Irrationality in Health Care

What Behavioral Economics Reveals About What We Do and Why

Douglas E. Hough

• Publisher: Stanford Economics and Finance
• Year: 2013
• ISBN:
  9780804785747
• Edition: 1
• Topic: Économie
• Subtopic: Microéconomie

The health care industry in the U.S. is peculiar. We spend close to 18% of our GDP on health care, yet other countries get better results—and we don't know why. To date, we still lack widely accepted answers to simple questions, such as "Would requiring everyone to buy health insurance make us better off?" Drawing on behavioral economics as an alternative to the standard tools of health economics, author Douglas E. Hough seeks to more clearly diagnose the ills of health care today.

A behavioral perspective makes sense of key contradictions—from the seemingly irrational choices that we sometimes make as patients, to the incongruous behavior of physicians, to the morass of the long-lived debate surrounding reform. With the new health care law in effect, it is more important than ever that consumers, health care industry leaders, and the policymakers who are governing change reckon with the power and sources of our behavior when it comes to health.

1

WHAT IS BEHAVIORAL ECONOMICS—AND WHY SHOULD WE CARE?

It is time, therefore, for a fundamental change in our approach. It is time to take account—and not merely as a residual category—of the empirical limits of human rationality, of its finiteness in comparison with the complexities of the world with which it must cope.

—Herbert Simon (1957)

The health care industry in the United States is peculiar. We spend close to 18 percent of our gross domestic product on health care, yet other countries seem to get better results—and we really don’t know why. Most health care products and services are produced by private organizations, yet federal and state governments pay for about half of these services. More starkly, those who consume health care do not pay for it, and those who do pay for that care do not consume it. That is, patients pay less than 15 percent of their care at the point of purchase, the rest being picked up by their employers, private insurance companies, Medicare, or Medicaid (which have no need for physician visits, medications, or surgeries themselves). Health care is also peculiar on the supply side. Unlike in every other industry, the people who fundamentally determine how resources are allocated—that is, the physicians—rarely have any financial stake (as owners or employees) in the resources that they control in hospitals, nursing homes, or other facilities.

It is no wonder, then, that economists like myself are fascinated by this industry and are turning our theoretical and empirical tools to all aspects of demand and supply. Although we have made some headway in understanding the health care industry, the standard tools do not seem to be helping us to understand much about the behavior of patients, physicians, and even society as a whole. In this book I will offer a new economic lens that I hope will provide more clarity in diagnosing the problems facing the business side of health care. This lens—behavioral economics—is helpful in understanding the “micro” decisions that we make as patients and that physicians make as they care for us. In addition, it yields insights into the “macro” decisions we make as a nation regarding how we organize and pay for health care. I will introduce the concepts of behavioral economics by discussing a series of what I call “anomalies,” that is, behavior—both individual and societal—that just does not seem to be rational. For example, we will consider:

• Why would requiring everyone to buy health insurance make everyone—including those who don’t want to buy health insurance—better off?

• Why do patients insist on getting a prescription, shot, test . . . when they go to a physician with an ailment—yet many patients do not adhere to their diagnostic and treatment regimens?

• Why do tens of thousands of patients die each year in the United States from central line–associated bloodstream infections—even though a simple five-step checklist used by physicians and nurses could reduce that number by two-thirds?

My point is not that these anomalies occur because people are stupid or naïve or easily manipulated. Rather, it is that we—as consumers, providers, and society—need to recognize the power of arational behavior if we are to improve the performance of the health care system and get what we pay for.

MAINSTREAM ECONOMICS AND ITS ASSUMPTIONS

Most economists practicing today learned their trade in what is known as the neoclassical tradition. We were trained in a school of economic thought that traces its heritage back to Adam Smith and The Wealth of Nations, published in 1776. In this world, markets—properly organized—allocate scarce resources to their highest and best use through the application of Smith’s famous “invisible hand.” The primary role of the government is to ensure that markets are properly organized and operated and then to get out of the way. Buyers and sellers, in seeking to further their own gains and with little or no conscious intent to improve public welfare, will be led to maximize their “utility” (economists’ term for happiness or satisfaction) or profit. In fact, using both graceful exposition and elegant mathematics, neoclassical economists have been able to prove what became known as the “fundamental theorem of welfare economics,” that a competitive market will generate a Pareto-optimal allocation of resources. That is, this market-generated allocation will yield the highest collective value of those resources. They proved that any deviation from that allocation would benefit some buyers and sellers only at the expense of others.

