Does your wellbeing program really work, or does it only feel that way?

HR professionals considering a wellbeing program are inundated with promises: reduced absenteeism, higher productivity, and more satisfied employees. However, the scientific literature of recent years, including some of the largest randomized studies ever conducted in this field, paints a more nuanced picture. While this picture is not discouraging for HR professionals, it is essential for selecting a program that actually delivers results.

In short:

• The two largest randomized studies on workplace wellness found improvement in self-reporting, but no measurable effect on blood pressure, cholesterol, healthcare costs, absenteeism, or productivity.

• Objective outcomes (absenteeism data, biomarkers, healthcare costs) and self-reported outcomes (questionnaires) do not always move together, and self-reporting on behavior such as physical activity often overestimates reality.

• The effects of individual programs such as mindfulness are real, but usually do not hold up in measurements taken more than thirteen weeks after completion.

• A large-scale British study of over forty-six thousand employees found no demonstrable benefit for participants for the most common individual interventions.

• Interventions aimed at the organization itself, such as workload and autonomy, show the most consistent effects in the literature.

• Not every source of stress originates in the workplace, so in addition to an organization-oriented approach, room is also needed for the individual situation of the employee.

Two types of outcomes, and why the difference matters

Researchers measure the effect of wellbeing programs in two ways. The first is objective: absenteeism from HR records, healthcare costs, blood pressure, cortisol levels, heart rate variability, or actual measured movement. The second is subjective: what employees themselves fill out on a questionnaire regarding their stress level, well-being, or behavior. Both are valuable, but they do not measure the same thing and do not always move together.

Two of the largest randomized studies on workplace wellness illustrate this sharply. The Illinois Workplace Wellness Study followed nearly five thousand employees for more than two years. The BJ's Wholesale study, published in JAMA, included thirty-three thousand employees across one hundred and sixty locations. In the BJ's Wholesale study, participants more frequently reported exercising regularly or actively managing their weight. In the Illinois study, even that did not change: researchers found no significant effect on self-reported health behavior, only on health beliefs and participation in screenings. Neither study found a measurable effect on blood pressure, cholesterol, healthcare costs, absenteeism, or productivity.

That pattern is no exception. It resurfaces time and again whenever researchers include an objective measure alongside a questionnaire. A recent German study among managers found a clear improvement in self-reported stress perception over six months following stress management training, but no change whatsoever in cortisol or salivary markers at any measurement point. Furthermore, research into self-reported physical activity shows that people systematically overestimate their own activity levels, sometimes by tens of percent compared to what a motion sensor registers. Therefore, part of the reported behavioral change after a program is possibly a more optimistic assessment, not actual different behavior.

Effects fade, and that applies to both types of outcomes.

Even where programs do indeed work, the question is for how long. Meta-analyses of mindfulness programs in the workplace, based on dozens of randomized studies, show a consistent pattern. Immediately after completion, there is improvement in stress, burnout, and well-being. In measurements taken between one and twelve weeks after completion, that effect usually persists. In measurements taken later than thirteen weeks, this is no longer the case.

A large-scale British study of over forty-six thousand employees found no demonstrable benefit for participants compared to non-participants for the most common individual interventions, including mindfulness, resilience training, and wellbeing apps. Only volunteering was associated with slightly higher well-being, and the author of the study herself emphasizes that this relationship may work the other way around: those who already feel better are more likely to volunteer.

What then

All this does not mean that investing in wellbeing is pointless. It means that the way many programs are structured and purchased does not align with what the strongest evidence shows.

Interventions focusing on the organization itself, such as reducing workload, giving employees more autonomy, or making adjustments to tasks and schedules, show the most consistent effects in the literature, particularly on burnout and absenteeism.

However, not every source of stress lies in the workplace itself. Someone with a strong sense of responsibility may take on too much work or continue to worry about it in the evenings, even if the position is perfectly manageable on paper. In psychology, difficulty setting boundaries or saying no is often linked to patterns that originated early in a person's life, long before an employer came into the picture. Drinking too much alcohol, worrying about a child, or anxiety disorders cannot be resolved with a workplace intervention.

For HR, this means that two areas of focus are required: the work context and the individual situation of the employee.

Individual programs such as mindfulness or stress management training do indeed have value, especially for employees who are already struggling with heightened tension. What such a program must meet to truly deliver that value follows directly from what is stated above. It does not measure solely with a questionnaire but combines this with an objective measure, so that you can see whether an improvement is also reflected outside of self-reporting. It does not stop at the end of the intervention but follows up with participants afterwards, because it is precisely during this period that most programs see their effect fade. And the result is greater when it is deployed specifically with employees who actually benefit from it.

This article is the result of a conversation lasting several hours between Claude and Sander Gremmen, using only scientific articles as sources. It is based on, among others, the Illinois Workplace Wellness Study (Jones, Molitor & Reif, Quarterly Journal of Economics, 2019), the BJ's Wholesale study (Song & Baicker, JAMA, 2019), the KMU-GO study on stress management training (BMC Public Health, 2026), meta-analyses of mindfulness programs in the workplace (Michaelsen et al., 2023, and Vonderlin et al., 2020, both in Mindfulness), and the British study by Fleming (Industrial Relations Journal, 2024).