Research

This study explores the temporal aspects of the weight loss process within the framework of time-to-event analysis (Survival analysis) using data from a Randomized Controlled Trial (RCT) in a weight control program. The experiment includes 418 participants (206 in the treated group and 212 in the control group) across two clinical sites (South Carolina and Vermont) and comprises two main phases: the initial weight loss phase (months 1-6) and the subsequent weight maintenance phase (months 7-18). Participants randomized into the treated group receive financial incentives based on their behaviors and outcomes during the study.

For the weight loss phase, we investigate whether and how much financial incentives are effective in reducing the duration to achieve weight loss and which participants benefit most from incentive structures. For the weight maintenance phase, we explore whether faster weight losers maintain better long-term outcomes and what long-term weight outcomes would look like under alternative intervention designs. Our analytical framework combines multiple complementary causal inference approaches to address identification challenges. We employ Cox models to estimate average treatment effects, causal survival forests to identify individual-level treatment effects, principal stratification analysis to estimate the effect of weight loss achievement time on weight maintenance duration, and sequential G-estimation to simulate counterfactual policy scenarios to measure the effects of alternative incentive timing strategies.

Our findings reveal that financial incentives increase the probability of achieving weight loss by approximately 37% at any given time, with substantial treatment effect heterogeneity across participants. Participants with higher baseline BMI demonstrate systematically larger benefits from financial incentives. Our weight maintenance analysis reveals that participants who become "fast weight losers" only under treatment conditions experience significantly longer maintenance benefits. Finally, our counterfactual analysis reveals that behavioral incentives combined with outcome incentives provide significantly higher success rates for weight maintenance than outcome incentives alone.

Our findings offer actionable policy recommendations for optimizing weight control interventions. Healthcare systems should prioritize targeting individuals with higher baseline BMI with financial incentives and develop comprehensive intervention designs combining behavioral incentives with outcome incentives for significantly higher success rates than outcome incentives alone.