By Martha Hostetter and Sarah Klein
Why do many people fail to exercise in spite of the well-known benefits of doing so, while others skip medications that could prevent serious illness and improve their quality of life? Why do some physicians stick to their usual care practices despite ample evidence of better approaches? And why have pay-for-performance programs not always succeeded?
Behavioral economists, who rely on psychology rather than conventional economic principles to explain human behavior, have a few theories. Chief among them is the belief that patients and physicians are "predictably irrational" in their decision-making because they succumb to consistent and recognizable human foibles.1 (Conventional economics, in contrast, is built upon the premise that humans are rational actors who consistently choose options that advance their self-interest.) Without an acknowledgment of this irrationality, interventions designed to encourage patients to adopt healthy behaviors or incentivize physicians to adopt evidence-based guidelines or deliver better coordinated care will, at best, achieve suboptimal results, behavioral economists believe.
To test this theory, researchers have been partnering with employers, health insurers, and pharmacy benefit managers to redesign wellness programs and patient engagement strategies so they’re in keeping with the principles of behavioral economics. New approaches to encouraging providers to engage in innovation are not far behind. The interventions are often described as "nudges" because they seek to preserve people's freedom to choose among a range of options while being guided toward choices that are in their best interest.2 They are informed by research showing that people often have difficulty making wise choices, particularly in the context of their health and care since these involve uncertainty, emotion, and complex trade-offs between current and future costs and benefits.
No Time Like the Present
One of the key principles behavioral economists rely on is "loss aversion"—that is, the tendency to be more sensitive to the prospect of loss than the possibility of gain. This principle is important in explaining why it’s been so difficult to encourage healthy lifestyles: there are immediate costs to eating less or exercising more (junk food tastes good and exercise is hard) while long-term benefits (better health and a little extra time at the end of life) that might counteract the perceived losses are often distant and intangible, says Douglas Hough, a faculty member at Johns Hopkins Bloomberg School of Public Health and author of the book Irrationality in Health Care: What Behavioral Economics Reveals About What We Do and Why (see Q&A with Hough). A related concept is "present bias," which leads people to pay more attention to immediate costs and benefits than equally significant ones that are likely to occur in the future.
In recognition of this, some employers are introducing more immediate losses for engaging in unhealthy behaviors (e.g., higher premiums for employees who smoke) or offering up-front rewards for abandoning them (e.g., a free gym membership). The Affordable Care Act, which beginning in 2014 will allow employers to dedicate up to 30 percent of the cost of their employee premiums toward such wellness incentives, may lead to their wide adoption.3
Yet, according to a recent RAND study, these wellness programs have to date led to only modest behavior changes and limited cost savings, at least in the near term.4 Some cleverly designed interventions are attempting to strengthen wellness programs by providing smaller but more consistent rewards for people to engage in healthy behaviors, such as entering them into a daily lottery, with participation predicated on weight loss. Other programs have participants put their own money or reputation at stake in "deposit contracts." Such approaches may be more effective than reductions to premium contributions or gym memberships since they provide both immediate gratification as well as immediate costs.
Watching Over Patients
Researchers are also using behavioral economic insights—combined with telemedicine tools—to monitor patients’ behavior and encourage them to adhere to treatment regimens. The Center for Medicare and Medicaid Innovation is funding a trial led by researchers at the University of Pennsylvania's Center for Health Incentives and Behavioral Economics (CHIBE) to improve medication adherence and health outcomes among privately insured, Medicare, and Medicaid patients recovering from acute myocardial infarctions in metropolitan Philadelphia and a number of other East Coast regions. Studies show that medication adherence in the year following heart attacks is often poor, in spite of the significant benefits of cardiovascular medication. Even making medication free may not be enough to drive adherence.
Participants in the trial are given a "smart" pill bottle that tracks and wirelessly transmits data about medication use. Each day, patients who have taken their medication are eligible to win cash prizes in a daily sweepstakes system. The next day, they are contacted by e-mail, text, or voice recording telling them whether they won a prize—or would have won if they had taken their medication. These small but frequent sweepstakes take advantage of people's tendency to prefer immediate rewards, and also play on their desire to avoid regret ("If only I had taken my medication, I could have won"). Family members or friends receive automated messages if people miss more than two medication doses—providing social pressure to succeed. Clinical social workers are deployed if people miss four consecutive doses.
Interventions that combine use of telemedicine with behavioral nudges have the potential to strengthen traditional care management approaches, says Kevin Volpp, M.D., Ph.D., founding director of CHIBE and professor of medicine and health care management at the University of Pennsylvania. "New technologies that enable us to track behavior and connect with patients are not in and of themselves going to change behavior in high-risk patients," he says. "For example, if you provide patients who have poorly controlled diabetes—who may have poor diet, a lack of physical activity, poor medication adherence—with a wireless glucometer, they probably will only be partly adherent to using that as well. You have to combine technology with an engagement approach that is really going to provide ongoing active participation for this technology to actually be useful."
Volpp and his colleagues describe this approach as "automated hovering"—a cost-effective way to monitor patients and deliver targeted feedback while they go about their lives. Such hovering can provide clues about why some patients are not adherent to medication regimens—reasons that can't be determined by the fact that a person doesn't refill a prescription. That alone would not reveal whether a patient simply forgot or didn't like the idea of being sick, yet each requires a different intervention, says Bob Nease, M.D., chief scientist at Express Scripts, a St. Louis–based pharmacy benefits manager. "Something that beeps when it’s time to take your medication is not going to help if you’re feeling medicalized," he says.
Decisions, Decisions, Decisions
Behavioral economists have also demonstrated that people don't always make decisions systematically, sifting through all of the available evidence, but instead use heuristics, or rules of thumb, to help them. They are also likely to be subject to "decision fatigue," whereby they may become mentally exhausted when presented with too many decisions and make ill-considered or inconsistent choices, or fall back on the status quo. (This is why candy is located at the cash register—many shoppers have decision fatigue at that point.6 )
Both have significant ramifications for health care providers, who are asked to make multiple decisions in the course of a day, often with great speed and in the face of perverse incentives (to do more is to earn more). Research led by Scott Halpern, M.D., Ph.D., assistant professor of medicine at the University of Pennsylvania Perelman School of Medicine, one of the directors of CHIBE and director of the Fostering Improvement in End-of-Life Decision Science (FIELDS) program within CHIBE, has shown that the average ICU physician makes 11 critical decisions per patient per day—or 1,500 critical care decisions in the course of a two-week rotation with a typical caseload. Given the evidence on decision fatigue, "it’s probable that when clinicians are making decisions over and over again it becomes harder to go against prevailing norms," Halpern says. "This speaks to the opportunity to use standardized care protocols and default care practices to alleviate that decision burden." It also suggests that promoting collaboration among health care providers and relying on multidisciplinary health care teams may improve medical decision-making processes. Group decision-making has been shown to be better than individual decision-making in some cases by reducing decision biases and the impact of the quirks of individual psychology, though there are also some pitfalls to this approach (such as the tendency of groups to fall prey to shared information biases and "groupthink").
In addition to decision fatigue, factors such as loss aversion may help explain why providers are slow to adopt new practices, even in the face of overwhelming evidence of their comparatively greater effectiveness. Physicians may be reluctant to try something new if they think it will slow them down or put their past success at risk—which may help explain why pay-for-performance programs have had limited effects to date.
Conventional economics also plays a role in discouraging new practice, as many incentives to reach certain benchmarks, or provide certain care, operate at the margins alongside the powerful incentives provided by fee-for-service payments. "There is already a giant physician incentive program, which is the fee-for-service program," says Volpp. "If we switch to provider payment modeled on accountable care, there's a lot of work that will need to be done to figure out how to design incentives for providers to no longer maximize revenue but keep patients healthy."
It's easy to see how patients who are asked to make complex decisions in situations when they may feel overwhelmed, confused, or afraid, may have trouble making reasoned decisions about their care. They tend to be persuaded by what their friends or neighbors say instead of reviewing the available evidence and making choices that are best for them. And since patients can't be expected to fully understand medical information, they are influenced by the way in which questions are framed by their physicians.
Framing is critically important when it comes to decisions about end-of-life. "All of our medical ethics and clinical standards are predicated on the notion that people have very strong preferences about the care they wish to receive," says Halpern says. "But end-of-life care is not like any other decision we make—we make decisions once, and we don't get any feedback about what the other option would have looked like."
