The prevention puzzle

Promoting prevention is an important goal of public policy. Fifty years ago, Ehrlich and Becker (J Polit Econ 80:623-648, 1972) proposed a simple model of prevention (or self-protection as they called it). Surprisingly enough, subsequent research, mainly within the expected utility paradigm, showed that it is hard to derive clear predictions within this simple model that can help to guide policy. This is what I refer to as the prevention puzzle: why is it so hard for economic theory to guide prevention decisions? In this article I try to shed light on this question. I review the existing literature and add some tentative new results under nonexpected utility. While the impact of risk aversion on prevention is complex, three factors seem to contribute unambiguously to underprevention: prudence, likelihood insensitivity, and loss aversion. I conclude by giving some ideas how empirical research may contribute to the understanding of prevention decisions and help to solve the prevention puzzle.

Insurance decisions under nonperformance risk and ambiguity

An important societal problem is that people underinsure against risks that are unlikely or occur in the far future, such as natural disasters and long-term care needs. One explanation is that uncertainty about the risk of non-reimbursement induces ambiguity averse and risk prudent decision makers to take out less insurance. We set up an insurance experiment to test this explanation. Consistent with the theoretical predictions, we find that the demand for insurance is lower when the nonperformance risk is ambiguous than when it is known and when decision makers are risk prudent. We cannot attribute the lower take-up of insurance to our measure of ambiguity aversion, probably because ambiguity attitudes are richer than aversion alone.

SUPPLEMENTARY INFORMATION

The online version contains supplementary material available at 10.1007/s11166-021-09364-7.

A comparison of individual and collective decision making for standard gamble and time trade-off

Quality-Adjusted Life-Years (QALYs) are typically derived from individual preferences over health episodes. This paper reports the first experimental investigation into the effects of collective decision making on health valuations, using both time trade-off (TTO) and standard gamble (SG) tasks. We investigated collective decision making in dyads, by means of a mixed-subjects design where we control for learning effects. Our data suggest that collective decision making has little effect on decision quality, as no effects were observed on decision consistency and monotonicity for both methods. Furthermore, QALY weights remained similar between individual and collective decisions, and the typical difference in elicited weights between TTO and SG was not affected. These findings suggest that consulting with others has little effect on health state valuation, although learning may have. Additionally, our findings add to the literature of the effect of collective decision making, suggesting that no such effect occurs for TTO and SG.

Ambiguity preferences for health

In most medical decisions, probabilities are ambiguous and not objectively known. Empirical evidence suggests that people's preferences are affected by ambiguity. Health economic analyses generally ignore ambiguity preferences and assume that they are the same as preferences under risk. We show how health preferences can be measured under ambiguity, and we compare them with health preferences under risk. We assume a general ambiguity model that includes many of the ambiguity models that have been proposed in the literature. For health gains, ambiguity preferences and risk preferences were indeed the same. For health losses, they differed with subjects being more pessimistic in decision under ambiguity. Utility and loss aversion were the same for risk and ambiguity. Our results imply that reducing the clinical ambiguity of health losses has more impact than reducing the ambiguity of health gains, that utilities elicited with known probabilities may not carry over to an ambiguous setting, and that ambiguity aversion may impact value of information analyses if losses are involved. These findings are highly relevant for medical decision making, because most medical interventions involve losses.

Discounting health and money: New evidence using a more robust method

This study compares discounting for money and health in a field study. We applied the direct method, which measures discounting independent of utility, in a representative French sample, interviewed at home by professional interviewers. We found more discounting for money than for health. The median discount rates (6.5% for money and 2.2% for health) were close to market interest rates, suggesting that at the aggregate level the direct method solves the puzzle of unrealistically high discount rates typically observed in applied economics. Constant discounting fitted the data better than the hyperbolic discounting models that we considered. The substantial individual heterogeneity in discounting was correlated with age and occupation.

Treatment decisions under ambiguity

Many health risks are ambiguous in the sense that reliable and credible information about these risks is unavailable. In health economics, ambiguity is usually handled through sensitivity analysis, which implicitly assumes that people are neutral towards ambiguity. However, empirical evidence suggests that people are averse to ambiguity and react strongly to it. This paper studies the effects of ambiguity aversion on two classical medical decision problems. If there is ambiguity regarding the diagnosis of a patient, ambiguity aversion increases the decision maker's propensity to opt for treatment. On the other hand, in the case of ambiguity regarding the effects of treatment, ambiguity aversion leads to a reduction in the propensity to choose treatment.

