Association for Public Policy Analysis & Management

Objective: In 2016, over sixty percent of large U.S. employers offered a employer wellness program (EWP), a strategy for reducing risk factors that may lead to chronic illnesses among employees. As the popularity of wearable tracking devices has grown, many employers have begun to offer physical activity tracking applications as part of their EWPs, allowing individuals to sync wellness data to various devices. We evaluate the effectiveness of a physical activity tracking application offered by an employer for improving individual biometric outcomes and reducing health expenditures. We also examine heterogeneous treatment effects of the program after the introduction of outcomes-based incentives. In an outcomes-based incentives structure, only individuals who meet specific BMI, cholesterol and blood pressure standards, or “within range individuals,” are directly rewarded. Individuals who do not meet these standards, “out of range individuals,” can only be rewarded by participating in wellness activities and/or showing health improvements.

Study Design: We used three data sources from a large employer, from 2011-2014: 1) health benefits eligibility data, 2) medical claims, and 3) wellness participation data. We employ a difference-in-differences (DID) analysis to estimate the average treatment effect of participation in the physical activity tracking program on biometric outcomes and monthly health care expenditures. We control for a number of demographic variables, health status characteristics, and lifestyle behaviors. In 2012, the employer transitioned to an outcomes-based incentive design, which only directly rewards individuals who fall within a pre-determined health range. We stratified the impact analysis by whether individuals were categorized as “within range” at their baseline biometric screen in order to assess the heterogenous effects of the program. We include all employees and dependents who were continuously enrolled in an eligible health benefit plan (n= 44878 person-years).

Findings: The DID results show that participation in the tracking program is associated with a reduction of 0.229 in BMI (p<0.001). This partially offsets the overall secular growth trend in BMI by about 26%. The program did not have a statistically significant effect however, on cholesterol, blood pressure, or medical and pharmaceutical expenditures. The outcomes-based incentives analysis results showed participation is associated with a statistically significant reduction in BMI, cholesterol and systolic blood pressure for only the individuals in the “within range” group.

Conclusions: The physical activity tracking program was associated with a statistically significant reduction in BMI, but the size of the effect was only large enough to partially offset the increasing secular trend in BMI. The outcomes-based incentives analysis results suggest that the benefits of the digital activity tracking program are concentrated amongst the individuals who had already achieved the healthy biometric ranges.

Implications for Policy: Employers should invest in understanding who participates in EWPs and the extent to which there are favorable selection effects (e.g., healthy individuals more likely to participate), particularly for programs which require regular exercise. Additionally, these findings suggest that financial rewards alone are not enough to impact biometric outcomes among unhealthy employees. Employers may want to consider more holistic strategies to engage employees in wellness activities.