Do Workplace Wellness Programs and Wearables Actually Work?

Not on the outcomes employers usually buy them for. In the Illinois Workplace Wellness Study, which randomly assigned nearly 5,000 university employees, the program raised health screening rates but had no significant effect on medical spending, health behaviors, productivity or self-reported health after more than two years. In a trial across 160 BJ's Wholesale worksites and 32,974 employees, more people reported exercising and managing their weight, with no significant change in 10 clinical measures, 38 spending and use measures, or absenteeism, tenure and job performance after 18 months. Neither program used wearables. Trials that did add a tracker found step counts rose while an incentive or game ran and fell back afterwards, and the largest of them found no improvement in weight, blood pressure or any other health outcome at 6 or 12 months. No independent study has tested any vendor's readiness, recovery or stress score against a workplace outcome. The savings and sick-day figures on vendors' employer pages rest on member surveys, an economic model, unpublished data or no stated source.

What the largest workplace wellness trials found

Randomized trials are rare in this field because most employers cannot assign some staff to a program and keep others out. The two that did it at scale ran in the US and reported in 2019.

The Illinois Workplace Wellness Study offered a comprehensive program, with screening, a health risk assessment and activities, to randomly chosen University of Illinois employees and paid some of them to take part. People who signed up already had lower medical spending and healthier habits in the year before the program began. That is the selection problem in one finding: comparing participants with non-participants makes a program look effective before it has done anything. Against the randomized control group, the program raised screening rates and moved nothing else measured, and its confidence intervals ruled out 84 percent of earlier estimates of savings on medical spending and absenteeism (Quarterly Journal of Economics, 2019). The 24-month follow-up of 4,834 employees found no significant effect on biometrics, diagnoses or medical use, while more people reported having a primary care physician (JAMA Internal Medicine, 2020).

The BJ's Wholesale trial randomized whole worksites. Twenty received eight modules on nutrition, physical activity and stress led by registered dietitians; 140 received nothing. Participation in surveys and screenings at the program sites ran from 36.2 to 44.6 percent, and the average employee completed 1.3 modules. Regular exercise was reported by 69.8 percent at program sites against 61.9 percent at control sites, and the program had no significant effect on any of the other outcomes it measured (JAMA, 2019).

The favorable figures that circulate in this field come from earlier, mostly observational work. A meta-analysis estimated that medical costs fell about 3.27 dollars and absenteeism costs about 2.73 dollars for every dollar spent (Health Affairs, 2010). A survey of US employers for the Departments of Labor and Health and Human Services found employers overwhelmingly believed their programs saved money, and only 2 percent had an estimate of the savings (RAND, 2013).

What happened when the program included a wearable

The largest workplace trial of a tracker gave a Fitbit Zip to 800 employees of 13 organizations in Singapore, with or without cash or charity incentives. At 6 months the cash group did 29 more minutes a week of moderate to vigorous activity than controls, and the tracker alone made no significant difference. At 12 months the pattern reversed: the tracker-only and charity groups were ahead and the cash group was not, after the incentives had stopped. No group improved on weight, blood pressure or any other health outcome at either point, which the authors said calls into question the value of these devices for health promotion (Lancet Diabetes and Endocrinology, 2016).

In the US, a trial enrolled 602 adults with overweight or obesity from 40 states through Deloitte and tracked their steps with wearable devices for a 24-week game and 12 weeks of follow-up. Competition added 920 steps a day over a control group that received device feedback only, support 689 and collaboration 637. During follow-up only the competition arm stayed significantly ahead, at 569 steps, and activity was lower in every arm than during the game (JAMA Internal Medicine, 2019). Among 146 sedentary US office workers who all wore a Fitbit Zip, adding a gamified app raised steps by 2,092 a day over the tracker alone across 10 weeks, an effect the authors reported decreased over time (Journal of the American Heart Association, 2018).

At Volkswagen, 129 employees with metabolic syndrome who had finished a six-month exercise program were randomized to a wearable with personal consultations, a wearable alone, or neither for six months of maintenance. Physical activity fell in all three groups with no difference between them (BMC Sports Science, Medicine and Rehabilitation, 2022).

A Cochrane review of 14 randomized trials and 4,762 employed adults, all testing pedometer-based workplace programs, found pedometers may have no effect on physical activity measured at least a month after the program, on very low certainty evidence. Body mass index changed by -0.64 (95 percent CI -1.45 to 0.18) and systolic blood pressure by -0.08 mmHg (-3.26 to 3.11). Only one of the 14 trials was at low risk of bias (Cochrane Database of Systematic Reviews, 2020). The broader evidence on whether trackers make people more active outside work is covered in whether wearing a fitness tracker makes you more active and whether fitness trackers improve health outcomes.

Why the step gains fade

Use of the technology drops off. A systematic review of 25 workplace studies of wearable activity monitors and smartphone apps found 14 reported a significant increase in physical activity and observed a large decline in technology use and engagement over time. Only 1 of the 25 studies was rated strong quality (Digital Health, 2019).

Even in a study built to collect data, wear was incomplete. Over nine months, 239 office and production workers wore devices with a median data-collection adherence of 61.9 percent, lower at production sites than in offices (Digital Health, 2026). A program that reports its average step count is reporting on the employees who kept wearing the device, who are not a random sample of the workforce.

Can a wearable predict or prevent burnout?

Not on current evidence. KnowYourPrime's HRV research desk rates whether higher occupational stress is linked to lower heart rate variability as Emerging, on six studies. A meta-analysis of continuous monitoring in doctors found HRV was lower during stress than during recovery periods, with RMSSD lower by a standardized mean difference of 0.63, in studies of 12 to 54 participants whose quality the authors rated moderate at best (Occupational Medicine, 2025). The one study that tested the link within the same people across a working day, teachers rating their demands six times a day, found the effect on exhaustion and not on HRV (Anxiety, Stress and Coping, 2020).

