Section 01The Mental-Health Spending Went Up. The Burnout Numbers Went Up Too.
Between 2019 and 2025, U.S. employers increased spending on workplace mental-health programs roughly 4.5×, from an industry-estimated $11B in 2019 to north of $50B in 2025 [1]. Over the same six years, the share of U.S. workers reporting burnout, on the most widely used clinical instruments, rose from 43% to 54% [2]. The most expensive workplace mental-health buildout in human history is correlated, almost exactly, with the worst sustained burnout numbers in the available record.
This is the paradox the second half of the 2020s has to solve. It is not solved by spending more. It is solved by understanding why the spending has not worked — and which interventions, much cheaper than the current ones, do.
Section 02What Burnout Is, According to the Person Who Defined It
Christina Maslach is the U.C. Berkeley psychologist whose 1981 instrument, the Maslach Burnout Inventory, has been the standard measure of workplace burnout for the four decades since. Maslach's definition has not changed materially in those four decades — but the way it is used in corporate health programs has drifted.
Burnout, in Maslach's construct, has three dimensions:
- Emotional exhaustion — feeling drained, depleted, unable to recover with normal rest
- Depersonalization (or cynicism) — emotional distance from the work, the clients, the colleagues; treating others as objects rather than people
- Reduced personal accomplishment — a falling sense of competence and meaning at work
The critical point in Maslach's lifetime of research — repeated, with growing impatience, in her recent work with Michael Leiter — is that burnout is a workplace condition, not an individual one [3]. The 2019 WHO ICD-11 codified it the same way: burnout is "an occupational phenomenon" resulting from "chronic workplace stress that has not been successfully managed." It is not a clinical illness inside the worker. It is a state produced by the working environment.
Burnout is fundamentally a result of an imbalance between job demands and job resources. It is not a flaw in the worker. It is a structural condition of the work. Christina Maslach & Michael Leiter, The Burnout Challenge (Harvard University Press, 2022)
This matters because where you locate the problem determines what you spend money on. If burnout is a workplace condition, you fix it by redesigning the work. If burnout is an individual condition, you fix it by giving workers tools to manage stress better. The corporate health industry, for reasons of scalability and revenue model, has spent the last decade selling the second framing. The first framing is what the research actually supports.
Section 03The Six Workplace Conditions That Produce Burnout
Maslach and Leiter's research identifies six job-design dimensions on which mismatch produces burnout. The names are unsexy. The diagnostic value is enormous.
| Dimension | What mismatch looks like | The intervention that addresses it |
|---|---|---|
| Workload | Sustained demand exceeds sustainable supply; recovery time between sprints is structurally insufficient | Workload caps; staffing levels; structural recovery time between intense periods |
| Control | The worker has no meaningful authority over how, when, or what the work is | Decision rights expanded into the worker's role; reduced micromanagement; predictable schedules |
| Reward | Financial, social, or intrinsic reward is misaligned with effort or outcome | Compensation calibration; recognition that ties to outcomes the worker actually produced |
| Community | The team relationships are conflictual, isolating, or absent | Team design; manager training; explicit conflict-resolution mechanisms |
| Fairness | Workload, opportunity, evaluation, or pay are perceived as inequitable | Transparent compensation; transparent promotion criteria; explicit fairness audits |
| Values | The work asks the worker to do things their values reject | Mission clarity; mechanisms for worker dissent and ethical objection |
The list above is the diagnostic. If a workforce is burned out, the right question is not "what mental-health resources have we deployed" but "which of these six dimensions is most mismatched." The answer is almost always Workload or Control. The most expensive solutions ignore both.
Section 04What the Industry Sells, and What the Evidence Actually Shows About It
The workplace mental-health industry is built around four flagship product categories. The evidence base on each is much weaker than the marketing implies.
Meditation and mindfulness apps
The most thoroughly evaluated of the four. The largest randomized controlled trial of workplace meditation apps to date — Bartlett et al., 2024, n=4,237 employees across 11 U.K. organizations — found no statistically significant improvement on burnout, productivity, or job satisfaction at 12 months compared to waiting-list controls [4]. The British Medical Journal's 2025 systematic review of workplace mindfulness interventions concluded that "evidence of benefit at the workforce level is weak, and existing positive findings are concentrated in small studies with significant risk of bias" [5].
This is not an argument that meditation does not help individuals who choose to practice it. It is an argument that meditation deployed as a workforce-level intervention to address structural workplace conditions does not produce the structural-workplace-level effect employers buying it expect.
Employee Assistance Programs (EAPs)
EAPs — the third-party hotline-and-short-term-counseling benefit standard across U.S. employers since the 1980s — have an industry-claimed utilization rate of 6.5%. Independent audits find true utilization closer to 2–4% [6]. The workers who most need EAPs use them least; the workers who use them most often have alternative care available regardless. The 96–98% of workforces that do not engage the EAP receive no benefit from its existence other than the optics of its provision.
