The idea of paying people to engage in healthy activities, and to refrain from unhealthy ones, gained some traction in the health policy discourse in several developed and developing countries toward the end of the 2000s. The concept itself is simple and is informed by one of the most basic features of standard economics, the relative price mechanism, that is, if you pay someone to do something, there is an expectation that the person is more likely to do it, and the higher the cash incentive, the greater the effect. Here, payments to people to encourage healthier behaviors are defined as user financial incentives (UFIs), and might be expected to have a positive effect if the utility that an individual gains from the payment outweighs the personal disutility consequent on the behavior change.
UFIs exist when an individual can expect a monetary transfer, which is made conditional on them for acting in a particular way. Mostly, they are positive rather than negative financial incentives, rewards not punishments. Taxes on certain harmful products, such as cigarettes, are, however, a form of negative financial incentive and can be quite effective in changing people’s consumption patterns. Moreover, this intervention has gained considerable exposure in recent policy debates internationally, with respect to, for example, the so-called ‘fat taxes’ (i.e., additional taxes on food with high calorific content) and ‘soda taxes.’ The imposition of such measures affects all those engaged in the targeted activity though: those who alter their behavior, by the very fact of them altering their behavior, and those who continue as before, when they are now subject to higher prices. It could be argued, therefore, that taxes are heavily paternalistic. UFI are more libertarian by comparison, in that those who continue to smoke, or eat too much, or refrain from exercise, are not directly affected by the payments at all. It is for this reason – even though UFI ought to be seen as a complement to, rather than a substitute for taxes – that the political leanings of those who currently govern much of the world may lead them to take an interest in positive financial incentives.
Aside from the simplicity of the theory that informs UFI, one must ask oneself, do they work? To answer this, one must turn to the evidence.
Evidence
Numerous experiments with UFI have been, and are being, conducted at the local practical policy level in a number of countries, including the UK, with the use of financial and payment in kind (iPods, hotel breaks, helicopter trips) incentives by healthcare purchasers, and the US, in employer based wellness programs. These local level pilots are, however, rarely well evaluated. Fortunately, there is a reasonable amount of evidence reported in the academic literature on the effectiveness of mostly quite small UFIs to enable us to reach some informed conclusions.
Behavior change can be categorized as either complex or simple. Complex behavior change requires sustained effort over a length of time; simple behavior change requires single actions at a point in time. Some preventive activities – for example, smoking cessation and healthier eating – clearly require a sustained effort, whereas others, such as attending doctors’ appointments and participating in vaccination programs, are more likely to be relatively simple. Sustained and ‘one shot’ behavior change require qualitatively different responses from the individual, and thus it makes sense to dichotomize the evidence according to these categories.
Sustained Behavior Change
The behavior changes discussed here are those associated with smoking cessation and weight loss. There have been systematic reviews in both areas, and the evidence is not auspicious. In relation to smoking, abstinence is measured through use of a biochemical test that records cotinine levels in saliva or urine. In 2008, Cahill and Perera reported a systematic review of studies that have analyzed the effectiveness of UFI for smoking abstinence, wherein only 1 of the 17 studies has demonstrated significantly higher cessation rates for those to whom incentives are offered as compared with those in control groups beyond 6 months from the start of the intervention. Unfortunately, cost information is usually absent from UFI studies and thus even when effectiveness is observed, the discernment of cost effectiveness is difficult.
A 2007 meta-analysis of nine randomized trials on the use of UFI to reduce obesity rates after 12 months following the initiation of the incentive as reported by Paul-Ebhohimhen and Avenell has concluded that an incentive of less than 1.2% of personal disposable income is associated with a zero mean weight change. Financial incentives of at least 1.2% of disposable income were, compared with no incentive, associated with a mean weight loss of 2.4 lb at 12 months and 1.5 lb at 18 months. Thus, the effect by 18 months postinitiation was small and dissipating, and cost-effectiveness information was, again, missing. Indeed, in the domain of weight loss interventions, it is probably difficult to gauge cost effectiveness without knowing the health implications of the loss in weight, which will occur (if at all) many years in the future, and are therefore likely to be confounded by an individual’s broader behaviors, environment, and genetic profile. Rewarding people for weight loss could also feasibly incentivize some quite unhealthy behaviors.
