans using comparative data on physicians, for example, grew from 4% to 6% (Kaiser Family Foundation, 2008). Although information on health plans and hospitals was seen by a larger share of the public, use of this information actually declined over this same period. Even those seriously contemplating change were deterred by the complexity of the information available (Kaiser Family Foundation, 2004).
To be sure, most people ignore this information because they would not even consider switching doctors or insurers; few are inclined toward this as recreational reading. Boundedly rational individuals simply do not have the time and attention, have an overly positive assessment of their current circumstances and have little desire to learn that their past choices were flawed. This lack of motivation is in itself part of the problem. The core paradox of medical consumerism is that it is hard for most patients to determine if their health care experiences are really problematic, even if the outcomes were disappointing. If they cannot discern this, they have little motive to seek out comparative performance data, even though this is precisely the information that might convince them that their experiences were not up to par (assuming that they would find these metrics coherent and salient).
Lack of motivation might be overcome if primary care clinicians reliably informed patients when their own care was inadequate. The American experience suggests that competing demands on physicians’ time, inadequate compensation for patient counseling and concerns about legal liability render this sort of physician assistance uneven at best, even in settings designed to facilitate choice (Bodenheimer, Reference Bodenheimer2008). But even if these logistical and motivational limitations could be overcome, it is not clear that physicians are appropriate for helping most patients make health care choices, particularly those involving high-stakes outcomes. Most patients will deploy an array of emotional heuristics and filters to ease choice. Clinicians are trained to be objectively dispassionate – possibly making them the least appropriate partners, since they are committed to a style of decision-making so at odds with how many patients approach their choices.
Recognizing these limitations, policymakers have engaged third-party ‘advocates’ (nurses, social workers and lay people with certificate-level training) to help patients make sense of information, choices, and health-care experiences.Footnote 14 These publicly funded infrastructures first emerged in the US to facilitate choices among insurers (McCormack et al., Reference McCormack, Schnaier, Lee and Garfinkel1996) but have in the past decade expanded to include ‘patient navigator’ programs easing patients’ transitions along the care continuum for chronic illnesses (Dohan and Schrag, Reference Dohan and Schrag2005). A comparable and in some ways more extensive infrastructure has emerged in the UK over the past ten years, in the form of the National Health Service (NHS)-Direct help-line and the Patient Advising and Liaison Service, though access to these supports remains checkered and their impact modest (Evans et al., Reference Evans, Powell and Cross2008).
The lack of academic attention to these third-party advisors is unfortunate since evidence is emerging that they improve patient choices and reduce confusion (O’Cathain et al., Reference O’Cathain, Goode, Luff, Strangleman, Hanlon and Greatbatch2005; Bowles et al., Reference Bowles, Tuzzio, Wiese, Kirlin, Greene, Clauser and Wagner2008). Most crucially, from the standpoint of bounded rationality, they have also provided the sort of emotional support required to reduce anxieties that may induce information avoidance and choice deferral (O’Cathain et al., Reference O’Cathain, Goode, Luff, Strangleman, Hanlon and Greatbatch2005; Ferrante et al., Reference Ferrante, Chen and Kim2008). However, despite this potentially crucial role for third-party advisors and advocates, we still know relatively little about the factors that shape their success.
Choice in context: mediating factors in the application of choice-promoting policies
The American experience suggests that the consequences of promoting choice will also depend on how choice-enhancing policies interact with other empowerment initiatives and with patients’ ability to learn from their medical experiences.
Two complementary strategies for empowering patients
Some choices in health care are made prospectively – for example, seeking a specialist for a newly emergent condition. But many are made in response to previous experiences, including those evoking patients’ ire or concern. The authors and editors of this issue recognize that when patients are so aggrieved, they might respond through exit or voice (Bevan et al., Reference Bevan, Helderman and Wilsford2010, this issue; Dixon et al., Reference Dixon, Robertson and Dixon2010, this issue). Consequently, the impact of policies promoting choice depends crucially on the prevailing level of patient voice and the policies designed to enhance it. But this interaction can be complex and subtle.
