The Impact of Employee Wellness Programs

Robert Wood Johnson Foundation, National Institute on Aging (NIA), Abdul Latif Jameel Poverty Action Lab, University of Chicago

There is great public and private interest in the use of workplace wellness programs to reduce health care spending, improve health outcomes, and enhance productivity for employees. However, there is little rigorous evidence on the effects of wellness programs. This study partners with a large multi-state U.S. employer (BJ's Wholesale Club) and an experienced wellness vendor (Wellness Workdays) to evaluate a multi-prong workplace wellness program, including components such as nutrition counseling, fitness challenges, and stress management workshops. The wellness program will be delivered by a team of experts including Registered Dieticians, and will include financial rewards for participation. The program will be available to employees in initially 20 of BJ's 200 worksites, and later expanded to 25 worksites. These worksites have been randomly selected, allowing a randomized controlled trial evaluation of the effects of the wellness program. Data will be collected on a wide array of outcomes from multiple sources, including on-site biometric screenings and surveys, employment records, and health insurance claims for employees at both treatment and control worksites.

Employees at treatment worksites will be given access to workplace wellness programming. Participation by employees will be voluntary, but all employees at treatment sites will be considered as part of the treatment group. Employees will also be invited to complete on-site biometric assessments and questionnaires. Data from secondary data sources (including employment records and health insurance claims) will be collected for employees at all BJ's worksites.

Employees at primary control worksites will be invited to complete on-site biometric assessments and questionnaires, but will not have access to the workplace wellness programming. Data from secondary data sources (including employment records and health insurance claims) will be collected for employees at all BJ's worksites.

Employees at secondary control worksites will not participate in in-person screenings or questionnaires, and will not have access to workplace wellness programming. Data from secondary data sources (including employment records and health insurance claims) will be collected for employees at all BJ's worksites.

Multi-prong workplace wellness program, with components such as nutrition counseling, fitness challenges, and stress management workshops, including supports and incentives.

Regular exercise, number of hours sitting per day, and actively managing weight, as indicated in responses to survey questions

Dollars spent on health care for employees covered by employer-sponsored insurance, from claims records

Dollars spent on health care and prescription drugs for employees covered by employer-sponsored insurance, decomposed into spending on inpatient, outpatient, emergency department, and prescription drugs, from claims records

Health care encounters and prescription drugs for employees covered by employer-sponsored insurance, decomposed into inpatient, outpatient, emergency department, and prescription drugs, from claims records

Affirmative engagement in active efforts to improve nutrition, as indicated in responses to survey questions

Short form 8 (SF-8) physical and mental summary scores, as indicated in responses to survey questions. Each question of the SF-8 uses a 5- or 6-point Likert scale. Its standardized scoring system combines responses into a score that can be interpreted as a continuous variable (analogous to a 0-100 scale), with higher scores denoting better self-reported health.

Inclusion Criteria: Employed in one of BJ's clubs during the period of the intervention Exclusion Criteria: Not employed in one of BJ's clubs during the period of the intervention

Song Z, Baicker K. Effect of a Workplace Wellness Program on Employee Health and Economic Outcomes: A Randomized Clinical Trial. JAMA. 2019 Apr 16;321(15):1491-1501. doi: 10.1001/jama.2019.3307. Erratum In: JAMA. 2019 Apr 17;: