The Washington PostDemocracy Dies in Darkness

Workplace wellness programs are big business. They might not work.

One of the most rigorous studies yet finds little impact on health — or expenses.

A doctor checks a patient’s blood pressure. (iStock)

When we get through the coronavirus crisis, we will continue to confront challenges to our health — and health-care system — posed by the epidemics of heart disease, diabetes and obesity. “Workplace wellness” programs have been enthusiastically embraced as a potentially cost-effective way to promote healthy diets, regular exercise and other behaviors that help keep such diseases in check.

These programs seem to make sense: Many people spend most of their waking hours on the job, so that’s a natural setting to try to promote changes in behavior. And employers have their own stake in improving employees’ health; many of them provide benefits like health insurance and sick leave, and workplace wellness programs could result in lower insurance costs as well as a healthier, more productive workforce. All of this has spawned a multibillion-dollar industry. In fact, most employers as of 2020 had wellness programs of some kind — including 53 percent of small firms (those with three to 200 employees) and 81 percent of large companies.

Some studies have suggested that these programs indeed result in better worker health and lower health-care costs. But how reliable are these results? A new study we conducted, with the cooperation of BJ’s Wholesale Club, was sobering. We randomly assigned some BJ’s worksites to adopt a new wellness program and tracked employees at sites with and without such programs. After nearly three years, we didn’t see any substantial effects on employment outcomes (such as fewer sick days), health-care spending or objective health measures. We did find some improvements in self-reported health behaviors. For example, the wellness program increased the share of employees who reported actively managing their weight by about 8.7 percentage points and the share of those who said they got regular exercise by about 11 percentage points. But we saw no effects on blood pressure, diabetes or obesity.

The key element of this study is the randomization. It’s easy to be misled by a casual examination of the health and employment outcomes of people who participate in wellness programs compared with those who don’t. For example, when an employer offers a free gym membership, people who take up the offer unsurprisingly tend to be people who like to go to the gym. People who use this employment benefit might well be healthier than those who don’t, but that probably would have been the case regardless of the wellness program.  

Many studies have used more sophisticated approaches to try to eliminate this kind of confounding factor. Some have used statistical tools to adjust for differences between participants and non-participants; others have used a different group of employees without a wellness program as a comparison. These are better approaches, but participants may still differ from non-participants in unobserved ways — having different health profiles or behaviors — making it difficult to sort out the effect of the wellness program itself.

Our partnership with BJ’s allowed us to use a truly randomized controlled trial design. BJ’s Wholesale Club employs roughly 32,000 people across some 220 stores in the United States. Working with the vendor Wellness Workdays, we implemented a multifaceted wellness program for employees at 25 randomly selected BJ’s stores, while monitoring workers in 135 randomly chosen stores as a control group. The program ran from early 2015 into 2017 and had different modules — consecutive mini-programs focused on exercise, healthy eating, health education and stress reduction. Some modules, for example, had group fitness challenges, others had nutrition counseling and healthful-cooking demonstrations, and others had meditation instruction. Registered dietitians administered each of these six-to-eight-week modules on-site, management encouraged attendance via email and other means, and participation came with $25 to $50 gift cards as an incentive to join. About 28 percent of employees in worksites with the program participated in at least one module.

We then gathered a wide range of data, including employment outcomes (like job tenure and absenteeism), self-reported health behaviors (like exercise and healthy eating), health-care spending (for those insured through BJ’s) and health outcomes (like blood pressure, blood sugar, diabetes and obesity). Again, the only changes related to wellness programs involved workers’ paying somewhat more attention to staying in shape. Health-care costs, underlying health measures and job-performance-related data were indistinguishable. (There was limited evidence that people in wellness programs got more of the medical tests recommended by their physicians, but the finding was not robust.)

These results build on our earlier analysis of the same program after roughly 18 months, when we saw similar effects (or lack thereof). One might have anticipated that the changes in behavior might lead to health improvements after three years, but that did not seem to be the case.

Partly because there’s such a huge array of wellness programs in use across a wide range of workplace settings, no one study can speak definitively to whether such programs “work.” There may be more effective program designs than the one we examined, or those modules might work better in other workplaces. That said, our study adds to a growing body of rigorous evidence that suggests tempering expectations for these interventions. Another trial that made use of a “random assignment” design, for example, found similarly limited effects. That study, based at the University of Illinois at Urbana-Champaign, randomly made some employees eligible for the program. (The design differed from ours by randomizing individual employees within a worksite, rather than entire worksites.) After two years, employees in the treatment group were more likely to have a primary-care physician and believed that they were healthier — that they had lower weight, blood pressure, cholesterol, etc. But in reality, researchers found no detectable changes in these physical health measures. Nor were there differences in medical diagnoses or in health-care use.

Does this mean that wellness programs are a waste of money? It depends on what you want to get out of them. Employees seemed to value the benefit, had heightened awareness of the importance of healthy behaviors and were trying to implement them. If employers are seeking to add benefits that workers value (or attract workers who value those benefits), the programs may be worth it. But if the goal is to save money by reducing health-care costs and absenteeism, or to improve chronic physical health conditions, the evidence so far is underwhelming. Whatever their positive effects, wellness programs don’t seem to be doing much to alleviate the nation’s persistent health problems.

Read more from Outlook:

Follow our updates on Facebook and Twitter.