A direct method for measuring discounting and QALYs more easily and reliably

Time discounting and quality of life are two important factors in evaluations of medical interventions. The measurement of these two factors is complicated because they interact. Existing methods either simply assume one factor given, based on heuristic assumptions, or invoke complicating extraneous factors, such as risk, that generate extra biases. The authors introduce a method for measuring discounting (and then quality of life) that involves no extraneous factors and that avoids distorting interactions. Their method is considerably simpler and more realistic for subjects than existing methods. It is entirely choice based and thus can be founded on economic rationality requirements. An experiment demonstrates the feasibility of this method and its advantages over classical methods.

An experimental test of the concentration index

The concentration index is widely used to measure income-related inequality in health. No insight exists, however, whether the concentration index connects with people's preferences about distributions of income and health and whether a reduction in the concentration index reflects an increase in social welfare. We explored this question by testing the central assumption underlying the concentration index and found that it was systematically violated. We also tested the validity of alternative health inequality measures that have been proposed in the literature. Our data showed that decreases in the spread of income and health were considered socially desirable, but decreases in the correlation between income and health not necessarily. Support for a condition implying that the inequality in the distribution of income and in the distribution of health can be considered separately was mixed.

How to Aggregate Health? Separability and the Effect of Framing

Background. Unweighted summation or quality-adjusted life year (QALY) utilitarianism is the most common way to aggregate health benefits in a cost-effectiveness analysis. A key qualitative principle underlying QALY utilitarianism is separability: those individuals unaffected by a policy choice should not influence the policy choice. Separability also underlies several of the alternatives for QALY utilitarianism that have been proposed. Objectives. To test separability and to test whether the support for separability is affected by the framing of the choice questions. Methods. In 2 experiments, 345 student subjects (162 in the first experiment, and 183 in the second experiment) were asked to select 1 of 2 possible treatments, with each treatment resulting in a different distribution of health across individuals. The only aspect that varied across choice questions was the state of the patients whose health was unaffected by the act of choosing a policy. In each experiment, we used 2 frames. In the implicit frame, it was implied but not plainly expressed what outcomes the treatments had in common. In the explicit frame, common outcomes of the 2 treatments were directly stated. The 2 experiments differed in the way the explicit frame was presented (verbal v. numerical). Results. The support for separability was significantly greater in the explicit frame. The proportion of violations in the implicit frame was 44% in Experiment 1 and 31% in Experiment 2, while in the explicit frame, the proportion of violations was 28% in Experiment 1 and 8% in Experiment 2. Conclusions. Framing affected the support for separability, raising issues as to whether it is possible to achieve a canonical representation of social choices.

The predictive validity of prospect theory versus expected utility in health utility measurement

Most health care evaluations today still assume expected utility even though the descriptive deficiencies of expected utility are well known. Prospect theory is the dominant descriptive alternative for expected utility. This paper tests whether prospect theory leads to better health evaluations than expected utility. The approach is purely descriptive: we explore how simple measurements together with prospect theory and expected utility predict choices and rankings between more complex stimuli. For decisions involving risk prospect theory is significantly more consistent with rankings and choices than expected utility. This conclusion no longer holds when we use prospect theory utilities and expected utilities to predict intertemporal decisions. The latter finding cautions against the common assumption in health economics that health state utilities are transferable across decision contexts. Our results suggest that the standard gamble and algorithms based on, should not be used to value health.

When are person tradeoffs valid?

The person tradeoff (PTO) is commonly used in health economic applications. However, to date it has no theoretical basis. The purpose of this paper is to provide this basis from a set of assumptions that together justify the most common applications of the PTO method. Our analysis identifies the central assumptions in PTO measurements. We test these assumptions in an experiment, but find only limited support for the validity of the PTO.