As you might imagine, this finding has been used to justify capitalism and the market economy. At the same time, it has been used to explain the evils of monopolies (because monopolies typically raise prices above what would be charged in a truly competitive market) and to defend the intervention of the government to limit pollution (because private markets typically do not factor in the costs of pollution to society).

This theory of economics rests on a number of critical assumptions about the structure of the market and the behavior of buyers and sellers in the market. It is important for the discussion here to describe these assumptions, why they are important, and how the theory can fail if the assumptions are not valid. The first—and most fundamental—assumption is that everyone is rational. That is, standard economics assumes that buyers and sellers, individuals and organizations, always act in their own best interests. If participants in the market are not always rational, then they will not make decisions that promote their well-being (either satisfaction/happiness on the part of consumers or profits on the part of sellers), and mainstream economists will be at a loss as to how to proceed.

Second, mainstream neoclassical economics assumes that all participants in the market know their preferences. Again, it would be difficult for a consumer to maximize his or her preferences without knowing what they were. Third, the theory assumes that all participants in the market have full information—about the products in the market, their features and drawbacks, and the prices being offered by various sellers. Understandably, if consumers are not aware of the alternatives that face them, it will be difficult for them to make the right decisions. Similarly, sellers need to know about the preferences of consumers and the range of products being offered by competitors if they are to offer the right product at the right price and sell their wares.

A somewhat less intuitive assumption of standard economics is that consumer preferences and decisions are path independent. The preferences that consumers have and the decisions that they make should not depend on how they arrive at those preferences or decisions. For example, a consumer’s willingness to buy a particular car should not depend on whether he saw a more expensive or less expensive car first or whether he saw a blue car (a color he loves) before or after a green car (a color he despises). If consumer preferences and decisions are based on these external and seemingly irrelevant factors, then one has to question the validity of his choices.

Finally, even mainstream economists admit that consumers and producers sometimes make mistakes. Even so, these economists assume that the mistakes are random and not systematic. So, if people miss the mark in making decisions that improve their situation, sometimes they will be above the mark, and sometimes they will be below—and we have no way to predict what mistakes they will make.

It may be pretty obvious that these assumptions do not accurately describe reality all of the time or, in fact, most of the time. People do not always act rationally; they often do not have full information about the products or services they may want to purchase; and occasionally they may not know exactly what they prefer. Mainstream economists have spent a lot of energy over the past several decades analyzing what happens when these assumptions are violated. Going into this work will take us too far afield. However, we should note a rather profound argument made by two prominent economists—Milton Friedman and Leonard Savage—regarding the importance of assumptions.

In an influential article written over sixty years ago, Friedman and Savage (1948) confronted the contention that bad assumptions lead to bad theory. They argued that economic theory does not assert that people act exactly as the assumptions claim that they do; instead, it is sufficient that people only act as if they were obeying the assumptions. Friedman and Savage argue that this “as if” nuance is crucial. They maintain that any theory should be evaluated on the accuracy of its predictions, not on the reality of its assumptions. If the assumptions allow the economist to develop a theory that yields results that outperform other theories, then the assumptions themselves are irrelevant.

Friedman and Savage support their argument by their now-famous analogy of a billiards player. A scientist might want to predict the path of a ball struck by an expert billiards player during a match. Friedman and Savage offer that it might be possible to develop mathematical formulas that predict the optimal force and direction of the cue and all the balls on the table. Such a theory of billiards behavior may require an assumption that the player knows and uses these formulas, more correctly that the player acts as if she knows and uses the formulas. As Friedman and Savage argue, “It would in no way disprove or contradict the hypothesis, or weaken our confidence in it, if it should turn out that the billiard player had never studied any branch of mathematics and was utterly incapable to making the necessary calculations” (p. 298), as long as the assumption was necessary for the development of the hypothesis and the theory predicted the results of the billiards shot better than any competing theory. The implication of this line of reasoning for economics is that the realism of the assumptions may not matter if the theory of behavior that uses these assumptions generates the most accurate predictions.