Recent evidence suggests that it is voice, rather than exit, that provides a more reliable marker for problematic patient experiences (Elbel and Schlesinger, Reference Elbel and Schlesinger2009). Repeated and severe problems with health care are more likely to induce complaints than they are to stimulate exit. Conditions expected to empower medical consumers are associated with increased voice in response to serious problems, but not increased exit. This suggests that if policymakers are truly concerned about making the health care system more responsive to patients’ concerns, they will emphasize arrangements that enhance voice. But what might one expect of policies that promote both exit and voice?
Health policy discourse typically portrays exit and voice as substitutes – flight vs fight (Annas, Reference Annas1997). The limited evidence from the US actually documents a positive relationship between voicing grievances and switching health plans (Schlesinger et al., Reference Schlesinger, Mitchell and Elbel2002), suggesting that perhaps the potential for exit can encourage voice. This might be the case if vocal patients risk alienating their doctor or insurer (or fear doing so), making it useful to have a handy escape route to avoid any subsequent retaliation.
Although the potential for exit might encourage voice, sometimes patients must choose between switching and complaining: at least which to do first. Bounded rationality also constrains patients’ ability to the choice between exit and voice, yet we know little about how it does so. Speculatively, one might suspect that the patients who were the most anxious (due to personality, frailty, limited personal resources or life-threatening conditions) would lean toward exit, since avoidance behavior is powerfully fear-driven. It may also be easier for disadvantaged groups to switch than to voice; the latter requires a better understanding of the presenting problem and greater capacity to articulate one’s concerns and expectations for remediation (Schlesinger et al., Reference Schlesinger, Mitchell and Elbel2002).
If this portrayal is accurate, it suggests the need for support systems that can enhance the capacity for voice among patients who are feeling most vulnerable. These parties might help patients cope with their anxieties to lessen the urge to flight. Or they might serve as advocates to reduce the personal burdens of voice. Either way, these concrete interventions would need to be guided by a thoughtful vision of the right balance between exit and voice, for health care generally as well as for particular services.
One crucial missing link: experience, learning and decision-making
Most policy analyses of health care choices implicitly treat them as one-off events. Certainly some of them are: few patients experience multiple polio vaccinations, tonsillectomies, hysterectomies or entries into hospice. Nonetheless, much health care involves repeat business. As people get older, they accumulate experiences with physicians and other caregivers; those who develop chronic conditions may have many encounters with the health care system each year. The capacity to make effective choices in these latter circumstances (which constitute the bulk of medical encounters in the US and European Union) depends crucially on consumers’ ability to learn from prior experience and draw appropriate inferences about their future needs and prospects.
Curiously, given all the effort and resources invested in an infrastructure to make informed choice possible in medical settings, policymakers have paid little attention to patients’ learning. The lessons of bounded rationality make this seem a troubling omission. To bolster their sense of self-confidence and self-efficacy in a domain of life fraught with uncertainty, people recall their past medical experiences that they have had a hand in choosing with an optimistic filter that renders their memories of treatment more positive than the actual experiences, which in turn makes patients too sanguine about the risks in future medical encounters (Armor and Taylor, Reference Armor and Taylor2002). Ironically, the more fraught with anxiety are the medical circumstances, the more distorting this optimistic bias is likely to be.
A second bias can sometimes cut in the opposite direction – it comes into play when people recall an extended series of past events, such as a sequence of treatments for a chronic condition. Their assessment of these experiences tends to be dominated by ‘peak’ episodes – discrete events that were particularly positive or negative (Gilbert et al., Reference Gilbert, Pinel, Wilson, Blumberg and Wheatley2002). Consequently, patients judge clinicians with whom they’ve had one memorably bad visit (even midst a long-run of positive interactions) more harshly than a clinician who is consistently ineffectual – even if the overall care provided by the latter is actually far worse.
Supporting effective consumer choice requires addressing these biases. Once again, patients’ primary physician could take on this role. But here again, clinicians may not be the most promising candidates. For therapeutic purposes, doctors are motivated to encourage optimistic biases, not dispel them. They may be even more reluctant to do so when past events involve problems with a patient’s medical care, which the physicians themselves may be loathe to acknowledge – either out of professional pride or fear of future legal ramifications.