One Shot Behavior Change
As noted above, various types of medical adherence require single, or limited, acts. In 1997, Giuffrida and Torgerson reported a systematic review of UFI to motivate medical adherence (including adherence to a tuberculosis medical regimen, dental care for children, immunization, postpartum appointments, etc.). They identified 11 randomized controlled trials, of which, 10 demonstrated a positive effect. There have been additional studies in the intervening years, many (although by no means all) of which have shown similar promise. For example, a 2003 study by Seal and colleagues reported a randomized controlled trial on a population of hard to reach intravenous drug users in San Francisco. All of the drug users were given the first of three required hepatitis B vaccine doses and then they were divided into two groups, an ‘outreach group’ and an ‘incentive group.’ The third vaccine dose was administered 6 months after the first dose, and the outreach group was assigned a weekly contact with an outreach worker; the incentive group, however, received a monthly US$20 monetary incentive if they remained in the vaccine program. It turned out that 69% of those in the incentive group received all three doses of the vaccine as against only 23% in the outreach group.
In another study, this time from 2005, Slater and colleagues administered two types of mail-based interventions to women aged 40–64 years to encourage them to undergo mammography. Both of the interventions offered a free mammogram if the respondent rang a toll-free number, with one of the interventions also offering a small financial incentive if a person actually underwent the mammogram within 1 year. More than four times as many calls were received for the mail-plus-incentive intervention than the mail-only intervention, and the subsequent mammogram rate was significantly higher in the former intervention than the latter intervention, which in itself produced a significantly higher rate than ‘do nothing.’ As with the interventions to encourage sustained behavior change, however, those that focus on medical adherence are generally silent regarding value for money.
Summary
There is evidence to suggest that UFIs are potentially useful in many areas of medical adherence, but in terms of policy areas that demand more sustained efforts from the targeted groups, the effectiveness of this intervention has been generally poor. In short, for smoking cessation and weight loss, any early success tends to dissipate when the incentives are no longer offered. It may be the case that many studies have been somewhat underpowered, in terms of the size of the incentives offered and the length of time they are offered for. Indeed, a large trial reported in 2009 by Volpp and colleagues at the Center for Health Incentives and Behavioral Economics at the University of Pennsylvania, would seem to suggest that larger incentives may work.
In the trial, 878 people were randomized to either a control group or an incentives group. At baseline, all participants tended to smoke approximately one packet of cigarettes daily. The participants in both groups received information regarding smoking cessation programs, but the incentives group additionally received US$100 for completing a program, a further US$250 for exercising abstinence for 6 months into the trial, and an additional US$400 if they remained abstinent until 6 months thereafter. At 12 months from the initiation of the trial, the cessation rate in the incentive group was significantly higher than that for the controls (14.7% vs. 5%), and although there was some relapse in both groups, this pattern persisted beyond the lifetime of the incentive at 18 months (9.4% vs. 3.6%). Moreover, using incentives in excess of US$100 over a 4-week period, Charness and Gneezy have provided some evidence that gym attendance may be sustained at a significantly higher rate than would otherwise be the case post trial, at least in the relatively short term.
Generally, however, offering larger incentives over longer periods may not be feasible, particularly if financed by the public sector and targeted at broad population levels, given the current global fiscal environment. Therefore, it would make sense to examine whether the payment mechanism in UFI can be redesigned so as to improve its effect.
Strengthening The Payment Mechanism
It may be possible to improve the strength of the UFI payment mechanism by appealing to the findings of behavioral economics. For instance, theoretically, requiring participants to commit their own money (a ‘deposit contract’), with the intention of receiving their money back if they achieve the target behavior, might be expected to improve effect. This is because in the behavioral economics literature, it has been observed that losses loom larger than gains; that is, people attach a greater magnitude of disutility to losing a particular good than the utility that they attach to winning the same good. Given this general observation of loss aversion, we might expect the loss associated with giving up money in deposit contracts to make them more effective than the conventional practice of simply giving people money if they meet their target behaviors. Moreover, some have highlighted the behavioral economic observation that people tend to be attracted to large rewards that have a small probability of occurring, and that therefore, instead of offering the target population a small financial incentive for certain, they ought to be offered, if they successfully change their behavior, a lottery that has the same expected value as the certain payment, but entails some probability of winning a relatively large monetary amount. In short, compared with conventional UFI, deposit contracts and lottery payment mechanisms do not necessitate increases in the average payment, but they may trigger cognitive effects that make respondents perceive the incentives to be more substantial.
