The studies were right. The yardstick was wrong.
The 2019 trials read as proof that workplace wellbeing does not work found something narrower and more useful. The problem was never the evidence. It was the yardstick.
For a decade, corporate wellbeing has carried an asterisk. In 2019, two of the most cited studies of workplace wellbeing programs returned results that sounded like a verdict, and the headlines reduced them to a daunting message. Workplace wellbeing programs do not work.
However, these studies deserve a closer examination. What they actually found is more precise than the verdict and more useful.
The first, published in JAMA, randomly assigned nearly 33,000 employees of a large warehouse retailer, worksite by worksite, to receive a comprehensive wellbeing program or none at all. After eighteen months, the researchers had measured roughly eighty outcomes, from cholesterol and blood pressure to medical spending, absenteeism and job performance. Two metrics improved significantly. Both were self-reported behaviors, more regular exercise and more active weight management.
The second, the Illinois Workplace Wellness Study, randomized access to a comprehensive program for university employees and followed them for thirty months. The authors reported that their confidence intervals ruled out 84% of the savings estimates previously published in the field. A well-designed experiment made most of the field's optimistic math statistically untenable.
Read as written, both results are narrower than the verdict built on them. The JAMA authors concluded that the findings "may temper expectations about the financial return on investment that wellness programs can deliver in the short term." The conclusion is on financial return in the short term. This is the null result's scope, in the authors' own words. Neither study found that employee health lacks business value. They found that a particular kind of program, judged by one yardstick over a short window, did not pay for itself in reduced medical claims.
The yardstick deserves the scrutiny. Return on investment, applied to workforce health, asks whether a program recoups its cost in claims data within a year or two. Almost nothing else a company invests in is asked to clear that bar. A 21-author consensus paper in the Journal of Occupational and Environmental Medicine made this point back in 2014, noting that there are "other yardsticks by which health promotion initiatives should be measured" and observing that health insurance itself, then costing employers more than $16,000 per family each year, has never been required to turn a profit.
The same 2014 paper drew a second distinction the headlines skipped. Some programs are well designed, grounded in evidence and properly executed, and many are none of those things. The randomized trials tested representative programs, which is exactly what made them fair tests, and exactly why their results describe the average program rather than the practice at its best.
So the evidence supports a narrower summary. Programs built to lower medical claims within eighteen months mostly do not. The business value that runs through productivity, engagement, safety and retention sat largely outside the studies' frame. Quoting the null results as a reason to stop investing draws a bigger conclusion than the data allows. Quoting the old optimistic ROI math draws the same oversized conclusion in the other direction.
The better response is a better yardstick. Defining what we measure and why is critical to the success of workplace wellbeing programs, and asking the right questions with appropriate analytics strategies keeps a workforce health initiative funded past its first budget cycle. The null results make that case sharper, not weaker. If the question was wrong, the work is to ask a better one, and to measure the answer with the same rigor these trials brought to the wrong one. That is where this series goes next.