New evidence of preference reversals in health utility measurement

A central assumption in health utility measurement is that preferences are invariant to the elicitation method used. This assumption is challenged by preference reversals. Previous studies have observed preference reversals between choice and matching tasks and between choice and ranking tasks. We present a preference reversal that is entirely derived from choices, the basic primitive of economics and utility theory. The preference reversal was observed in two studies regarding health states after stroke. Both studies involved large representative samples from the Spanish population, interviewed professionally, and face-to-face. Possible explanations for the preference reversal are the anticipation of disappointment and elation in risky choice and the impact of ethical considerations about the value of life.

Aversion to health inequalities and priority setting in health care

Traditionally aversion to health inequality is modelled through a concave utility function over health outcomes. Bleichrodt et al. [Bleichrodt, H., Diecidue E., Quiggin J., 2004. Equity weights in the allocation of health care: the rank-dependent QALY model. Journal of Health Economics 23, 157-171] have suggested a "dual" approach based on the introduction of explicit equity weights. The purpose of this paper is to analyze how priorities in health care are determined in the framework of these two models. It turns out that policy implications are highly sensitive to the choice of the model that will represent aversion to health inequality.

New tests of QALYs when health varies over time

This paper performs new tests of the QALY model when health varies over time. Our tests do not involve confounding assumptions and are robust to violations of expected utility. The results support the use of QALYs at the aggregate level, i.e. in economic evaluations of health care. At the individual level, there is less support for QALYs. The individual data are, however, largely consistent with a more general QALY-type model that remains tractable for applications.

Health Utility Bias: A Systematic Review and Meta-Analytic Evaluation

Background. A common assertion is that rating scale (RS) values are lower than both standard gamble (SG) and time tradeoff (TTO) values. However, differences among these methods may be due to method specific bias. Although SG and TTOs suffer systematic bias, RS responses are known to depend on the range and frequency of other health states being evaluated. Over many diverse studies this effect is predicted to diminish. Thus, a systematic review and data synthesis of RS-TTO and RS-SG difference scores may better reveal persistent dissimilarities. Purpose. The purpose of this study was to establish through systematic review and meta-analysis the net effect of biases that endure over many studies of utilities. Methods. A total of 2206 RS and TTO and 1318 RS and SG respondents in 27 studies of utilities participated. MEDLINE was searched for data from 1976 to 2004, complemented by a hand search of full-length articles and conference abstracts for 9 journals known to publish utility studies, as well as review of results and additional recommendations by 5 outside experts in the field. Two investigators abstracted the articles. We contacted the investigators of the original if required information was not available. Results. No significant effect for RS and TTO difference scores was observed: effect size (95% confidence interval [CI]) = 0.04 (−0.02, 0.09). In contrast, RS scores were significantly lower than SG scores: effect size (95% CI ) =−0.23 (−0.28, −0.19). Correcting SG scores for 3 known biases (loss aversion, framing, and probability weighting) eliminated differences between RS and SG scores: effect size (95% CI ) = 0.01 (−0.03, 0.05). Systematic bias in the RS method may exist but be heretofore unknown. Bias correction formulas were applied to mean not individual utilities. Conclusions. The results of this study do not support the common view that RS values are lower than TTO values, may suggest that TTO biases largely cancel, and support the validity of formulas for correcting SG bias.

Willingness to pay for reductions in health risks when probabilities are distorted

We study the willingness to pay for reductions in health risks when people do not evaluate probabilities linearly, as is commonly assumed in elicitations of willingness to pay, but weight probabilities, as is commonly observed in empirical studies of decision under risk. We show that for the levels of baseline risk typically considered, probability weighting strongly affects willingness to pay estimates and may lead to unstable monetary valuations of health.

A nonparametric elicitation of the equity-efficiency trade-off in cost-utility analysis

We performed an empirical elicitation of the equity-efficiency trade-off in cost-utility analysis using the rank-dependent quality-adjusted life-year (QALY) model, a model that includes as special cases many of the social welfare functions that have been proposed in the literature. Our elicitation method corrects for utility curvature and, therefore, our estimated equity weights are not affected by diminishing marginal utility. We observed a preference for equality in the allocation of health. The data suggest that the elicited equity weights were jointly determined by preferences for equality and by insensitivity to group size. A procedure is proposed to correct the equity weights for insensitivity to group size. Finally, we give an illustration how our method can be implemented in health policy.