In a selection of small studies that have not been specifically informed by behavioral economics, the performance of deposit contracts and lottery payment mechanisms in motivating sustained behavior change has been mixed, at best. There has, however, been at least one UFI for weight loss study that was informed directly by behavioral economics, with some interesting, if not spectacular, results. In this 2008 study by Volpp and colleagues, participants were assigned to one of three arms: (1) a weight monitoring program that required a monthly weigh-in; (2) the weight monitoring program plus a deposit contract, where at the beginning of each month participants could deposit between 1 cent and US$3 per day, with the deposited amount matched by the investigators in addition to a fixed payment of US$3 per day, with all refundable if the participant met the targeted weight loss at the end of each month; and (3) the weight monitoring program plus a lottery incentive, where, if they met their weight loss target, participants played a daily lottery that had an expected value of approximately US$3, with some of the lotteries comprising of a large payoff with small probability but most comprising of a small payoff with larger probability. The trial end point target weight loss for all participants was 16 lb at 16 weeks. At 16 weeks, the mean weight losses in the control group, the deposit incentive group and the lottery incentive group were 3.9, 14, and 13.1 lb, and the percentage of those in each group achieving the 16 lb weight loss target were 10.5%, 47.4%, and 52.6%, respectively: At that point in time, the weight loss was statistically higher in both incentive groups, as compared with the controls. However, at 7 months, although both the incentives groups still on average weighed significantly less than they did at the initiation of the study, the mean weight losses in the control group, the deposit incentive group, and the lottery incentive group were now 4.4, 6.2, and 9.2 lb, respectively, a nonsignificant difference across the three groups. In both incentives arms, therefore, the earlier effectiveness had dissipated considerably.
More experimentation with different UFI payment mechanism designs, informed by the findings of behavioral economics, is warranted. If they can be shown to effect sustained as well as simple behavior change, then they might prove to be a very useful addition to the preventive health policy armory. For this, however, evidence of effect is necessary, but not sufficient. There are many moral and practical objections to the use of UFI that should not be ignored.
Objections
Moral Objections
Some believe that UFIs are unethical. One argument is that trading money for health (or, presumably, health-related behaviors) involves incommensurable values, in the same way that, for example, selling a child for an electoral vote is unacceptable. The commodification of health-related behaviors, according to this view, may lead to their denigration, devaluation, and/or corruption. UFI may denigrate the person’s choices by failing to respect sufficiently the decision that the individual has reached, assuming that he or she has taken into account all of the pros and cons of their actions, which might be particularly dangerous in the field of mental health, where the voice of the patient has often traditionally been ignored.
Thus, there are concerns that UFI, by potentially interfering with the rights of self-governance, may undermine some conceptions of fairness and justice. Moreover, in the area of personal lifestyle behaviors, such as smoking, diet, and exercise, it is possible that the general public will resent general tax revenues being used to reward people for doing what the majority believe they should be doing anyway. However, the general acceptance of UFI may prove context specific. For example, paying people to take their medications might garner general societal support, especially if it is perceived that patients underestimate the pros and/or overestimate the cons of medication. If patients do not fully appreciate the benefits of medication, and if their carelessness toward medication poses avoidable harm for themselves or for the communities in which they live, then the societal view may be that the use of UFI is justifiable.
If UFIs are targeted at the relatively poor, as they often are, they could generate further ethical problems, because even small monetary rewards may deter poor patients from terminating treatment when they feel that it is causing them harm. Furthermore, the offer of a financial incentive could be judged as a bribe when directed toward those with limited means. Conversely, it may be contended that UFIs, whereby those targeted can refuse to participate for any reason, are a model of respectful and equal exchange, that the notion of coercion is more usually associated with punishment than reward, and that it is hard for some to accept how a transparent financial inducement to take medicine in order to remain well undermines society’s notion of fairness and justice.
Unintended Consequences Of UFI
Other objections to UFI focus less on whether they are morally wrong, and more on their potential to have undesirable unintended effects. For example, some worry that the introduction of UFI for some aspects of medication adherence will encourage patients to stop taking their treatment so as to receive the payment for taking their treatment again. Similarly, in relation to broader lifestyle behaviors, it is possible that a few will temporarily initiate unhealthy activities, such as smoking, so as to receive payment for quitting.
A number of experts, across a range of disciplines, have warned that external rewards to act in a particular way may in the long run crowd out the intrinsic desire to alter one’s behavior. For instance, if rewards are offered to people to quit smoking, those payments could potentially become an expectation, and thus people might be less willing to abstain under their own personal motivation. Thus, those who are targeted for UFI must be made aware that any financial incentive is there to serve as merely temporary support to help them achieve a personal goal. Similarly, perhaps, financial incentives to discourage particular unhealthy actions might crowd in other undesirable activities: for instance, an ex-smoker motivated by money rather than health may be more likely than a health-aware reformed smoker to substitute doughnuts for cigarettes.
