The state of
outdoor work
in Australia

Why outdoor work needs its own conversation#

Australia does a reasonable job of counting workplace harm. Safe Work Australia publishes fatality figures every year and workers compensation data every year. What the country does not do is cut any of it by whether the work happens under a roof.

That gap matters, because the hazards that define outdoor work do not respect industry boundaries. A linesman, a fruit picker, a concreter, a courier and a park ranger sit in five different divisions of the Australian and New Zealand Standard Industrial Classification. They share almost nothing on an organisational chart. They share a great deal on a summer afternoon: the same ultraviolet radiation, the same heat load, the same moving plant, the same distance from a hospital, the same weather that decides whether today is a paid day.

Three forces are converging#

One. The physical risk has not fallen the way the headline suggests

Australia's work fatality rate has dropped 24 per cent since 2014, from 1.7 to 1.3 deaths per 100,000 workers. That is a real achievement. But the decline has flattened, and it is unevenly distributed. Agriculture, forestry and fishing still runs at 13.7 deaths per 100,000 workers, more than ten times the national rate. Serious claims have gone the other way: up 34.5 per cent in a decade.

Two. The environment itself is becoming a hazard

The Australian Government's first National Climate Risk Assessment, released 15 September 2025, named mining, construction, agriculture and power generation and transmission as the sectors most exposed to heat-driven productivity loss. It projected between 700,000 and 2.7 million additional working days lost to heatwaves every year by 2061, and a hit to national economic output of $135 billion to $423 billion by 2063. Heat has moved from a comfort issue to a balance sheet issue.

Three. The legal floor has risen sharply

Between 2022 and 2026, industrial manslaughter became an offence in every Australian jurisdiction, engineered stone was banned outright, psychosocial hazards became an express duty, silica rules extended to every industry, and in New South Wales codes of practice became enforceable in their own right. A safety failure in 2026 carries consequences that did not exist in 2020.

Sizing the outdoor workforce#

There is no official count of Australian outdoor workers. The Australian Bureau of Statistics classifies employment by industry and by occupation, and neither classification asks whether the work happens outside.

The workable approximation is to take the industry divisions in which the majority of frontline work is performed outdoors. On the ABS Labour Account for the March quarter of 2026, the four largest such divisions employ just over 2.7 million people.

Source: ABS Labour Account Australia, March quarter 2026, employed persons. Shares are calculated against Labour Account total employed persons of 15,109,700, not against the Labour Force Survey total, because mixing the two collections produces a misleading ratio.

Where this estimate is wrong, and in which direction#

Honesty about the boundary matters more than a precise-looking number, so here is what the figure above gets wrong in both directions.

• It over-counts. Transport, postal and warehousing includes distribution centre staff and freight schedulers who never leave a building. Mining includes control room operators and corporate functions. Construction includes estimators, contract administrators and site office staff.
• It under-counts. It excludes emergency services personnel sitting inside public administration and safety, groundskeepers inside education, couriers and delivery riders inside retail and accommodation, and the large body of outdoor labour hired through labour hire firms.
• It misses insecure work entirely. Gig-model delivery riders, seasonal harvest labour and short-term contractors are systematically hard to capture in either employment or compensation data.

The two distortions run in opposite directions and roughly offset. Treat 2.7 to 2.9 million as the defensible range for the Australian outdoor workforce, or somewhere between 18 and 19 per cent of everyone in a job.

The demographic profile#

Outdoor work in Australia skews heavily male, skews older at the top end, and concentrates in regional and lower socioeconomic areas. Those characteristics run straight through the harm data. Men account for 96 per cent of work-related deaths and carry a fatality rate of 2.4 per 100,000 workers against 0.1 for women. Workers aged 65 and over have a fatality rate of 4.3 per 100,000, more than three times the national figure. The Australian Work Exposures Study found that workers exposed to solar ultraviolet radiation were more likely to be male and to live in regional and lower socioeconomic areas, with farming, painting and plumbing the most exposed occupations of all.

The fatality ledger#

In 2024, 188 Australian workers were killed by traumatic injury at work. Both that count and the five-year average it is measured against cover all Australian workers, indoors and outdoors together, because Safe Work Australia does not publish a separate series for outdoor work. On that all-industry basis, 188 deaths sits slightly below the five-year average of 191 deaths a year, and the fatality rate of 1.3 per 100,000 workers sits slightly below the five-year average of 1.4. The concentration in outdoor industries is set out below.

