Abstract
Dr Goetzel’s comment on the Illinois Workplace Wellness Study and the iThrive wellness program makes 3 main points: a different wellness intervention or longer time horizon might produce larger effects, the study didn’t address cost-effectiveness, and statistical methods other than randomized controlled trials (RCTs) are better suited to evaluating wellness programs.
The iThrive wellness program was designed to be representative of typical comprehensive wellness programs offered by employers. The wellness program ran for 2 years, a time horizon over which prior studies have argued effects can emerge. 1 Yet after 30 months, medical spending, productivity, health behaviors, and self-reported health did not significantly differ between the treatment and control groups. Dr Goetzel argues that a higher participation rate, a more intensive intervention, and a longer follow-up might have produced larger effects. We completely agree that a different intervention might have produced different results, and in ongoing work, we are evaluating the effects of the program over a longer time horizon.
Dr Goetzel also writes that our “article does not discuss how much iThrive costs,” making it “hard to determine whether the program realized a return on investment (ROI).” In fact, our article describes the cost of the program ($271 per participant, on average) and calculates the implied ROI. 1 As we discuss in the introduction of the paper, the ROI for iThrive rules out a widely cited ROI for these programs. 2
Finally, Dr Goetzel argues that statistical methods such as multivariate regression are better suited than RCTs to evaluate workplace wellness programs. We disagree. Our study underscores the value of using RCT designs to evaluate workplace wellness programs by demonstrating that an observational regression analysis of our data yields incorrect results, even after we adjust for selection bias using a comprehensive set of demographics and baseline risk factors. 1,3 A lesson from the Illinois Workplace Wellness Study is that research design matters and that more researchers should use randomized evaluations to study workplace wellness programs.