That difference decides what a workplace deployment can claim. An association that appears when groups of workers are compared supports a statement about a population. A burnout alert is a prediction about one person, and no study found here validates one. A scoping review of 10 studies in which health care workers wore devices, nine of them wrist-worn, found no physiological measure was reliably associated with burnout or anxiety; heart rate and HRV tracked acute stress (Journal of Medical Internet Research, 2024).

One randomized trial points the other way, with a caveat. Among 184 physicians at two US academic centers, six months of wearing a smartwatch with access to its sleep, step and heart rate data was followed by burnout in 41.2 percent against 50.5 percent of controls, a difference that was not significant unadjusted (P = .21) and was significant in the adjusted model (odds ratio 0.46, 95 percent CI 0.21 to 0.99) (JAMA Network Open, 2025). The watch was not used to predict anything; the trial tested whether wearing one changed how physicians fared. What stress scores themselves measure is covered in whether a wearable stress score measures stress.

Readiness, recovery and other vendor scores in an employer program

Vendor scores are where the evidence is thinnest. Readiness, recovery, strain, stress and energy scores are proprietary composites, their weighting is not published, and KnowYourPrime's recovery desk rates whether consumer readiness and recovery scores measure what they claim as Uncertain, on five studies. That rating applies across the category. No independent study found here has tested any vendor's score against absenteeism, burnout, health care cost or any other workplace outcome; see whether recovery and readiness scores are validated and how much a recovery score moves from day to day.

The signals underneath the scores are a different matter, and they vary. Heart rate from the wrist is well tested: the activity desk rates wearable heart rate accuracy as Established on 61 studies. A chest strap matches an ECG for HRV, rated Established, while wrist and ring HRV is Uncertain on 30 studies. Sleep duration and stages are Uncertain, and calorie estimates are Uncertain, with errors above 30 percent for every brand in one large review (what the calorie number means). Step counting undercounts badly for people who use a walker or cane (mobility aids and step accuracy), which matters for any step target set for a whole workforce.

What the published evidence says about each brand's scores is on its own page: Oura, WHOOP, Garmin, Apple Watch, Fitbit, Samsung, Ultrahuman, RingConn, Amazfit and Polar.

What the wearable brands offer employers, and what they claim

Five of the ten brands checked sell a program to organizations. Oura for Organizations offers dashboards, an enterprise API and data export. WHOOP Unite gives an organization a people hub and an insights area. Google Health Enterprise, formerly Fitbit Enterprise, supplies devices and lets the employer build aggregated dashboards. Polar for Business offers the screenless Polar 360 and an API for the employer's own wellness system, and Ultrahuman for Companies describes an enterprise dashboard for stress, recovery and wellbeing. Garmin supplies devices and API access and leaves reporting to partner platforms. Samsung sells devices to businesses without a wellness platform, Apple has no employer program and supports third-party ones, and no employer program was found for RingConn or Amazfit.

What an employer can see differs and is often not stated. WHOOP is the only brand whose documentation describes the choice: each member picks individual data or team-average data at invitation, and WHOOP labels individual data as recommended. Oura describes configurable de-identification and role-based permissions without saying what the default is. Google describes aggregated dashboards. The Garmin, Polar and Ultrahuman employer pages say nothing about whether an employer sees an individual employee's data.

The outcome figures on these pages do not come from randomized trials. Oura's employer page reports that 90 percent improved overall health and 79 percent improved work productivity, from internal data and a survey of 474 Oura members in their first 30 days. Google reports 6.5 million dollars in potential savings over two years for 8,000 participants, a figure from an economic model of a hypothetical insured cohort posted as a preprint, along with engagement and HRV figures it footnotes as unpublished internal data. Polar's corporate page lists a 56 percent reduction in sick days and a threefold reduction in health care costs with no source attached. Ultrahuman's partners page, which carries its offer for companies, says monitoring HRV, heart rate and sleep can reduce injury or burnout risk over time, also without a source. Set against the randomized trials above, none of these figures measures what a program did to employees compared with employees who did not get it.

What a savings claim has to overcome

Self-selection. The Illinois study showed that the employees who join a wellness program are already the healthier, lower-spending ones, so any comparison of joiners with non-joiners credits the program with a difference that existed before it began. The randomized result was null, and it ruled out most earlier estimates.

The one wearable-specific cost study found here has that design. A health plan's wearable-linked rewards app was matched against 56,816 non-enrollees and associated with about 10 dollars per member per month lower medical spending, among people enrolled for at least seven months who chose to join (Journal of Medical Internet Research, 2023). The authors were employed by Aetna and CVS Health, which ran the program. No randomized trial found here measured the effect of a wearable program on medical spending, sick leave or absenteeism.

What this does not settle

Neither large randomized trial of workplace wellness used wearables, and the wearable trials are smaller, shorter and focused on steps. The biggest ran in Singapore, the burnout trial enrolled physicians rather than a general workforce, and none followed employees for more than two years. A program built differently, run for longer or aimed at a narrower group could produce a different result, and none of the evidence here tests the newest devices or scores. What it does establish is that participation, step counts and engagement are not evidence of better health or lower cost, and that the trials designed to find savings did not.

Common questions

KnowYourPrime Evidence Registry: Finding KYP-ER-0005: Readiness and Recovery Score Validity · Finding KYP-ER-0006: Work Stress and HRV · Finding KYP-ER-0012: Do Activity Trackers Make People More Active?