Self‐Weighing Increases Weight Loss in Free‐Living Adults: A Double‐Blind Randomized Field Trial among 200,000 Health App Users

Abstract

Self‐weighing is closely associated with sustained weight loss, but randomized trials to test causality have been limited to small samples in controlled settings subject to Hawthorne effects. This study employs a simple behavioral nudge to induce more frequent self‐weighing among new users of the Lose It! app, randomly assigned among all individuals who enrolled over a four week period in 2014. The study population is similar in age, sex and socio‐economic status to other groups seeking weight‐loss assistance in the United States and Canada. There were no differences between treatment (T) and control (C) other than the nudge, and subjects were not told about differences between T and C. Among the 184,955 users who recorded at least one plausible data point, we find mean self‐weighing frequency (days) was T=4.27±0.07 and C=3.85±0.07, and mean weight loss (lbs) was T=1.95±0.04 and C=1.90±0.04, for a difference in weight loss of 0.054±0.052 over the first three months after enrollment. Using two‐stage least squares to estimate the causal effect of self‐monitoring, we find that each additional weigh‐in led to an average increase in weight loss of 0.131±0.119 lbs. These results suggest that encouraging frequent self‐weighing can improve the efficacy of weight loss programs at a national level, at very low additional cost.