The economic value of informal care: a study of informal caregivers' and patients' willingness to pay and willingness to accept for informal care

We provide a new test of the feasibility of using contingent valuation to value informal care. We start with a theoretical model of informal caregiving and derive that willingness to pay depends positively on wealth and negatively on own health, whereas the effect of other's health is sign-ambiguous. These predictions are tested in two new data sets on patients' and caregivers' willingness to pay (WTP) and willingness to accept (WTA) for informal care. The data are generally consistent with the theoretical predictions: wealth generally has a positive impact and own health a negative impact. Other's health has a mixed effect. We find only small differences between WTP and WTA. Our findings suggest that contingent valuation may be a useful technique to value informal care in economic evaluations of health care.

A new and more robust test of QALYs

Previous empirical tests of quality-adjusted life-years (QALYs), the most widely used outcome measure in economic evaluations of health care, generally yielded negative results. These tests were, however, for the most part based on expected utility, which is now widely acknowledged to be descriptively inaccurate. The observed violations might, therefore, have been caused by violations of expected utility. We performed a new test of QALYs, which is valid under expected utility and under the two most influential non-expected utility theories, rank-dependent utility and prospect theory, and found considerable support for the QALY model. Our findings suggest that QALYs may be valid if nonexpected utility formulas are used to compute health state utilities.

Equity weights in the allocation of health care: the rank-dependent QALY model

This paper introduces the rank-dependent quality-adjusted life-years (QALY) model, a new method to aggregate QALYs in economic evaluations of health care. The rank-dependent QALY model permits the formalization of influential concepts of equity in the allocation of health care, such as the fair innings approach, and it includes as special cases many of the social welfare functions that have been proposed in the literature. An important advantage of the rank-dependent QALY model is that it offers a straightforward procedure to estimate equity weights for QALYs. We characterize the rank-dependent QALY model and argue that its central condition has normative appeal.

A consistency test of the time trade-off

This paper tests the internal consistency of time trade-off utilities. We find significant violations of consistency in the direction predicted by loss aversion. The violations disappear for higher gauge durations. We show that loss aversion can also explain that for short gauge durations time trade-off utilities exceed standard gamble utilities. Our results suggest that time trade-off measurements that use relatively short gauge durations, like the widely used EuroQol algorithm, are affected by loss aversion and lead to utilities that are too high.

The effect of comorbidities on treatment decisions

Medical decision analyses typically focus on one disease, that is, on one source of risk. In many medical decisions multiple sources of risk co-exist, however. This paper analyzes the effect of such comorbidities on treatment decisions. The effect of comorbidities on treatment decisions depends primarily on the way in which the patient's attitude to health status risks varies with duration. In the QALY model comorbidities do not affect treatment decisions. This property of the QALY model can be used as a diagnostic test of its descriptive and prescriptive validity.

A new explanation for the difference between time trade-off utilities and standard gamble utilities

This paper gives a new explanation for the systematic disparity between standard gamble (SG) utilities and time trade-off (TTO) utilities. The common explanation, which is based on expected utility, is that the disparity is caused by curvature of the utility function for duration. This explanation is, however, incomplete. People violate expected utility and these violations lead to biases in SG and TTO utilities. The paper analyzes the impact on SG and TTO utilities of three main reasons why people violate expected utility: probability weighting, loss aversion, and scale compatibility. In the SG, the combined effect of utility curvature, probability weighting, loss aversion, and scale compatibility is an upward bias. In the TTO these factors lead both to upward and to downward biases. This analysis can also explain the tentative empirical finding that the TTO better describes people's preferences for health than the SG.

A proposal to solve the comparability problem in cost-utility analysis

In cost-utility analysis it is assumed that health state valuations are directly comparable across individuals. Instead, health state valuations may be relative and related to people's expectations and abilities. Then health state valuations are not fully comparable across people and, consequently, cost utility analysis cannot be applied in full. The present paper analyzes this comparability problem and proposes a method to solve it.

