Burnout statistics for 2026 — the seven numbers that actually matter
The numbers in your wellness vendor's pitch deck are usually outdated, often misattributed, and frequently irrelevant to your sector. These seven aren't.
By Chris Davis, M.S., Co-Founder, Pivot Training & Development
Every burnout vendor's homepage cites the same five-or-so statistics, and most of them are out of date or distorted in transit. This is a curated list of seven numbers that, as of mid-2026, hold up — and an honest description of what each one is and isn't telling you.
1. 76% of US workers experience burnout at least sometimes (Mercer Health on Demand, 2024)
What it means: roughly three-quarters of the US workforce reports burnout symptoms with some frequency. Mercer's survey methodology is reputable; the 'at least sometimes' threshold is broad, which inflates the number relative to clinically significant burnout.
What it doesn't mean: that 76% of your team needs a clinical intervention. The clinically significant burnout rate is closer to 25–35% in most office settings and 50–60% in clinical/helping professions. Use the 76% as a 'this is a workplace issue, not an individual issue' framing — not as your treatment population estimate.
2. $322B — global cost of burnout-related lost productivity and turnover (Gallup, 2024)
What it means: Gallup's modeling combines lost productive output (presenteeism) with replacement costs (turnover) to get a global annual figure. The number is large enough to communicate scale but not granular enough to inform a specific business case.
What it doesn't mean: that your organization is losing $322B / 8 billion-employees-globally per employee. For a specific business case, use your own organization's voluntary turnover rate × replacement cost per role. That's the number a CFO will actually respond to.
3. 49% of physicians and 56% of nurses report burnout symptoms (AMA + AHA, 2024–2025 data)
What it means: about half of physicians and well over half of nurses are at clinically significant burnout levels, post-pandemic recovery notwithstanding. Both numbers improved 3–5 points from the pandemic peak but remain ~10 points above the pre-pandemic baseline.
What it doesn't mean: that the situation is improving. Recovery from pandemic peak doesn't mean the pre-pandemic baseline was sustainable. The pre-pandemic numbers reflected an already-unsustainable system; the current numbers are 'unsustainable plus pandemic residual.'
4. 1 in 5 employees has been diagnosed with anxiety or depression (CDC + private surveys, 2024)
What it means: clinically diagnosed (not just self-reported) mental health conditions are at historic levels. Many of these are comorbid with burnout, but not all — depression and anxiety can exist independently of work conditions. About half of cases involve some workplace contribution per longitudinal studies.
What it doesn't mean: that your wellness program will fix this. Clinical conditions need clinical care. The workplace contribution to your team's mental health load is the part you can affect; the underlying conditions need EAP referrals, benefits coverage, and time off.
5. Managers report 23% higher burnout rates than ICs (multiple studies, 2024–2026)
What it means: middle management is the most-burned-out layer in most organizations. The consistent finding across 5+ studies in the last 24 months: managers carry the dual load of doing their own work plus absorbing their reports' emotional load, and most organizations don't account for the second half in capacity planning.
What it doesn't mean: that managers will tell you. Managers are the least likely cohort to surface their own burnout because their professional identity depends on appearing capable. See: 'Manager burnout — how to spot it before retention numbers go.'
6. Replacement cost is roughly 50–250% of annual salary, depending on role complexity (SHRM, 2024)
What it means: for entry-level roles, replacement cost is roughly half of annual salary. For specialized clinical, technical, or senior roles, it's 2–2.5x. The variance comes from recruiting costs, ramp time, productivity loss during the gap, and the institutional-knowledge cost.
What it doesn't mean: that you should use the average. Use the role-specific multiple. A 1% improvement in retention of a 50-person specialty team is worth 5–10x what the same percentage means at a generalist call center, because the replacement cost difference is that large.
7. Workplace driver scores predict burnout 6–9 months ahead of symptom scores (Maslach & Leiter longitudinal data)
What it means: the Areas of Worklife dimensions — workload, control, reward, community, fairness, values — show degradation before the symptom dimensions (exhaustion, detachment, reduced efficacy) start moving. If you're only measuring symptoms, you're seeing the problem too late to prevent it.
What it doesn't mean: that drivers alone are enough. You need both — the leading indicators (drivers) for prevention, the lagging indicators (symptoms) for intervention. This is why BurnoutIQ measures 9 dimensions instead of 3.
How to use these in a leadership briefing
Pick the three most relevant to your situation. Don't recite all seven — that's the wellness-vendor playbook and it doesn't land. The CHRO who cares about retention wants stat 6 + your org's data. The CMO who cares about patient outcomes in a hospital system wants stat 3 + HCAHPS correlations. The board member who wants the case for sustained investment wants stat 2 + your own P&L math.
And if you want the org-specific version of all this, BurnoutIQ ships sector-benchmarked readings — your numbers vs your sector's published norms — so you can have the same conversation with data that's actually yours.