Given this criterion, mainstream neoclassical economics has held up very well over the past several decades compared to other competing economic theories. It has dispatched radical political economics (aka Marxism) following the fall of the Soviet Union and the Berlin Wall. It has proved to be more discriminating than institutional economics, which has failed to generate much interest since John Kenneth Galbraith retired. Evolutionary economics, despite the best efforts of renowned economists such as Sidney Winter and Richard Nelson, has not yet generated testable hypotheses that rival neoclassical economics.

THE CHALLENGE OF BEHAVIORAL ECONOMICS

Then came behavioral economics. This field was formed largely through the work of Daniel Kahneman and Amos Tversky in the 1970s. Ironically, Kahneman and Tversky (who died in 1996) are behavioral psychologists, not economists. Kahneman and Tversky first became known to most economists through a 1979 article, “Prospect Theory: An Analysis of Decision Under Risk,” in the highly respected and mathematically rigorous journal, Econometrica (Kahneman and Tversky 1979). In that article, they presented the first explication of the tenets of behavioral economics. Despite its scarcity of mathematics, the article is the most frequently cited article ever published in Econometrica and the second most frequently cited article in the economics literature in the past forty years (Kim, Morse, and Zingales 2006).

What Kahneman and Tversky termed prospect theory has since evolved into what is generally referred to as behavioral economics. It has offered an interesting, and often compelling, alternative to mainstream neoclassical economics. First, it makes assumptions about human behavior that have greater face validity to both economists and laypeople. For example, it acknowledges that not everyone is rational, at least not all the time. Rather, buyers and sellers, individuals and organizations, do not always act in their own best interests. In addition, behavioral economics assumes that people do not have what neoclassical economists call a utility function, which maps all of the available goods and services and other contributions to happiness into a permanent set of preferences for each individual. Instead, behavioral economics assumes that people learn their preferences through experience, via trial and error. In addition, they make their decisions based on their current situation (which acts like a reference point), not from some overarching utility function.

A further assumption is that incomplete information abounds. Neither buyers nor sellers have all the information that they would like. Sometimes, the buyers do not have enough information, such as when they are buying a used car from the original owner. Sometimes it is the sellers that lack information, such as an insurance company writing a life insurance policy for an individual; the buyer knows his behavior, health history, and other risk factors, but the insurer does not.

Behavioral economists have found that preferences are, indeed, path dependent. Economic decisions are often influenced by factors independent of the individual: Buyers buy differently if they are shown a more expensive house before a less expensive house, a fully equipped car before a stripped-down model, a fifty-two-inch LCD television before a more modest set.

Not only does behavioral economics assume that buyers and sellers are not always rational; the theory also has a fundamental tenet that deviations from rational choice are systematic and can be predicted. This aspect of behavioral economics ultimately sets it apart conceptually from mainstream neoclassical economics and provides the focal point for testing the relative effectiveness of the two approaches for viewing human behavior.

I should note one final difference between mainstream neoclassical economics and behavioral economics. Neoclassical economics has largely been a self-contained discipline, developed by economists, for economists. If neoclassical economists have borrowed concepts from another field, it’s been mathematics. On the other hand, behavioral economics owes a very large intellectual debt to the discipline of psychology, especially behavioral psychology. By the nature of its assumptions, behavioral economics depends fundamentally on the perspective, hypotheses, and empirical studies of behavioral psychologists. In fact, there will be times in this book in which it is not clear whether we are looking at behavioral economics or behavioral psychology phenomena—and that is by design.