Time Preference for Health: A Test of Stationarity versus Decreasing Timing Aversion

This paper provides a new and more robust test of the descriptive validity of the constant rate discounted utility model in medical decision analysis. The constant rate discounted utility model is compared with two competing theories, Harvey's (1986) proportional discounting model and Loewenstein and Prelec's (1992) hyperbolic discounting model. To compare the various intertemporal models, previous studies on intertemporal preferences for health assumed a specific parametric form of the utility function for life-years and no discounting within the time periods that health states are experienced. The present study avoids such confounding assumptions by focusing on the axiomatic structure of the discounting models. The present study further differs by using choices instead of matching to elicit intertemporal preferences. The experimental results provide support for decreasing timing aversion, the condition underlying the proportional and the hyperbolic discounting model, but they violate stationarity, the central condition of the constant rate discounted utility model. There is some ambiguity whether the violations of stationarity are primarily caused by an immediacy effect. The results confirm violations of stationarity in choice-based elicitations tasks, in contrast with the results from Ahlbrecht and Weber (1997) which supported stationarity in choices over monetary outcomes. Copyright 2001 Academic Press.

Life-cycle preferences over consumption and health: when is cost-effectiveness analysis equivalent to cost-benefit analysis?

This paper studies life-cycle preferences over consumption and health status. We show that cost-effectiveness analysis is consistent with cost-benefit analysis if the lifetime utility function is additive over time, multiplicative in the utility of consumption and the utility of health status, and if the utility of consumption is constant over time. We derive the conditions under which the lifetime utility function takes this form, both under expected utility theory and under rank-dependent utility theory, which is currently the most important nonexpected utility theory. If cost-effectiveness analysis is consistent with cost-benefit analysis, it is possible to derive tractable expressions for the willingness to pay for quality-adjusted life-years (QALYs). The willingness to pay for QALYs depends on wealth, remaining life expectancy, health status, and the possibilities for intertemporal substitution of consumption.

Probability Weighting and Utility Curvature in QALY-Based Decision Making

Quality-Adjusted Life-Years (QALYs) are currently the most important utility model in medical decision making. QALYs are calculated by adjusting years of life for the utility of the health state in which these years are spent. For normative reasons the standard gamble is the preferred method to measure health state utilities, but concern exists about its descriptive properties. Recent theoretical work has suggested that probability weighting can explain anomalies in standard gamble measurement. This paper shows that applying probability weighting in standard gamble measurement increases the consistency of QALYs with individual preferences. The consistency of QALYs with individual preferences is not significantly increased further if utility curvature is also taken into account. Copyright 1999 Academic Press.

An experimental test of a theoretical foundation for rating-scale valuations

A major advantage of using a rating scale in health-utility measurement is its practical applicability: the method is relatively easy to understand, and various health states can be assessed simultaneously. However, a theoretical foundation for rating-scale valuations has not been established. The primary aim of this paper is to present a theoretical foundation for rating-scale valuations based on the theory of measurable value functions and to provide a consistency test to see whether rating-scale valuations do indeed elicit a measurable value function. If rating-scale valuations elicit a measurable value function, then Dyer and Sarin have shown how they are related to von Neumann-Morgensterm (vNM) utilities. The appropriate technique to measure vNM utilities is the standard gamble. Torrance has suggested that rating-scale valuations and standard-gamble valuations are related by a power function. A secondary aim of this paper is to examine the relationship between rating-scale valuations and standard-gamble valuations hypothesized by Torrance. An experiment was designed to test consistency of rating-scale valuations and the relationship between rating-scale valuations and standard-gamble valuations. The experiment tested whether rating-scale valuations are independent of the context in which they are elicited, as they should be if they elicit points on a measurable value function. 80 Swedish and 92 Dutch respondents participated in the experiment. The results showed that rating-scale valuations depend on the number of preferred alternatives in the task and thus violate a basic property of measurable value functions. The estimation of the power function did not result in stable results: parameter estimates varied, in some cases there was indication of misspecification, and in most cases there was indication of heteroskedastic errors. The implications of these findings for the common use of rating-scale valuations in cost-utility analysis are serious: the dependency of the rating-scale valuations on the other health states included in the task casts serious doubts on the validity of the rating-scale method.

Standard gamble, time trade-off and rating scale: experimental results on the ranking properties of QALYs

This paper compares the relative performance of quality adjusted life years (QALYs) based on quality weights elicited by rating scale (RS), time trade-off (TTO) and standard gamble (SG). The standard against which relative performance is assessed is individual preference elicited by direct ranking. The correlation between predicted and direct ranking is significantly higher for TTO-QALYs than for RS-QALYs and SG-QALYs. This holds both based on mean Spearman rank correlation coefficients calculated per individual and based on two social choice rules: the method of majority voting and the Borda rule. Undiscounted TTO-QALYs are more consistent with direct ranking than discounted TTO-QALYs.