Those national figures are the good news, and they are genuinely good: the rate has fallen 24 per cent since 2014. The bad news is where the remaining deaths are concentrated.

Deaths by industry#

Five industry divisions where the work is predominantly performed outdoors, namely transport, agriculture, construction, mining and utilities, accounted for 150 of the 188 deaths recorded in 2024. That is 79.8 per cent of every Australian killed at work last year.

Deaths by rate, which is the number that should worry you#

Counts reward large industries. Rates reveal danger. On a per-worker basis, agriculture, forestry and fishing is in a category of its own: 13.7 deaths per 100,000 workers, more than ten times the all-industry rate of 1.3.

What actually kills outdoor workers#

The mechanisms are not exotic. Vehicles dominate, and they dominate overwhelmingly.

• Vehicle incidents caused 79 deaths, 42 per cent. A vehicle was directly involved in 66 per cent of all worker fatalities, or 124 deaths. Half of those, 62, involved at least one truck. Critically, 72 per cent of vehicle-involved deaths were single-vehicle incidents, meaning fatigue, terrain, rollover and speed rather than collision with another road user.
• Falls from a height caused 24 deaths, 13 per cent, the second largest mechanism and a persistent construction and maintenance problem.
• Being hit by moving objects caused 17 deaths, 9 per cent, a category that includes workers on foot struck by plant and vehicles, which is the defining hazard of any active worksite that mixes people and machines.

Who dies#

Two findings stand out. Machinery operators and drivers die at 6.7 per 100,000, over five times the national rate, and account for almost a third of all deaths. And the fatality rate rises steeply with age: workers aged 65 and over die at more than three times the national figure. As the outdoor workforce ages and older workers stay on in agriculture and transport, this is a curve moving in the wrong direction.

The farm data tells a harsher story than the national data#

Safe Work Australia counts workers. AgHealth Australia at the University of Sydney counts everyone who dies on a farm, including children, family members and bystanders, drawing on media reports. The two datasets are not equivalent, and the farm figures are the more alarming of the pair.

Across 2024, side-by-side vehicles (14), quad bikes (10) and tractors (8) were the most frequently reported causes of on-farm death, and 87.5 per cent of those killed were male. Over the longer window from 2013 to 2023, tractors (118) and quad bikes (117) were the two leading causes of work-related farm fatality, followed by farm utilities (52). Ninety-four per cent of victims were men and almost half were aged 60 or over. More than 75 per cent of the reported 2024 injuries occurred in Queensland and New South Wales.

Research published in 2025 put a number on the cost for the first time: about $355 million a year in combined death and injury compensation across Australian agriculture, of which roughly $300 million is work-related, and around $1.8 billion in total for all farm fatalities across 2013 to 2023.

The injury ledger#

Deaths are the smallest part of the harm. In 2023-24 there were 146,700 serious workers compensation claims in Australia, defined as claims involving at least one working week of time lost. That is more than 400 serious claims every day.

And unlike the fatality rate, this number is going the wrong way. Serious claims have risen by 37,600, or 34.5 per cent, over the past decade. Adjusted for hours worked, the frequency rate still rose 12.6 per cent.

Outdoor industries carry a quarter of the national claim burden#

The five predominantly outdoor divisions generated 36,600 serious claims in 2023-24, close to a quarter of the national total from around a fifth of the workforce.

Frequency rate is serious claims per million hours worked. Claim counts and frequency rates are 2023-24 preliminary. Median weeks lost and median compensation are 2022-23, the most recent non-preliminary period, because some 2023-24 claims remain open. Source: Safe Work Australia.

Agriculture, forestry and fishing records the highest serious claims frequency rate of any industry in Australia at 10.0 claims per million hours worked, 46.9 per cent above the national average. Construction and transport sit close behind. Mining is the outlier in the group: its frequency rate is below the national average, but when a mining injury does occur it costs more than double the national median at $36,400 and takes 11.1 weeks off work.

Occupation is a sharper predictor than industry#

If you want to know who gets hurt, ask what someone does rather than who they work for. Labourers record a serious claims frequency rate of 23.1 per million hours worked, nearly three and a half times the all-occupation average of 6.8. Machinery operators and drivers sit at 12.1, technicians and trades workers at 8.1. Together those three manual and trade groups hold 53.5 per cent of all serious claims.

