The Impact of Employee Wellness Programs: A Randomized Controlled Trial

How to cite

Song, Zirui; Baicker, Katherine. 2015. "The Impact of Employee Wellness Programs: A Randomized Controlled Trial." AEA RCT Registry. AEARCTR-0000586. https://doi.org/10.1257/rct.586-6.6

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Zirui Song* , Katherine Baicker

* Primary investigator

Registry ID
AEARCTR-0000586
DOI
https://doi.org/10.1257/rct.586-6.6
First registered
Last updated
Status
On going
Trial dates
2015-02-09 – 2021-06-30
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/586

Abstract

Health care costs are a major concern for policymakers, patients, and employers. Many employers across the United States are implementing or considering employee wellness programs in an effort to control costs while improving employee productivity and health. The Affordable Care Act also established a variety of incentives for employee wellness. Despite the popularity of wellness programs, little rigorous evidence exists on their effects on health, productivity, and health care costs or utilization. Such evidence is crucial for employers and policymakers alike.

We will evaluate the effect of a year-long employee wellness intervention at a large multi-state firm in the Eastern United States, using a randomized controlled trial design. The wellness intervention will consist of several components, including team-based wellness challenges, nutrition counseling, stress reduction, and individual-focused physical activity. The intervention has been designed with and will be fielded through an experienced and large-scale vendor.

Randomizing this intervention across a set of treatment and control sites, we will evaluate its impact on 5 categories of outcomes gathered from both primary and administrative data sources: (1) self-reported health and well-being, (2) biometrics (such as body mass index, blood pressure, cholesterol, and blood sugar), (3) worker absenteeism and turnover, (4) health care spending and utilization, and (5) return on investment. Better information about the effectiveness of such programs should inform both employer investment and public policy.

Intervention

We will evaluate the effect of a two and a half year employee wellness intervention at a large multi-state firm. The intervention is based on a multi-part platform implemented by an established vendor, including a pre-intervention and post-intervention personal health assessment along with regular screenings. The staged program components include individual nutrition coaching, disease management, stress-reduction (using behavioral health coaching), and a team-based wellness challenge (capitalizing on social supports to improve diet and exercise).

Experimental design

The study design centers on the random assignment of employer worksites to treatment and control arms. The wellness intervention will be implemented in treatment arm sites, while control arm sites will operate as usual. Primary data will be collected from all employees (program participants and non-participants) at both treatment and control arm sites in parallel. This random assignment will allow causal inferences about the effect of the intervention on a range of outcomes.

Design details

Randomization method
The employer has approximately 200 worksites. We will first eliminate the fewer than 20% of worksites that are geographically difficult to reach for researchers or have atypical insurance coverage. We then randomly assigned 25 of the worksites to the treatment arm and 25 to the control arm for primary data collection. Additional secondary data will also be collected from all worksites.
Randomization unit
Worksite
Clusters
50 worksites for primary data and secondary data (25 treatment and 25 control) 120 worksites for additional secondary data
Observations
Primary data: 14,000 employees; Secondary data: additional 34,000 employees.
Arms
Primary data will be collected from 50 randomly assigned worksites (25 sites in the treatment arm and 25 sites to the control arm), with about 100 employees at each site at any given time. Additional data will be collected from secondary sources for these 50 plus approximately 120 additional sites.
Primary outcomes
Summary of outcome measures 1. Self-Reported Health and Behavior Source: Surveys Outcomes: Self-reported physical and mental health; anxiety and stress; satisfaction with care; life satisfaction and happiness; level of physical activity; healthful behavior; disease diagnosis and management 2. Clinical Health Source: Surveys Outcomes: Mental health (depression, anxiety); Diagnoses Source: Biometric Assessments Outcomes: Blood assays (total & HDL cholesterol, HbA1c); Blood pressure; BMI (height and weight) 3. Employment and Productivity Source: Employment Records Outcomes: Productivity (absenteeism; job performance evaluations; promotion), Employment (hours worked; wages/salary; retention and employment tenure) Source: Surveys Outcomes: Satisfaction with employment; on-the-job stress 4. Health Care Utilization and Spending Source: Insurance Claims Data Outcomes: Spending and utilization overall and by category of care (inpatient, outpatient, emergency department, prescription drugs, preventive); and condition (chronic diseases, mental health) 5. Return on Investment Comparison of complete program costs to monetized benefits

Cited by

No citation data available for this registration yet.

Keywords