Noting that many people fail to complete advance directives, Halpern and his colleagues theorized that medicine's default option for end-of-life care—providing aggressive treatment until someone explicitly chooses otherwise—was perpetuating a situation that doesn't promote what terminally ill patients say they want: namely, comfort and a good quality of life until the end. To test this theory, they led a research trial with three groups of patients: one group completed an advance directive for which the default option selected was a comfort-oriented goal, even if it meant shorter life; the second group's directives were defaulted to a life-extending goal, even if it meant more pain and suffering; and the third group's directive didn't have an embedded default. The result was that, among the three groups, twice as many people chose the comfort goal when provided with the "nudge" to do so, compared with those who didn't receive the nudge or who were defaulted to choose a life-sustaining option.10
This is an example of changing the "choice architecture"—the context in which choices are presented—to help people make decisions that are more in line with their interests or professed desires. Another way to do this is to increase the salience of certain factors by drawing people's attention to them, or by listing specific, simple steps. For example, Harvard researchers increased adherence to statins among Oklahoma Medicaid beneficiaries with Type II diabetes (a major predictor of cardiac disease) in part by sending them letters with a vivid description of the consequences of remaining untreated, and a clear and simple call to action.
As patients are asked to become more involved in their care (and shoulder more of the costs), this research offers insights for those who design educational materials and decision aids. "Patient choice" is important, but patients need help in sorting through their choices and support in achieving their goals.
Insights from behavioral economics have the potential to strengthen efforts to engage patients and providers. But while behavioral economics has inherent appeal—it seeks to work with human limitations instead of fighting against them—much more research is needed to put theories into practice. "We don’t know if behavioral economics in health is the next big thing. You therefore don’t want to put all of your eggs in that basket. Still, it offers promise because it does provide many important insights and policy implications that cannot be drawn from traditional economic models of behavior," says Thomas Rice, professor of health policy and management at UCLA Fielding School of Public Health.
There are also ethical questions to be considered when attempting to influence behavior, such as: to what extent should people be pushed and prodded into doing things that are good for them? In a free society, should people be allowed to be obese, smoke, skip their medications, or pursue medical care that may be of little value to them?
"The central tension for me boils down to this: as a society we wish to support people’s authority to make medical decisions, such as the type of care they wish to receive at the end of their life," says Halpern. "At the same time, we also want patients to make choices that promote their own interests. But in most cases patients can’t achieve that on their own—they procrastinate, get tripped up by well-documented foibles of human decision-making. So do we let patients make decisions that divert them from their own goals, or do we intervene? Our group’s standpoint is that as long as the nudges are true nudges, where a decision-maker is free to choose otherwise, then we are likely to achieve a great deal more good by helping overburdened people make decisions that are likely to be consistent with their goals."
1 D. Ariely, Predictably Irrational: The Hidden Forces That Shape Our Decisions (New York: HarperCollins, 2008).
2 R. H. Thaler and C. Sunstein, Nudge: Improving Decisions About Health, Wealth, and Happiness (New Haven: Yale University Press, 2008).
3 Or 50 percent at the discretion of the U.S. Department of Health and Human Services.
4 S. Mattke, H. Liu, J. P. Caloyeras et al., Workplace Wellness Programs Study: Final Report, RAND Corporation, 2013, available at http://www.rand.org/content/dam/rand/pubs/research_reports/RR200/RR254/RAND_RR254.sum.pdf.
5 N. K. Choudhry, J. Avorn, R. J. Glynn et al., "Full Coverage for Preventive Medications After Myocardial Infarction," New England Journal of Medicine, Dec. 1, 2011 365(22):2088–97.
6 T. Rice, “The Behavioral Economics of Health and Health Care,” Annual Review of Public Health, March 2013 34:431–47.
7 See https://www.boundless.com/management/decision-making/managing-group-decision-making--5/advantages-and-disadvantages-of-group-decision-making/.
8 Institute of Medicine, Crossing the Quality Chasm (Washington, D.C.: National Academies Press, 2001).
9 G. Flodgren, M. P. Eccles, S. Shepperd et al., "An Overview of Reviews Evaluating the Effectiveness of Financial Incentives in Changing Healthcare Professional Behaviours and Patient Outcomes," Cochrane Database of Systematic Reviews, July 6, 2011 (7):CD009255. Also see S. Woolhandler, D.Ariely, and D. Himmelstein, "Will Pay for Performance Backfire? Insights From Behavioral Economics," Health Affairs Blog, Oct. 11, 2012.
10 Changing the default on employer-based retirement programs so that employees must opt out of participation rather than in has greatly increased employee participation in these programs. See J. J. Choi, D. Laibson, B. C. Madrian et al., "For Better or for Worse: Default Effects and 401(k) Savings Behavior," from D. Wise, Perspectives on the Economics of Aging (Cambridge, Mass.: National Bureau of Economic Research, June 2004).