How outdoor workers get hurt#

Four mechanisms account for 83.8 per cent of all serious claims nationally, and all four are everyday features of outdoor work.

• Body stressing: 50,600 claims, 34.5 per cent. Lifting, carrying, repetitive strain and awkward posture. The single most common mechanism in every major occupation group. Median 9.2 weeks lost.
• Falls, trips and slips: 32,000 claims, 21.8 per cent. Of these, 68.3 per cent were falls on the same level, not from height. Uneven ground, wet surfaces and site clutter, not scaffolding.
• Being hit by moving objects: 23,400 claims, 16.0 per cent. Plant, tools, materials and vehicles in motion around workers on foot.
• Mental stress: 16,800 claims, 11.5 per cent. Covered in section 7. The most common triggers were harassment and workplace bullying at 33.2 per cent, work pressure at 24.2 per cent, and exposure to violence and harassment at 15.7 per cent.

The tail is where the money goes#

Most claims are short. The cost is not. In 2022-23, claims involving more than thirteen weeks off work made up 21.9 per cent of total accepted claims but consumed 74.8 per cent of all compensation paid, some $5.4 billion. The proportion of claims involving no time lost has fallen steadily over the past decade while the long-tail share has grown, and median time lost has risen 35.1 per cent in ten years.

The hazard nobody measures#

Read sections 3 and 4 together and one pattern runs through all three of the largest harm categories in Australian outdoor work.

Australian workplaces have spent thirty years engineering out the hazard: rollover protection, exclusion zones, high visibility clothing, ground condition assessment. Those controls work, and section 9 argues for more of them. But every one of them assumes a worker who is looking. None of them measures whether the worker is.

What the research actually shows about phones#

There is no Australian study of phone use in outdoor work. There is a substantial body of work on each of the three mechanisms above, and it points in one direction.

Driving, which accounts for 42 per cent of deaths

The systematic review evidence links problematic phone habits directly to distraction behind the wheel. Reviewing the international literature, Oviedo-Trespalacios and colleagues found the association between maladaptive phone use and distracted driving to be consistent across studies, and concluded that interventions should target the habit rather than the moment.

Australian work on why drivers do it anyway is unusually good. Studying young Australian drivers who use phones illegally, Koppel and colleagues documented how riders normalise the behaviour through what they call risk-compensatory strategies, holding the phone low, using it only at lights, telling themselves it is "wrong, but not that wrong". Kaviani and colleagues at Monash went further and showed that attachment to the phone, measured as nomophobia severity, predicts illegal use better than the perceived risk of being caught does.

Walking and slipping, which accounts for 68.3 per cent of the 32,000 falls claims

This is where the evidence gets specific, and Australian. Pelicioni and colleagues at the University of New South Wales and Neuroscience Research Australia put 50 young adults on a 10 metre walkway with controlled slip hazards and measured what texting did to their ability to recover.

Their conclusion, in their words: texting while walking "significantly impaired postural balance recovery when slipping". Note the structure of the result. Texting had no significant effect on trunk angle during normal walking or when participants merely anticipated a hazard. It mattered at the moment the slip actually happened, which is precisely the moment that turns a stumble into one of those 32,000 claims.

Independent biomechanical work published in 2025 found the same thing for obstacles rather than slips: phone tasks impaired frontal-plane dynamic stability and increased fall risk during obstacle crossing. A paddock, a trench edge and a half-built slab are obstacle-crossing environments.

Being struck, which killed 17 workers and generated 23,400 claims

The closest Australian evidence comes from the road rather than the worksite. Osborne, Horberry and Young at the Monash University Accident Research Centre combined a literature review, an observational study in Melbourne, pedestrian interviews and an expert workshop. They observed 20 per cent of pedestrians using smartphones while crossing roads at city locations, and recorded significantly higher rates of critical events, near misses and collisions, among those users.

Their framing is worth borrowing. They argue countermeasures belong inside a Safe System approach, one that "recognised that road users such as pedestrians will make mistakes", and therefore builds an error-tolerant environment rather than relying on individual vigilance. Swap "pedestrian" for "worker on foot" and "road" for "active worksite" and the argument transfers without modification.

Industry already acts as though the link is real#

Walbridge, one of the larger construction firms in the United States, prohibited talking, texting and device use while walking on its sites in 2010, restricting them to trailers and break rooms. Its vice president of safety put the reasoning plainly: "It's a bad idea to talk on the phone or text while you're walking around at a construction site. You not only put yourself in danger, but you can endanger others."