Resilience training
Workshops or courses that teach workers cognitive techniques for managing workplace stress. The U.S. Army's $100M+ Comprehensive Soldier Fitness program is the most thoroughly evaluated case — the program's effect on the relevant outcomes was small and contested even in the studies the program itself commissioned [7]. In the civilian sector, resilience-training meta-analyses show small short-term effects that decay within months.
The frame this category gets right is that workers benefit from skills to navigate stress. The frame it gets wrong is that giving a worker stress-management skills is a substitute for fixing the conditions producing the stress.
Wellness platforms and step-counting programs
The University of Illinois's 2018 RCT of a comprehensive corporate wellness program — n=4,834 employees, two years — found no significant differences in clinical outcomes, productivity, or health spending between treatment and control groups [8]. The 2020s have produced several attempted re-litigations of this finding. The underlying pattern has held.
Section 05What the Evidence Says Does Work
The interventions with the strongest evidence base for actually reducing workplace burnout share two features: they are structural rather than individual, and they are operational rather than therapeutic. Five categories, in order of evidence strength.
1. Workload-cap policies that managers cannot override
The clearest evidence comes from the medical profession, where caps on resident work hours (the 2003 ACGME 80-hour-week rule and its subsequent revisions) produced measurable reductions in burnout indicators among trainees [9]. The lesson generalizes: a workload limit that the worker has the authority to invoke and the manager cannot waive produces benefits that no amount of meditation does. Where this is being adapted in the modern knowledge-economy: "fair-week" scheduling at large retailers, the EU Working Time Directive's after-hours email norms, the on-call rotation overhauls inside SRE-heavy tech firms.
2. Manager training, specifically on the six conditions above
Workplace burnout co-varies with manager quality more strongly than with almost any other variable. Targeted manager training — not generic "people leadership" training, but specific training on the six Maslach-Leiter conditions and the manager's authority to address them — produces team-level burnout reductions of 0.3–0.5 SD in the available studies [10]. The cost is roughly $1,000–$3,000 per manager. It is the highest-ROI intervention in this category.
3. Predictable scheduling
For shift-based and frontline workforces, schedule unpredictability is one of the largest documented contributors to burnout (and to physical health outcomes). The Seattle and San Francisco "fair workweek" laws, the chain-by-chain retail commitments that followed, and the multi-year RCT at Gap Inc. on this exact question all produced consistent results: predictable scheduling reduces burnout, reduces turnover, and — surprising to many of the firms — has no negative effect on store productivity [11].
4. Genuine autonomy expansions
The "Control" dimension of the six conditions is the dimension most cheaply addressed. The interventions are surprisingly low-cost: allowing workers to choose how to sequence their tasks, to determine their own schedule within wide bounds, to invoke "no-meeting" blocks. The evidence is consistent across knowledge work, clinical care, and skilled trades.
5. A clinical-grade EAP that is actually used
EAPs are not without value — but only when (a) the access is friction-free (single login, mobile-first, no insurance navigation), (b) the clinical quality is independently audited, and (c) usage data is treated as a signal of conditions to be fixed, not as a stigma. The 2-4% utilization rate doubles when the friction comes down, and the 4-8% utilization rate doubles again when leaders themselves publicly use the resource.
Section 06An Honest Burnout Budget
A reallocation of a typical $50–per-employee-per-month workplace mental-health spend toward the interventions the evidence supports:
Where the money should actually go
- 40%: Manager training on the six conditions, refreshed annually
- 25%: Structural workload protections — staffing, recovery time, hours caps where applicable
- 15%: A clinical-grade, friction-free EAP from a clinically audited provider
- 10%: Predictable scheduling infrastructure (HRIS investment for shift workforces)
- 10%: Optional individual resources (apps, programs) that workers can choose to use
The ratio that most organizations are running today is closer to the inverse of this — 60% on individual resources (apps, EAPs, wellness platforms), 30% on resilience and training, and 10% on the structural and operational moves that produce the largest effects.
Section 07If You Are Burned Out and Reading This
Three things the literature supports for the individual worker, even in a workplace that has not yet addressed the structural conditions.
- Diagnose, do not blame yourself. Run the six-conditions diagnostic on your own situation: which of Workload, Control, Reward, Community, Fairness, Values is most mismatched in your specific role? Burnout is rarely a sign of moral failure or weakness. It is the predictable result of certain configurations of work.
- Name the cost. The conversation with your manager or HR will go better if the cost is concrete — "I am on track to attrit within six months at the current workload" reads as data; "I am stressed" reads as venting. The first opens a budget conversation. The second does not.
- Recovery is a workplace condition, not a vacation. Vacations partially restore an individual; they do not change the conditions that produced the depletion. If you take a two-week break and return to identical workload and identical control conditions, you will be burned out again within 8–12 weeks. Vacation is not the intervention. Changing the conditions is.
If you are a leader in a position to do something about the structural conditions, the six-conditions diagnostic is the place to start. The interventions that actually work are within most organizations' authority and budget. The harder part is the cultural shift from "the worker's resilience" to "the system's design." Most of the cost of the burnout decade has been in delaying that shift.
If your mental-health spending has gone up, your burnout numbers should have gone down. They didn't. Either the wrong things were bought, or the right things weren't. The evidence is unambiguous about which.