THE CONTRIBUTIONS—SO FAR—OF BEHAVIORAL ECONOMICS

In the next chapters we will be exploring the full range of explanations and predictions that behavioral economics can offer. However, I want to illustrate the relative power of behavioral economics here by examining three areas in which this theory provides explanations that are superior to mainstream neoclassical economics: decision biases, the power of the default, and the special value of zero.

First, let’s take a look at decision bias. Everyone makes mistakes, even neoclassical economists. What behavioral psychologists have demonstrated—and behavioral economists have used—is that people tend to make bad decisions in a particular way. (Note that I used the word tend; what behavioral psychologists have found are tendencies, not immutable laws of behavior. People cannot be as predictable as atoms, so psychology and economics cannot be as definitive as physics.) For example, a host of psychology studies have found that most people are overconfident about their abilities. In a famous study, university students in Oregon and Stockholm were asked to rate their driving compared to other students (Svenson 1981). About 80 percent of the American students thought that they were safer than the median student driver, and about 75 percent thought they were more skilled than the median student driver. Only 12.5 percent thought that they were less safe than the average, and only 7.2 percent thought that they were less skilled than the average. (The Swedish students were somewhat less confident than their American counterparts.) Perhaps the researcher just stumbled on a group of NASCAR protégés, but more likely this finding is an example of a “Lake Wobegon” effect, where all the children are thought to be above average.

We could write these results off as the wishful thinking of inexperienced, callow youth. But a recent study of attorneys shows that overconfidence bias may be endemic. Jane Goodman-Delahunty and her colleagues (Goodman-Delahunty et al. 2010) surveyed 481 litigation attorneys in the United States. They were asked two questions about a current case: “What would be a win situation in terms of your minimum goal for the outcome of this case?” and “From 0 to 100%, what is the probability that you will achieve this outcome or something better?” Sixty-four percent of the attorneys gave confidence estimates that exceeded 50 percent, which—as Goodman-Delahunty and her colleagues note—may not be surprising given that attorneys are trained to be zealous advocates for their clients and thus are likely to be optimistic about the case’s outcomes.

The attorneys were then recontacted when their cases were resolved. Fifty-six percent of the time, the outcomes of the cases met or exceeded the attorney’s minimum goals. However, in 44 percent of the cases the actual outcomes were less satisfactory than the minimum goals that the attorneys had set earlier. (By the way, there were some attorneys who were underconfident, but they were far outnumbered by the correctly confident and overconfident ones.) Ironically, only 18 percent of the attorneys said that they were very disappointed or somewhat disappointed in the outcome of the case. Now, that’s confidence.

In breaking down the results, Goodman-Delahunty and her colleagues found the greatest gap between the prior estimate and the subsequent outcome was for those attorneys who expressed the highest confidence that their goal would be achieved. That is, those who were the most confident before the case was decided were the most likely to be wrong.

Finally, one would hope that attorneys would learn from their mistakes—but they did not, according to this study. More senior, experienced attorneys were as overconfident as their junior colleagues. In an attempt to assist the respondents in calibrating their confidence estimates, the researchers asked 212 of the 481 attorneys in the initial survey to provide reasons why their litigation goals might not be achieved. Those who gave reasons were just as likely to be overconfident as those who did not.

If rampant overconfidence were not enough, Justin Kruger and David Dunning conducted a variety of experiments that supported Charles Darwin’s statement in his The Descent of Man, “Ignorance more frequently begets confidence than does knowledge” (Kruger and Dunning 1999). Groups of undergraduate volunteers from Cornell University were given various tests of logical reasoning and grammar. At the end of the test, the participants were asked to estimate how their score would compare with that of their classmates. As with the studies already described, participants overestimated their ability and performance, placing themselves on average between the sixty-fifth and seventieth percentile. The unique contribution of Kruger and Dunning is that they disaggregated the respondents by performance quartile and analyzed those responses. Those in the highest two quartiles of performance accurately predicted their ability and performance. However, those in the bottom half dramatically overestimated their results. For example, in the grammar test, those in the bottom quartile scored at the tenth percentile but estimated their ability in grammar in the sixty-seventh percentile and their perform...