That policy is sixteen years old. It was made, like almost every phone rule on almost every Australian site, without a single measurement of how much phone use was actually occurring, or whether the rule changed it.

The part that is now measurable#

For most of the period covered by this report, measuring phone use on a worksite would have meant standing somebody at the gate with a clipboard, which is why nobody has ever done it properly. That is no longer the constraint. Every modern smartphone carries an accelerometer and a gyroscope that between them report how the device is moving through three axes, several times a second, and they do it whether an app asks or not.

The useful part is that a phone riding in a trouser pocket and a phone held up in front of a face are doing completely different things in that coordinate system. One is being swung through an arc by a leg at roughly two steps a second. The other is being held almost still, with small corrections as a thumb moves. You do not need to know anything about what is on the screen to tell those two states apart, and PocketMode does not try to.

The second half of the method matters more than the first, because it is what makes the measurement acceptable to the person being measured. Monitoring that runs all day is a different proposition to monitoring bounded to the shift.

What we can and cannot say#

That gap is not unusual in Australian rural safety research, and it has a close precedent. Reviewing the fatigue literature, Summers, Peachey and Lower found that the studies they could locate "unanimously agreed that fatigue contributes to occupational injury in agriculture", but that "a scarcity of specific literature to Australia or agriculture" prevented any firm conclusion. Fatigue is a hazard everyone agrees is important and nobody has quantified on an Australian farm. Distraction is in the same position, one step further back.

Why closing the gap is now worth doing#

Two things changed recently that make this measurable in a way it was not ten years ago.

• The measurement is solved. Detecting whether a phone is being handled, from the phone's own accelerometer and gyroscope, is now established method. Published work in 2025 demonstrated real-time detection of distracted walking from smartphone inertial sensors alone, with no camera, no microphone and no content capture. The same sensing has been running at scale in insurance telematics for years.
• The safety system is asking for exactly this kind of number. Harrison and colleagues, including researchers at AgHealth Australia and James Cook University, are running a modified Delphi study to develop lead indicators for Australian agriculture, measurable things that predict injury before it happens rather than counting it afterwards. Minutes of handheld phone use per tracked shift-hour is a candidate lead indicator of exactly that shape.

We think it should be a university, not a vendor. The technical edition sets out a proposed research agenda, including a study design that could be run as an industry-linked PhD, and PocketMode has offered its sensing platform and its field access to whichever group wants to run it.

The slow hazards: sun, heat and climate#

A fall from a roof produces an incident report. Twenty-five years of ultraviolet exposure produces a diagnosis, usually long after the worker has left the employer who caused it. Both are workplace harms. Only one of them shows up in the safety statistics on time.

Ultraviolet radiation: the hazard Australia is worst at#

Australia experiences among the highest ambient ultraviolet radiation levels in the world and, as at 2020, had the highest age-standardised melanoma incidence rate of any country. Outdoor workers take in roughly 10 per cent of available ambient ultraviolet radiation, and receive between five and ten times more exposure per year than indoor workers.

Cancer Council Australia estimates that around 200 melanomas and 34,000 non-melanoma skin cancers each year in Australia are caused by occupational exposure. These are modelled attributable-fraction estimates rather than registry counts, because non-melanoma skin cancer is not notifiable in Australia. Meta-analyses put the squamous cell carcinoma risk for outdoor workers at an odds ratio of 1.77 against indoor workers, and basal cell carcinoma at 1.43.

The protection gap is the part employers can act on immediately, and it is enormous. The Australian Work Exposures Study found that about 95 per cent of exposed outdoor workers used at least one form of sun protection. Only 8.7 per cent were classified as fully protected, meaning they used a hat, sunscreen, clothing and shade for more than half of their outdoor working time.

Put differently: almost everyone does something, and almost nobody does enough. An earlier national survey, the National Hazard Exposure Worker Surveillance survey of 2008, found only 21 per cent of workers in direct sunlight scheduled work outside peak ultraviolet hours, and 17 per cent reported that neither they nor their employer did anything at all to prevent sun-related harm. Those two figures are eighteen years old and come from a different instrument to the AWES numbers above; no more recent national measurement exists.

