Illinois Workplace Wellness Study

Our treatment consists of a baseline survey among all control and treatment group members. Treatment group members will then be offered the opportunity to participate in a biometric screening and health risk assessment (biometric screening + HRA), and -- conditional on completing an biometric screening + HRA -- up to two, semester-long wellness programs (e.g. walking challenge, active living class, smoking cessation hotline). Treatment group members will be offered varying levels of cash reward for completing the biometric screening + HRA ($0, $100, $200), and an additional cash reward for completing each wellness activity ($25, $75), for a total of 6 treatment cells. An additional treatment group will feature a clustered treatment, based on self-reported peer groups measured in the baseline survey. Follow-up surveys and biometric screenings will be administered one year later, among a subset of control and treatment group members.

2016-08-08

2017-12-31

(1) biometric screening and HRA completion; (2) wellness activity participation and completion; (3) health utilization based on medical claims data; (3) absenteeism based on human resources (HR) data; (4) biometric health screening data; (5) health risk assessment data; (6) self-reported health status on baseline survey; (7) self-reported health status on follow-up survey; (8) expectations of health status prior to screening

Medical claims data: total costs, costs to employer, inpatient visits, outpatient visits, ER visits, diagnosis codes
Biometric data: height, weight, waist circumference, resting blood pressure, total cholesterol, total cholesterol ratio, HDL cholesterol, LDL cholesterol, triglycerides, glucose levels
Health Risk Assessment data: self-reported wellness, health status, nutrition, healthy activities, desire to improve health, preventative health measures
Baseline Survey data: self-reported health status, healthy activities, preventative health measures
Health expectations: expected weight, height, cholesterol, glucose, blood pressure, body mass index

We will draw a sample of all benefit-eligible employees at the university (approximately 12,000). An invitation will be sent to each employee, asking them to participate in a baseline survey. Employees will receive a gift card of $30 for completing the survey. Respondents to the baseline survey will comprise our core sample.

The core sample will be divided randomly into either a control group, or one of 4 treatment groups, Treatments A - D. Treatment groups A, B and C will receive a cash incentive of $0, $100, and $200, respectively, for completing the biometric screening + HRA. Within each group, half of the members will also be offered $25 for each of up to two, semester-long wellness activities that are completed during the year. The other half will be offered $75 for completing up to two, semester-long wellness activities during the school year.

The final treatment group, D, will feature a clustered design. The baseline survey will elicit information about workplace social networks. We will use standard methods to map those networks among those employees that do respond to the baseline survey. Provided that we are able identify a sufficient number of non-overlapping, local neighborhoods, we will systematically assign a share of neighborhood members to either the control group or the treatments of groups A, B and C with equal probability. Treatment group D will feature a higher share of control group members in a given social network than members of treatment groups A, B and C.

Upon completion of the baseline survey and treatment group assignment, employees will be invited to schedule a biometric screening. Once the biometric screening is completed, an online HRA will be made available for participants. Following completion of the HRA, participants will be allowed to enroll in up to two wellness activities.

One year later, members of the control and treatment groups will be invited to complete an additional follow-up survey and biometric survey. Incentives will be increased in the follow-up period in order to minimize attrition. Participants will be offered either $25 or $35 dollars for completing the follow-up survey, and either $150 or $250 for completing a follow-up biometric screening. In order to minimize attrition, these incentives may be modified based on information gained from the first year of the experiment.

Our ability to implement the peer-effects design, i.e. treatment group D, depends on the structure of social networks in our data. We cannot observe these prior to collecting baseline survey data. Once we have done so, we will be able to do preliminary power analysis. In the event that including a peer effects group significantly reduces our statistical power, we will eliminate treatment group D, and focus on an individual level design.

Randomization will be done on an office computer using Stata's pseudo-random number generator with a preset seed of 11.

Our peer effects design will randomize treatment at the self-reported social network level.
For the rest, treatment will be randomized at the individual level, and stratified by age, gender, and employee class.

Yes

no peer effect design: 6,000 employees
peer effect design: 5,600 employee/social networks

6,000 employees

no-peer effect design: 2,700 employees control; 1,100 employees treatment group A; 1,100 employees treatment group B; 1,100 employees treatment group C
peer effect design: 2,300 employees control; 900 employees treatment group A; 900 employees treatment group B; 900 employees treatment group C; 200 social networks treatment group D

Participation MDE: - 5.7 percentage points (baseline 50% participation) when comparing groups B + C to group A - 6.6 percentage points if we are separately comparing group B to group A, and group C to group A. Selection Regression MDE: approximately 3.8 percentage points, for an attribute that is held by 40% of the population. Short-Run Effects Regression MDE: - 0.61 for BMI (mean = 26.3%, SD = 3.9%) - 1.21 for Systolic blood pressure (mean = 120, SD = 7.5) - 0.97 for diastolic blood pressure (mean 80, SD = 6) - 0.23 for absenteeism (mean = 4, SD = 2)

Analysis Plan

Workplace wellness programs cover over 50 million U.S. workers and are intended to reduce medical spending, increase productivity, and improve well-being. Yet limited evidence exists to support these claims. We designed and implemented a comprehensive workplace wellness program for a large employer and randomly assigned program eligibility and financial incentives at the individual level for nearly 5,000 employees. We find strong patterns of selection: during the year prior to the intervention, program participants had lower medical expenditures and healthier behaviors than nonparticipants. The program persistently increased health screening rates, but we do not find significant causal effects of treatment on total medical expenditures, other health behaviors, employee productivity, or self-reported health status after more than two years. Our 95% confidence intervals rule out 84% of previous estimates on medical spending and absenteeism.

Damon Jones, David Molitor, Julian Reif, What do Workplace Wellness Programs do? Evidence from the Illinois Workplace Wellness Study, The Quarterly Journal of Economics, , qjz023, https://doi.org/10.1093/qje/qjz023

https://doi.org/10.1093/qje/qjz023