Nudging Smokers

Cass R. Sunstein, J.D.

N Engl J Med 2015; 372:2150-2151May 28, 2015DOI: 10.1056/NEJMe1503200

In the past 40 years, we have seen a revolution in thinking about thinking.1,2 The central idea is that human beings depart, in systematic ways, from standard economic approaches to rationality. Because the departures are systematic and predictable, they can be taken into account by researchers, clinicians, and others who want to improve health and reduce premature mortality.

Behavioral scientists have shown, for example, that people are “loss averse”; they tend to dislike losses more than they like corresponding gains. A 5-cent tax on the use of a grocery bag is likely to have a much greater effect than a 5-cent bonus for bringing one’s own bag. People also suffer from “present bias”: they tend to focus on the short term and sometimes see the future as a kind of foreign country (and their future selves as strangers). Most people tend to be unrealistically optimistic, at least about their own prospects.3 People are sometimes aware of their own biases and are willing to precommit to courses of action that will counteract them.4

These and related findings help to explain preventable health problems and also suggest a wide range of potentially promising interventions.2 Some such interventions involve economic incentives, in the form of subsidies and penalties. Others involve “nudges,” in the form of choice-preserving interventions that do not impose any such incentives; information, warnings, reminders, and default rules are examples of nudges. Still other interventions combine the two, as with nudges designed to encourage people to enter programs that impose economic incentives.

In this issue of the Journal, Halpern and his colleagues5 investigate the role of financial incentives and nudges in promoting smoking cessation, with particular emphasis on the importance of loss aversion and precommitment. They explore two kinds of interventions for CVS Caremark employees and their relatives and friends. The first is a “deposit program,” in which smokers deposit $150, which they can get back if they stop smoking, along with $650 extra. The second is a “reward program,” in which people receive $800 if they stop smoking. Halpern et al. are interested above all in two questions. First, when will smokers actually enter a cessation program? Second, which of these cessation programs will be most effective? They asked these questions both for individuals and for targeted six-person groups.

As compared with usual care (involving information and, for participants enrolled in the CVS Caremark health care plan, free smoking-cessation aids), both the deposit program and the reward program had significant effects, in the form of higher abstinence rates through 6 months. But the authors also found major differences between the two programs. (The results were essentially the same for individuals and for groups.)

In brief, the deposit program was much less attractive but much more effective. Only 13.7% of the participants assigned to the deposit program chose to enroll in it, whereas 90.0% of those assigned to the reward program chose to enroll. By contrast, 52.3% of those enrolled in the deposit program had sustained abstinence for 6 months, as compared with just 17.1% of those enrolled in the reward program. It is tempting to question the second finding on the grounds that those who are willing to enroll in a deposit program are especially determined to quit, and if so, it is no surprise that they show higher cessation rates. Halpern et al. tried to control for this selection effect by examining smokers who would accept either program. Even with this control, they found a significantly larger effect from the deposit program.

Halpern et al. ended up demonstrating the importance of loss aversion in two different ways. The more obvious is that smokers are far more likely to quit if they stand to lose money if they fail. The more subtle is that the very prospect of incurring losses makes people far less willing to enter a smoking-cessation program. Despite the greater comparative effectiveness of the deposit program, the reward program is likely to be more successful, because far more people will sign up for it.

There are implications for both future research and public policy. With respect to research, it would be valuable to know whether a smaller deposit might increase participation without reducing efficacy. With respect to public policy, deposit programs, which enlist loss aversion, are the better way to help people to quit smoking (and perhaps to alter many kinds of health-related behavior). The challenge is to find a way to nudge people to enroll in such programs. If that challenge cannot be met, reward programs are much better bets.

Disclosure forms provided by the author are available with the full text of this article at NEJM.org.