Heat: the hazard that is growing fastest#

Heat harm is under-recorded for a different reason. Heat, electricity and other environmental factors accounted for only 1,900 serious claims in 2023-24, just 1.3 per cent of the total, with a median of 2.4 weeks lost and $3,800 paid. Nationally, 1,774 heat-related claims were accepted across the entire decade from 2009-10 to 2018-19, and of those, 441 were for heat stroke or heat stress specifically.

What the climate assessment actually projects#

In September 2025 the Australian Government released the first National Climate Risk Assessment. It named mining, construction, agriculture, power generation and transmission, and some service occupations as particularly at risk from labour productivity reductions due to higher temperatures.

The range is wide because it depends on emissions pathway and adaptation. The direction does not. Every projection has the number of workable outdoor hours in an Australian summer falling, in every state, for the rest of the century. Note that the days-lost figures are additional days above baseline, not totals.

The mental load nobody rostered for#

The fastest growing category of workplace harm in Australia is not a physical injury. Serious claims for mental health conditions have risen 161.1 per cent in ten years, the largest increase of any nature of injury group by a wide margin.

Mental health conditions now make up 12.0 per cent of all serious claims, up 14.7 per cent in a single year. They are also, by a distance, the most expensive category of workplace harm in the country.

Why this lands hard on outdoor sectors#

The mental health claim data is not published split by whether work is indoors or outdoors, so what follows is inference from the surrounding evidence rather than a measured finding. But the drivers documented in the claims data map closely onto the conditions of outdoor work.

• Harassment and workplace bullying drives 33.2 per cent of mental stress claims. Male-dominated, hierarchical crews with strong informal cultures are a recognised setting for exactly this.
• Work pressure drives 24.2 per cent. Weather-compressed schedules, liquidated damages on construction programmes and seasonal harvest windows concentrate pressure into short, unavoidable bursts.
• Income volatility is structural. When rain stops work, income stops. Shift work research consistently finds that difficulty managing fluctuating income and unpredictable scheduling are the leading complaints of younger shift workers.
• Isolation and remoteness compound everything. Lone work, fly-in-fly-out rosters and long-haul driving remove the informal support that indoor workplaces provide by default.
• Exposure to trauma is routine. Emergency services, road transport and agriculture all involve witnessing serious incidents as a foreseeable part of the job.

The compliance floor has moved#

If your safety system was designed before 2022, it was designed for a different legal environment. Five changes in quick succession have raised the floor for every Australian employer of outdoor labour.

One: industrial manslaughter now applies everywhere#

Industrial manslaughter is an offence in every Australian state and territory and at the Commonwealth level, carrying a maximum penalty of imprisonment for individuals. The Australian Capital Territory legislated first in 2004 and Queensland followed in October 2017, but the bulk of jurisdictions arrived recently: the Northern Territory and Victoria in 2020, Western Australia in March 2022, the Commonwealth and South Australia on 1 July 2024, New South Wales in September 2024, and Tasmania in October 2024.

Several jurisdictions express corporate maximums in penalty units rather than dollars, and the unit value is indexed annually. Victoria's 100,000 units equate to roughly $20.9 million at the 2026-27 unit value. Verify current values against the relevant Act before relying on them.

The practical shift is who is exposed. These offences reach officers personally, not just the corporate entity. A director who cannot demonstrate that they exercised due diligence, meaning that they acquired knowledge of the hazards, ensured resources were available and verified that controls worked, carries individual criminal risk.

Two: engineered stone is banned, and silica rules now reach every industry#

On 1 July 2024 Australia became the first country in the world to prohibit engineered stone. Every jurisdiction amended the model work health and safety regulations to make it an offence for a person conducting a business or undertaking to carry out, direct or allow work involving the manufacture, supply, processing or installation of engineered stone benchtops, panels or slabs. A separate importation ban commenced 1 January 2025. Removal, repair, minor modification and disposal of legacy engineered stone remain permitted subject to controls and prior notification to the regulator.

Screening studies of Australian stone benchtop workers have found silicosis in a substantial minority of those tested. The wider lesson is the one worth carrying: a hazard with a long latency period was allowed to accumulate for years before it became visible, and then the regulatory response was total prohibition rather than gradual tightening.

Three: psychosocial hazards are an express duty#

Covered in section 7. The short version is that managing work pressure, bullying, fatigue, isolation and exposure to trauma is now a regulated obligation on the same footing as managing a physical hazard. Safe Work Australia's model code of practice on managing psychosocial hazards at work dates from July 2022; jurisdictions have adopted their own versions on different dates, and the Commonwealth made a separate code in November 2024 that adds fatigue, intrusive surveillance and job insecurity as named hazards.

