Workplace safety is one of the under-sung triumphs of the last century, and it has run out of the thing it was built to fight. The hard hat, the harness, the guard on the machine and the sign about the wet floor were designed for the acute injury, the falling object and the broken limb, and against that enemy they have largely won; Australia's serious-claim rate has fallen by roughly a quarter over the past decade. The trouble is that the acute injury is no longer what shortens most working lives. The slow harm is, and the workplace, which sees a working adult more hours of the week than any clinic or any family does, does little about it.
By slow harm I mean the exposome of ordinary work: the diesel particulates in the loading dock, the carbon dioxide that climbs through a sealed afternoon in an open-plan office, the heat that is turning Australian summers into an occupational hazard in their own right, and the slow creep of blood pressure, blood sugar and the odd shadow on a scan that few think to check for. The World Health Organization puts the toll of ambient and household air pollution at around seven million deaths a year, a number closer to the scale of tobacco than of industrial accidents. We regulate the dust and the noise, because a previous generation of reformers fought to, and we leave the air and the body to chance. It is a strange place to have drawn the line.
The reframe I would offer is that this is not a wellness problem to be outsourced to an app, but a safety problem one category wider than the one we already accept. Michael Marmot's Whitehall studies, which followed eighteen thousand British civil servants over decades, established that the workplace is not a neutral backdrop to health but one of its strongest determinants, the gradient of status and condition written directly onto the body. A duty of care that stops a pallet crushing a foot can, with the same logic and the same insurer, stop the heart attack that the foot's owner is unknowingly marching towards. Two moves get you most of the way, and neither needs new legislation to begin.
- Measure and clean the air, as we already do the dust. Treat a workplace's air and environment as a health exposure with limits and a sensor, extending the monitoring we already mandate for hazardous substances to fine particulates, carbon dioxide and heat. The Safer Air Project and the broader clean-indoor-air movement that grew out of the pandemic have shown this is now cheap, continuous and unobtrusive; the science is closer to settled than the practice.
- Check the worker, not just the site. A systematic annual health check for every employee, with the results owned by the worker and routed to their own GP, catches the hypertension and the pre-diabetes at the point where a change of habit still works rather than a stent. Japan has required employer-provided checks under its Industrial Safety and Health Act since 1972, the annual ippan kenshin and the more thorough ningen dock, and the evidence on early detection is one reason Japanese working-age mortality sits where it does.
Japan is the through-line worth dwelling on, because its preventive-health culture is not an accident of temperament but of design, built into the obligations of the firm rather than left to the conscience of the individual. I have spent a good part of my working life learning from Japanese institutions, and the lesson that keeps recurring is that they make prevention a shared structural duty rather than a personal virtue, which is exactly the move an exhausted Australian primary-care system cannot make on its own. Mariana Mazzucato writes about the state shaping markets towards a mission; here the mission is already legislated, and the firm is simply asked to carry a sliver of it on behalf of the public good it benefits from.
The annual check is the leg I would build most carefully, because the evidence does not reward simply scanning everyone and hoping. The Cochrane reviews of general health checks found that inviting a whole population in for routine screening did little for overall mortality while carrying a cost of its own: the false alarm, the incidental finding, the biopsy and the months of worry that follow a shadow that would never have harmed anyone. A universal full-body sweep earns that objection. The version worth funding is narrower, the handful of conditions where catching them early demonstrably changes the course, blood pressure, blood sugar, the exposure-linked harms the job itself creates, routed to the worker's own GP as Japan does rather than turned into a hunt for whatever a scan happens to show. Of the ideas in this set this is the one with the thinnest margin between help and harm, and it should be made to earn its place rather than asserted.
Which leaves the question every one of these ideas has to answer, the one Rozetta rightly insists on: who pays, and why would they. The answer is that the employer already does, for the narrower version, and the arithmetic of the wider one is the same arithmetic.
- The employer, on the safety logic it already accepts. Fewer sick days, higher productivity, lower turnover and a smaller tail of occupational-disease liability all land on the same ledger that already justifies the harness and the training day.
- The self-insurer, most sharply of all. A licensed workers' compensation self-insurer carries its own chronic-disease claims directly, and chronic occupational disease is the fastest-growing and most expensive line it faces; a dollar of prevention shows up on its own book, not on some distant public account it will never see.
- The health and life insurers, as co-funders. Where an early-caught condition is a claim that never matures, the insurer whose claim it would have been has an obvious reason to help pay for the catching.
The other honest objection, and the one that can sink the whole idea, is not cost but trust. A health check your boss pays for is a short step from a health check your boss reads, and from there to the discreet non-renewal of the worker whose scan looked wrong; the history of workplace medicine includes enough of that to earn the suspicion. So the design has to hold on exactly this point. The worker owns the data and the clinician holds the duty of care; nothing flows to hiring or firing, by law if it comes to that; and the whole thing has to be built to avoid the twin failure modes of surveillance on one side and wellness theatre, the fruit bowl and the mindfulness webinar, on the other. Get that wrong and the idea deserves to fail. Shoshana Zuboff's warnings about the instinct to turn every human signal into a managed asset are the right thing to keep pinned above the desk while building it.
None of this is speculative technology. The sensors exist, the Japanese template is fifty years old, and Australia's work health and safety framework already mandates exposure monitoring for the hazards a previous generation named; the step is only from the dust and the noise to the air and the whole body.