‘Workplace wellness’ doesn’t work

Have you participated in a “workplace wellness program” or known somebody who tried one? Did you notice any difference? If not, a team of enterprising policy analysts in Illinois may have found an explanation. It ran an experiment on employees of the University of Illinois at Urbana-Champaign and recently wrote up the results for the National Bureau of Economic Research.

First, 12,459 benefits-eligible university employees were invited to complete a 15-minute online survey about health and wellness. Roughly 4,800 bothered to fill out the survey, at which point they were randomly divided into seven groups. Six of the groups were offered varying levels of incentives to participate in a wellness program. The seventh group was told it might be contacted for future surveys and was otherwise left alone.

After the program had run its course, researchers compared those who were offered the wellness program with those who were left alone. Thirty-nine wellness-related outcomes were measured, from sick days to health care spending. The researchers found no significant effect in 37 of the 39 outcomes. Two variables showed “significance”: One was “an increase in the number of employees who ever received a health screening,” the paper reports, and the other “an increase in the number who believe that management places a priority on health and safety.”

The result from Illinois contrasts with some of the earlier literature on these programs. A 2010 meta-analysis found that medical costs fell by $3.27 for every dollar spent on wellness programs, while the cost of absenteeism declined by $2.73.

But that earlier literature was vulnerable to a problem known as selection bias: The companies that offer wellness programs, and the employees who use them might not be representative of the population as a whole.

For example, we know that higher-wage employees are more likely to be offered wellness programs. There also can be selection effects between those obnoxiously fit folks who leap at the opportunity to fine-tune their health and those who are too sick or uninterested to participate. That’s what the Illinois study found: Average annual medical spending for those who didn’t participate in the program was about $1,400 higher than for people who participated, even before they joined.

Which is why the study needed to be designed as a randomized experiment. If you look at the results of wellness programs without controlling for the types of people who use them, it can seem as if they’re making people healthier when they’re just scooping up healthy employees and running them through useless hoops. Companies might even see real benefits in terms of health care costs — but only because the incentives offered for wellness-program participation make their workplaces more attractive for healthy people than sick ones.

If selection bias is driving the apparent benefits, then the incentives often offered for participation essentially redistribute wealth and resources upward, from the sick to the healthy and, by extension, from lower-earners to the more affluent. Government policies encouraging wellness programs should be discontinued. But even without government help, companies might decide it is in their interest to keep them going. Government policy has given employers a financial interest in your health, between your health insurance, your workers’ compensation premiums and the liability problems that sick employees can create.

That’s not to say employers are cynically using the programs to discriminate against the unhealthy. Selection bias is seductive that way: It can make useless or even counterproductive interventions look as if they’re working spectacularly. But it does mean that no matter what future research shows, employers will probably find reasons to keep these programs going. Because even if they don’t produce any actual wellness, they might still be very healthy for the bottom line.

Megan McArdle is a Washington Post columnist.