Four: in New South Wales, codes of practice are now enforceable#

This is the change with the widest reach and the least publicity. From 1 July 2026, section 26A of the New South Wales Work Health and Safety Act creates a comply-or-justify obligation across roughly 48 approved codes. Failing to follow a code is now itself a breach, a Category 3 offence, rather than merely evidence that a duty was not met.

Five: heat regulation is arriving, jurisdiction by jurisdiction#

There is still no national heat standard. Safe Work Australia publishes a Guide for managing the risks of working in heat, updated to reflect revised first aid guidance for heat stroke, but a guide is not a regulation.

As at August 2026 the Australian Capital Territory is the only jurisdiction with an approved code of practice: the Work Health and Safety (Managing the Risks Associated with Extreme Temperatures Code of Practice) Approval 2025, notified 6 November 2025 and in force from 14 November 2025, covering both heat and cold. It sets no numeric temperature threshold, deliberately, and creates no new legal obligations; it is guidance on meeting existing duties. New South Wales, Victoria and Queensland have not introduced heat codes. The Australian Council of Trade Unions began campaigning for national heat rules in February 2026, seeking mandated rest breaks, work scheduling and stoppages at defined temperature and humidity thresholds.

Sitting above all five is the Australian Work Health and Safety Strategy 2023-2033, whose stated vision is safe and healthy work for all. The direction of travel across every jurisdiction is consistent: broader duties, higher penalties and personal accountability for officers.

The workforce squeeze#

Everything above lands on a workforce that is already too small and getting smaller. Australia cannot staff the outdoor work it has committed to, and the shortfall is widening faster than the forecasts anticipated.

Infrastructure Australia's 2025 Infrastructure Market Capacity Report, released November 2025, projects a national construction workforce shortfall of more than 300,000 workers by mid-2027, against a shortfall of 141,000 recorded as at October 2025. Peak workforce demand has been revised up to 521,000 workers from 417,000 in the previous year's assessment. Trades workers and labourers peak at a shortage of 126,000 in mid-2027.

Regional shortages are forecast to grow faster than metropolitan ones, from 38,200 in October 2025 to a peak of around 181,000 in 2027, roughly a fourfold increase, as renewable energy generation and transmission projects move outside the capital cities. This is significant for safety as well as delivery: regional and remote work carries longer emergency response times, more lone work, more driving and higher heat exposure.

The pipeline is not refilling#

Jobs and Skills Australia projects demand for 195,800 additional technicians and trades workers over the ten years to May 2035, and technicians and trades workers make up over half of all occupations in continuous shortage. The supply side is moving in the opposite direction.

• The base has fallen sharply. Apprentices and trainees in training peaked at 514,780 in June 2012. As at 31 December 2025 the figure was 282,430, down 8.6 per cent year on year. Trade apprentices account for 208,950 of that.
• Commencements are volatile and trending down. Over the twelve months to September 2025, trade apprenticeship commencements fell almost 10 per cent and traineeship commencements fell just over 18 per cent, though the December 2025 quarter alone showed trade commencements up 12.5 per cent year on year.
• Completion is the real leak. The trade individual completion rate is 63.4 per cent at six years for the 2019 commencing cohort. On a contract basis, which counts a worker who changes employer as a non-completion, it is 58.7 per cent.

And the workforce is ageing into higher risk#

The two curves compound. Fewer young workers enter, so the average age rises. And the fatality rate climbs steeply with age: 4.3 deaths per 100,000 workers among those aged 65 and over, against a national rate of 1.3. Claim severity follows a similar pattern, with median time lost rising from 4.0 weeks for workers under 25 to a peak of 10.3 weeks in the 55 to 64 band, and median compensation from $7,500 to a peak of $22,900 across the same span.

The ABS Work-related Injuries survey adds a useful nuance: the highest rate of work-related injury is in the 55 to 59 age group at 4.7 per cent, and the rate then falls away sharply for those 65 and over. Read together with the severity data, that suggests the oldest workers are less likely to be injured but considerably more likely to be seriously injured when an incident occurs.

