840,000 Deaths a Year. And We're Still Talking About Wellness Apps.
The ILO just released its most comprehensive global report on workplace psychosocial risk. The headline figure is stark: more than 840,000 workers die every year from cardiovascular disease and mental disorders directly linked to how their work is designed, organised, and managed.
That is not a mental health statistic. That is a work design statistic.
The distinction matters enormously, because most organisations have responded to the wrong problem.
The Risks Are Structural. The Responses Are Not.
The ILO identifies five primary drivers behind those deaths: job strain (high demands, low control), effort-reward imbalance, job insecurity, long working hours, and workplace bullying and harassment.
None of these originate in the individual. All of them originate in decisions made by organisations about how roles are structured, how workloads are set, how managers are trained, and what behaviours are tolerated.
35% of workers globally exceed 48 hours per week. 23% have experienced violence or harassment in their working life.
These are not edge cases. They are operating norms in organisations that simultaneously publish wellbeing strategies and run EAP programs. The problem is not that organisations are doing nothing. The problem is that what they are doing addresses the person, not the system that is harming them.
The report is explicit: priority must go to "organisational and collective measures that address root causes, including workload management, role clarity, communication, participation and leadership practices." Individual interventions, such as counselling, resilience training, and mindfulness, are described as supplementary. They should "complement, not replace, action on organisational conditions."
Most organisations have that the wrong way around.
The New Risks
The report flags what is emerging alongside the familiar patterns.
Digitalisation and AI are transforming how tasks are coordinated, monitored, and assessed. New employment arrangements, including platform work, varied contracts, and remote and hybrid models, are redefining supervision, expectations, and working time. The report notes these developments may "exacerbate psychosocial risks" if not proactively managed.
Digital monitoring sits squarely in the frame. Surveillance of workers through algorithmic management tools is now explicitly recognised as a psychosocial hazard, something that directly undermines autonomy, control, and dignity at work. Organisations deploying these tools while publishing values statements about psychological safety are not just inconsistent. They are creating measurable, documented risk.
The Uncomfortable Calculation
The economic cost sits at 1.37% of global GDP annually. Nearly 45 million disability-adjusted life years are lost every year. The ILO's own analysis shows that the cost of prevention is dwarfed by the cost of inaction.
And yet national legislation still frames employer duties primarily around protecting "mental health," language the report notes "may encourage individualised responses" rather than structural ones. Only 18% of cross-border corporate agreements between 2000 and 2025 explicitly address mental health or psychosocial factors at all.
The gap between what the evidence demands and what organisations actually do is not a knowledge problem. The evidence has been available for years.
It is a will problem.
Organisations know that workloads are unsustainable. They know that management quality varies wildly. They know that certain roles carry chronic stress and that their reporting systems undercount harm. They know, and they continue to book the yoga instructor.
What the Report Actually Recommends
The ILO structures its recommendations across three areas. They are worth reading carefully, because each one implicitly indicts the way most organisations currently operate.
On research and data: The report calls for psychosocial indicators to be embedded in national OSH monitoring systems rather than tracked separately, if at all. Right now, most organisations cannot tell you their actual exposure levels. They do not measure workload systematically. They do not track the relationship between management quality and worker health outcomes. You cannot manage what you do not measure, and most organisations have made a quiet decision not to measure this.
On policy and regulation: The report calls for OSH laws to explicitly address psychosocial risks, and flags a specific trap: legislation that focuses on "mental health at work" as an outcome rather than on work design as a cause tends to produce individualised responses rather than structural ones. The framing matters. Calling something a mental health problem sends it to HR and the EAP. Calling it a job design problem sends it to operations, leadership, and governance. The report is clear that the latter is the correct address.
On workplace action: This is where the report is most direct. It calls for workload review, role clarity, revised supervision practices, and genuine worker participation in how risks are identified and managed. It names leadership responsibility specifically: prevention must be "embedded in everyday management decisions," not delegated to a wellbeing team. It also draws a line that most organisations will find uncomfortable. Individual measures like counselling and resilience training should supplement organisational action, not substitute for it.
That last point is the one most organisations will quietly ignore. Individual interventions are cheaper, faster, and do not require anyone to examine whether the way work is structured is itself the problem. Organisational change requires leaders to look at their own decisions as a source of harm. That is a harder conversation to start, and an even harder one to finish.
The report does not pretend these changes are simple. But it is unambiguous about the direction of travel.
Fix the work. Not just the worker.