Income-related inequalities in health: some international comparisons

This paper presents evidence on income-related inequalities in self-assessed health in nine industrialized countries. Health interview survey data were used to construct concentration curves of self-assessed health, measured as a latent variable. Inequalities in health favoured the higher income groups and were statistically significant in all countries. Inequalities were particularly high in the United States and the United Kingdom. Amongst other European countries, Sweden, Finland and the former East Germany had the lowest inequality. Across countries, a strong association was found between inequalities in health and inequalities in income.

Health utility indices and equity considerations

The aim of this paper is to propose methods that incorporate equity concerns into cost utility analysis. The focus of the paper is on QALYs, but the results apply to health utility indices in general. Two interpretations of QALYs are considered: QALYs as (von Neumann Morgenstern) utilities and QALYs as measures of health. A justification is provided for aggregating consistently scaled "QALYs as utilities" over individuals. The conditions underlying unweighted aggregation of QALYs are identified. These conditions exclude two common types of equity concern. Algorithms are proposed that take into account equity concerns and that are relatively easy to apply.

The validity of QALYs: an experimental test of constant proportional tradeoff and utility independence

Pliskin, Shepard, and Weinstein identified three preference conditions that ensure that quality-adjusted life years (QALYs) represent preferences over gambles over chronic health profiles. This paper presents an experimental test of the descriptive validity of two of these preference assumptions: utility independence and constant proportional tradeoff. Eighty students at the Stockholm School of Economics and 92 students at Erasmus University Rotterdam participated in the experiment. The results of the experiment support the descriptive validity of constant proportional tradeoff: both within groups and between groups constant proportional tradeoff could not be rejected. The results are less supportive of the descriptive validity of utility independence. Within-groups utility independence was rejected. Between-groups utility independence could not be rejected, but this may have been due to a lack of statistical power. Analysis of the individual responses revealed that without adjustment for imprecision of preference, 39 respondents (22.8%) satisfied constant proportional tradeoff. Twenty-three respondents (13.4%) satisfied utility independence without adjustment for imprecision of preference. However, because of the relative unfamiliarity of the respondents with both the health states to be evaluated and the methods of health-state-utility measurement, it is likely that the respondents' preferences were imprecise. Adjusted for imprecision of preference, the upper estimates of the proportions of respondents who satisfied constant proportional tradeoff and utility independence, respectively, were 90.1% (155 respondents) and 75.6% (130 respondents). Pliskin et al. further derived that if an individual's preferences satisfy both constant proportional tradeoff and utility independence, then these preferences can be represented by a more general, risk-adjusted QALY model. Without adjustment for imprecision of preference, ten respondents (5.8%) satisfied both constant proportional tradeoff and utility independence. Adjusted for imprecision of preference, the upper estimate of the proportion of respondents who satisfied both constant proportional tradeoff and utility independence was 68.6% (118 respondents). The results of this study indicate that constant proportional tradeoff holds approximately. The evidence is much weaker for utility independence, however. This has important implications for the use of QALY-type measures in medical decision making.

Time preference, the discounted utility model and health

The constant rate discounted utility model is commonly used to represent intertemporal preferences in health care program evaluations. This paper examines the appropriateness of this model, and argues that the model fails both normatively and descriptively as a representation of individual intertemporal preferences for health outcomes. Variable rate discounted utility models are more flexible, but still require restrictive assumptions and may lead to dynamically inconsistent behaviour. The paper concludes by considering two ways of incorporating individual intertemporal preferences in health care program evaluations that allow for complementarity of health outcomes in different time periods.

QALYs and HYEs (healthy year equivalents): under what conditions are they equivalent?

The paper examines what restrictions have to be imposed on the individual's preference structure for QALYs and HYEs to yield identical results. It is shown that using QALYs involves imposing three additional restrictions. Empirical evidence suggests that these restrictions cannot be expected to hold in all applications. The main problem in using HYEs appears to be practical. An alternative index is proposed, that may help to bridge the gap between QALYs and HYEs by combining to some extent the advantages of the two measures.