A Randomized Controlled trial of Premium-based vs. Lottery Incentives for Workplace Weight Loss

Background: Employers commonly use adjustments to health insurance premiums as incentives to encourage healthy behavior, but their effectiveness is unclear. The typical manner in which this is done (bundled in with premiums and delayed in time) might be less effective than alternative ways of providing incentives that are both more salient and provide more frequent ongoing feedback.

Methods: 197 employees of a mid-size company who were already enrolled in a workplace wellness program were given a weight loss goal equivalent to 5 percent of their baseline weight. They were randomly assigned to a control arm, with no financial incentive for achieving the goal, or to one of three intervention arms offering an incentive valued at $550. Two intervention arms used health insurance premium adjustments, beginning the following year (delayed) or in the first pay period after achieving the goal (immediate). A third arm used a lottery incentive separate from premiums that was run on days participants came into work (schedules varied since many of the employees were nurses on different shifts).

Results: Mean BMIs at baseline were 36.9. At twelve months there were no statistically significant differences in mean weight change either between the control group (whose members had a mean gain of 0.1 pound) and any of the incentive groups (delayed premium adjustment, -1.2 pound; immediate premium adjustment, -1.4 pound; daily lottery incentive, -1.0 pound) or among the intervention groups . The 5 percent weight loss goal was achieved by 37 participants in the study (18.8 percent), but there were no significant differences in percentages meeting the goal across the study groups. There were no significant differences in demographic characteristics between participants who did and those who did not achieve the 5 percent weight loss goal. On average, participants who achieved the goal did not regain weight before the end of the study (mean change in weight after meeting goal: -0.13 pounds). The adjusted multivariate model revealed no significant differences in mean weight loss among any of the intervention groups and control.

Limitations: Weights were only tracked within the workplace and some employees reported feeling uncomfortable with having weights tracked at work. We did not have work schedules so could not incorporate any regret messaging in the lotteries. There was a significant amount of attrition over time.

Conclusions: The apparent failure of either delayed or immediate payment of incentives to lose weight through premium reductions suggests that this widely used approach by employers is not effective. However, an alternative approach to using episodic lotteries to help employees lose weight was no more effective. Further work is needed to design a more effective strategy to using benefit design to help obese employees lose weight.