A monetary payment, if administered from the doctor to the patient, might also crowd out the traditional trust-based nature of the doctor–patient relationship, possibly damaging the principal–agent interaction in this regard if some patients no longer follow advice in the absence of financial rewards. Moreover, attempts to change people’s lifestyle choices are potentially patronizing and condescending to the targeted group; the obese, for instance, may feel unfairly stigmatized beyond the levels to which they have already been stigmatized. There are clearly a host of objections to the use of UFI; whether or not they are insurmountable in all contexts requires a broad public policy debate.
Conclusion
For some aspects of medical adherence, modest financial incentives can have an effect on people’s behavior, although little is yet known regarding the cost effectiveness of these interventions. However, there is currently little evidence of a positive sustained effect on changing lifestyle behaviors when associated with smoking or weight loss. Nonetheless, all hope in this domain should not yet be abandoned; further research that tests the differential impact – and value for money – of different incentive mechanisms that are informed by the findings of behavioral economics is warranted.
There are many ethical and practical objections to the use of UFI. Although the use of UFI is not intended to threaten individual liberty in that participants self-select into incentives programs, which target behavior change that they themselves are meant to desire, there is perhaps a legitimate concern that these interventions violate spheres of privacy and autonomy. For instance, if UFI are used as a public policy tool, it seems not unreasonable to worry that they may patronize people, and might single out a subset of the population (e.g., the obese), stigmatize them further, and treat them as if they lacked full use of reason. Clearly, more public debate is required in order to reach a consensus on the limits of acceptability to the use of UFI, assuming of course that they prove to demonstrate both reasonable effect and value for money in the areas toward which they are targeted.
References:
- Bryan, G., Karlan, D. and Nelson, S. (2012). Commitment devices. Annual Review of Economics 2, 671–698.
- Burns, T. (2007). Is it acceptable for people to be paid to adhere to medication? Yes. British Medical Journal 335, 232.
- Cahill, K. and Perera, R. (2008). Competitions and incentives for smoking cessation. Cochrane Database of Systematic Reviews, 2, 1–31.
- Charness, G. and Gneezy, U. (2009). Incentives and exercise. Econometrica 77, 909–931.
- Cookson, R. (2008). Should disadvantaged people be paid to take care of their health? Yes. British Medical Journal 337, a589.
- Giuffrida, A. and Torgerson, D. J. (1997). Should we pay the patient? Review of financial incentives to enhance patient compliance. British Medical Journal 315, 703–707.
- Gneezy, U., Meier, S. and Rey-Biel, P. (2011). When and why incentives (don’t) work to modify behavior. Journal of Economic Perspectives 25, 1–21.
- Jochelson, K. (2007). Paying the patient: Improving health using financial incentives. London: The King’s Fund.
- Kane, R. L., Johnson, P. E., Town, R. J. and Butler, M. (2004). A structured review of the effect of economic incentives on consumers’ preventive behavior. American Journal of Preventive Medicine 27, 327–352.
- Marteau, T. M., Ashcroft, R. E. and Oliver, A. (2009). Using financial incentives to achieve health behaviour. British Medical Journal b1415.
- Paul-Ebhohimhen, V. and Avenell, A. (2007). Systematic review of the use of financial incentives in treatments for obesity and overweight. Obesity Reviews 9, 355–367.
- Popay, J. (2008). Should disadvantaged people be paid to take care of their health? No. British Medical Journal 337, a594.
- Sandel, M. J. (2012). What money can’t buy: The moral limits of markets. New York: Farrar, Straus and Giroux.
- Seal, K. H., Kral, A. H., Lorvick, J., et al. (2003). A randomized controlled trial of monetary incentives versus outreach to enhance adherence to the hepatitis B vaccine series among injection drug users. Drug and Alcohol Dependence 71, 127–131.
- Shaw, J. (2007). Is it acceptable for people to be paid to adhere to medication? No. British Medical Journal 335, 233.
- Slater, J. S., Henly, G. A., Ha, C. N., et al. (2005). Effect of direct mail as a population-based strategy to increase mammography use among low-income underinsured women ages 40 to 64 years. Cancer Epidemiology, Biomarkers and Prevention 14, 2346–2352.
- Sutherland, K., Christianson, J. B. and Leatherman, S. (2008). Impact of targeted financial incentives on personal health behavior. Medical Care Research and Review 65, 36S–78S.
- Volpp, K. G., Troxel, A. B., Pauly, M. V., et al. (2009). A randomized, controlled trial of financial incentives for smoking cessation. The New England Journal of Medicine 360, 699–709.