Once more, this makes a clear case for a third-party infrastructure to facilitate choice (though this support system could also be based in clinicians’ offices, as in some conceptions of an expanded ‘medical home’). A focus on patients’ learning from their own medical experiences alters the expectations for advocacy and advice programs. Currently, these are designed primarily to answer discrete queries or respond to short-term crises; they lack the capacity to maintain continuity of relationships with patients for extended periods. Yet this is precisely what would be required to overcome biased learning. Some of the patient navigator experiments in the US do provide continuity, but only over the course of treatment for a specific medical condition (Dohan and Schrag, Reference Dohan and Schrag2005).
Conclusion
Our brief tour of the literature on bounded rationality suggests that health care choices ‘don’t come easy’. It is not easy to induce patients to seek information or engage in choice. It is not easy for those who are choosing to do so well – that is, in a manner that helps them to both understand and pursue their own best interests. Nor is it easy for policymakers, having embarked on choice-oriented strategies, to discern how well they are working or how best they can be refined.
All these challenges are exacerbated by the recent vintage of choice-enhancing health policies. But even in the US, with its more extensive track record, policymakers remain largely clueless about these initiatives. This was evident in the implementation of Part D of the Medicare program, covering prescription drugs for older and disabled Americans. Designed from the start to rely on consumer choice among private insurers, the number of choices proliferated (with government encouragement) so that each Medicare beneficiary was choosing among scores (in some cases hundreds) of insurers, making choice overload certain and enhancing the risks of status quo bias suppressing the very choice that these numerous insurers were intended to promote (Hanoch and Rice, Reference Hanoch and Rice2006; Thaler and Sunstein, Reference Thaler and Sunstein2008).
Similar shortcomings can emerge in subtler ways. For example, the patient choice policy in the NHS initially (2005) offered a choice of four or five secondary providers (Dixon et al., Reference Dixon, Robertson and Dixon2010, this issue). A choice set this small can be assessed by most consumers. And yet within three years, it had expanded to ‘any provider in England who met national standards and billed at the same rate as NHS facilities’. Surely, this seemed like a benign move to its advocates, at least to those enamored with the purported benefits of enhanced patient empowerment. Yet bounded rationality suggests otherwise: expanding the choice set likely undermined the capacity for coherent choice among at least some patients.
If policymakers understand how people actually make choices among uncertain prospects, they can encourage better choices and enhance patient security. Of course no set of choice-oriented policies, however, sensibly designed or thoughtfully implemented, can ever fully expunge choice pathologies. They cannot eliminate the disparate impact of these biases on patients with more limited cognitive capacities, personal resources, perceived self-efficacy or predisposition to anxiety. As a result, not all patients will be sufficiently supported to make choices consistent with their needs or preferences. Choice does not equate with either freedom or responsiveness. Nor to greater equity, if the most vulnerable patients are least able to choose in ways that can actually promote their own well being.
Policy aspirations need to be more circumspect and realistic. Choice is no panacea, but it can make a difference in terms of empowerment (particularly through its positive synergies with voice) for at least some patients under certain circumstances. For that reason, though it would be easy to become disillusioned when the impact of choice falls short of its proponents’ promises (as it surely will), it would be unfortunate if choice-oriented policies are too quickly abandoned. Policymakers instead should take seriously the challenges of supporting patients in their choices (including those that might encourage voice more than exit), a task that will require learning over time about how best to implement this infrastructure.
Providing an infrastructure that can advise patients’ in their choices, and potentially act as advocate on their behalf when they have difficulty putting choice into practice, would more reliably protect patients’ interests and render the health care system more responsive to their preferences. To be sure, patient advocate, navigators and help-lines have, to date, demonstrated only moderate benefits. They cannot be seen as panaceas to documented pathologies of choice, nor can choice-based policies be made the sole foundation of policies to enhance health system performance, even when augmented by this scaffolding, Yet pursued in this manner, choice-oriented policies can become a useful and constructive element of a portfolio of interventions through which government can enhance health system performance.
Acknowledgements
Thanks for Brian Elbel for his continued collaboration and thinking on these matters of bounded rationality, as well as to helpful comments from Gwyn Bevan, Adam Oliver and Larry Brown.