A Randomized Controlled Trial of strategies for Weight Loss

Presenter: Kevin Volpp, University of Pennsylvania

Abstract

Objectives: Identifying effective strategies for treating obesity is both a clinical challenge and a public health priority due to the health consequences of obesity and its contribution to health disparities. This study tests whether common decision errors identified by behavioral economists such as prospect theory, loss aversion, and regret could be used to design an effective weight loss intervention.

Methods: 57 subjects at Philadelphia VA Medical Center (mean BMI 34.9, 42% African American) enrolled in a 3-arm randomized controlled trial (RCT) in which all subjects were given weight loss targets of 16 pounds in 16 weeks, a counseling session with a nutritionist, and subjects were randomized to: 1) monthly weigh ins; 2) entry into a daily lottery (L) with a 20% chance of winning $10 and a 1% chance of winning $100, with earnings paid only if subjects are below target weights; 3) an option to deposit $.01-$3.00 per day of their own money (D_, matched 1:1 by the investigators, in which subjects receive the sum of both amounts plus a fixed payment of $3.00 each day if below their target weight, but are paid nothing (and lose their deposit) if not. All incentives were disbursed monthly once weights are confirmed in clinic. All participants were given a weight loss goal of 1 pound per week for 16 weeks, and results were analyzed using intention-to-treat analysis of variance models.

Results: Participants in both incentive groups lost significantly more weight than participants in the control group (3.9 pounds); (Lottery = 13.1 lbs; p-value for lottery vs. control .018, 95% CI of the difference in means [1.95, 16.40]); deposit contract = 14.0 lbs, p-value vs. control .006, 95% CI of the difference in means [3.69, 16.43]). 47.4% (95% CI [24.5%, 71.1%]) of deposit contract participants and 52.6% (95% CI [28.9%, 75.6%]) of lottery arm participants met the 16-pound weight loss goal compared to 10.5% (95% CI [1.3%, 33.1%]) in the control group (p-value = .014). By the end of 7 months, substantial amounts of weight were regained. While the net weight loss between enrollment in the study and the end of 7 months was larger in the incentive arms (L = 9.2 lbs; DC = 6.2 lbs) than in the control group (WM = 4.4 lbs), these differences were not statistically significant (L: t = 1.21, p = .23, 95% CI [-3.20, 12.66]; DC: t = 0.52, p = .61, 95% CI [-5.17, 8.75]). However, incentive participants weighed significantly less at 7 months than at the study start (L: p value = .010, D: p value = .027) whereas controls did not.

Conclusion: The use of economic incentives produced significant weight loss during the 16 weeks of intervention. This weight loss was not fully sustained. The longer-term use of incentives should be evaluated, as this approach could have a major impact in reducing the incidence of obesity-related illnesses in both the VA and US populations.