There is no conflict of interest.
Highlights
- The study's objective was to examine the impact of a workplace wellness program on health and employment outcomes.
- The study used a randomized controlled trial to assign individuals to the workplace wellness program (treatment group) or the control group. Using survey and administrative data, the authors conducted statistical models to compare the outcomes of the treatment and control groups before and after the intervention.
- The study did not find statistically significant relationships between participation in the workplace wellness program and self-reported health, clinical measures of health, or employment outcomes.
- This study receives a low evidence rating. This means we not confident that any estimated effects would be attributable to the workplace wellness program; other factors would likely have contributed. However, the study did not find any statistically significant effects.
Features of the Intervention
Workplace wellness programs aim to reduce healthcare costs, enhance employee health, and boost productivity. The program examined in this study was created and implemented by Wellness Workdays, an established wellness vendor. The workplace wellness program ran for three years and included twelve modules, each lasting four to eight weeks. The modules covered various topics such as physical activity, stress management, and nutrition, and were led by registered dietitians through individual and team activities and challenges. Participants received small rewards for completing the modules, totaling around $300 for the entire program. This initiative was for employees at a large warehouse retail company. While all employees at the participating worksites could join, participation was not mandatory.
Features of the Study
The study used a randomized controlled trial design. A workplace wellness program was introduced in 20 out of 160 worksites that had similar health insurance coverage. These 20 worksites were randomly selected for the program. The control group included 140 worksites, with 20 serving as primary controls involving in-person data collection, and 120 as secondary controls collecting administrative data. After 18 months, five additional worksites from the secondary controls were randomly assigned to the treatment group, and another five to the primary control group. Employees were categorized into treatment or control groups based on their worksite status at the time of randomization or upon starting employment. The sample consisted of 7,288 employees at 25 treatment worksites, 7,377 employees at 25 primary control worksites, and 33,999 employees at 110 secondary control worksites. Each site had an average of 116 employees at any given time.
About half of the employees were full-time; 25% identified as Black and 17% as Hispanic. In-person data collection was voluntary and included a survey of self-reported health and behaviors and clinical biometric screening by a nurse. In-person data collection occurred at two stages: halfway through the study (after 18 months) and at the end of the study (after 3 years). Administrative data collection for employment outcomes was ongoing. Health outcomes included self-reported health (summary scores for physical health, mental health, unmanaged stress, unmanaged depression, and stress at work) and clinical measures of health (total cholesterol, glucose, blood pressure, and body mass index). Employment outcomes included absenteeism (percentage of scheduled hours missed) and tenure (days employed during the treatment period). The authors used statistical models to compare the outcomes of treatment and control group members.
Findings
Health and safety
- The study found no statistically significant relationships between participation in the workplace wellness program and self-reported health or clinical measures of health.
Employment
- The study found no statistically significant relationships between participation the workplace wellness program and absenteeism or tenure.
Considerations for Interpreting the Findings
During the study, when more funding became available, the authors randomly selected five sites from the secondary control sites to become treatment groups and five sites from the secondary control sites to become primary control sites. This compromised randomization. Also, the authors used statistical models that applied weights to account for differences in age, sex, and race/ethnicity between the treatment and control groups but did not account for pre-existing differences in health and employment outcomes prior to randomization. Therefore, the study receives a low causal evidence rating.
Causal Evidence Rating
Research Guidelines
Review Protocol: Living Systematic Annual Search and Review Protocol
Review Guidelines: Causal Evidence Guidelines