The data nobody has ever collected#

Every number in the preceding nine sections describes harm after it has happened. Australia counts fatalities carefully, counts accepted claims carefully, and knows almost nothing about the conditions that produced either. The coronial file will tell you that a worker rolled a side-by-side on a slope. It will not tell you what the temperature was, how many consecutive days that person had been rostered on, whether the crew was short-handed that week, or how many times something similar had nearly happened on the same slope and gone unreported.

What a rostering app already knows#

This is where a workforce app turns out to be an unusually good research instrument, almost by accident. It is already holding the roster, so it knows the shift pattern and how many days in a row someone has worked. It is already holding clock-in and clock-out, so it knows the real hours rather than the planned ones. It knows roughly where the crew was, because that is how the timesheet gets approved. It can pull the forecast and the observed weather for that location and that day, because rostering outdoor work without the forecast is guesswork. And if you give people a simple way to report a near miss, it collects the thing that is missing from every official dataset in the country.

Individually none of those is remarkable. Together, across enough crews and enough seasons, they are the longitudinal exposure dataset that Australian occupational safety research has never had. The injury data has always been a numerator without a denominator: we know how many people were hurt, and almost nothing about the hours of ordinary risk that produced them.

Who it is worth something to#

• Workers, most of all. The point of collecting it is a safer shift, but the immediate return is much more mundane and much more useful day to day: an honest record of hours worked, shifts published further ahead, a fair account of a disputed week, a warning when tomorrow afternoon is going to be dangerous, and a personal view of their own patterns including step count, time on foot and time in the heat. None of that requires anyone else to see their individual data.
• Employers. A benchmark that is theirs rather than an industry average, and early sight of the patterns that precede incidents on their own sites rather than on somebody's national average.
• Insurers and schemes. The parties that actually carry the cost of Australian workplace injury currently price risk on industry code and claims history. Conditions data would let them price on what is happening rather than on what has already happened, which is the same shift that transformed motor insurance.
• Researchers. An anonymised, consented, longitudinal dataset of this kind does not exist for Australian outdoor work. Section 7 of the technical edition documents exactly how large that hole is.

Why this is a rostering problem before it is a safety problem#

Read back through the nine sections above and a pattern emerges that does not look like a safety pattern at all. Heat illness is what happens when a crew starts at ten instead of six on a day the forecast had already called. Fatigue is what happens on the eleventh consecutive day, and somebody wrote that eleventh day into a roster three weeks earlier without noticing. Lone working is a rostering decision. Being on foot in a yard while a truck reverses is a rostering decision. The psychosocial hazards in section 6 are, in a majority of the cases the regulators describe, about unpredictability of hours and of income, which is to say about how far ahead the roster gets published.

What that looks like in practice#

It is deliberately unremarkable software. Shifts get published further ahead, which is the single most requested thing by the people doing this work and the cheapest retention measure in section 9. The manager can see who is on their eleventh day without building a spreadsheet to work it out. Clock-in and clock-out come from a geofence that runs for the length of the shift rather than a button people forget to press, so timesheets are accurate and the pay run stops being a Thursday night job. The weather comes in against the site rather than against the postcode, and if tomorrow afternoon is going to be dangerous the alert lands on the dashboard the evening before, when moving the start still costs nothing.

Then, because the app is already open in the pocket of everyone on site, two things become possible that have never been practical before. A near miss can be reported in about twenty seconds, anonymously if that is what it takes, and the psychosocial ones can be reported the same way. And the smartphone's own motion sensors can tell a pocketed phone from one held in a hand, which turns the hazard in section 5 from an anecdote into a number a site can actually manage.

None of the controls in the middle column are wrong. They are simply downstream of the decisions in the right hand column, and the right hand column is a rostering system. An employer who improves the roster gets the safety improvement, the retention improvement described in section 9, and the dispute reduction described at the end of it, from one change rather than three programs.

Nine things employers can do this quarter#

None of the following requires new legislation, new technology or a large capital budget. Each maps directly to a finding in this report.

Sources and limitations#

Known limitations#

• Outdoor work is not measured directly by any Australian statistical agency. The 2.7 million figure is an industry-based proxy that both over-counts indoor roles within outdoor industries and under-counts outdoor roles within indoor industries.
• Occupational disease is systematically under-represented. Skin cancer, silicosis and hearing loss have latency periods measured in decades and are excluded from fatality statistics entirely.
• Only 31 per cent of Australians who sustain a work-related injury receive workers compensation, so claims data captures under a third of the real injury burden. That survey reports 2021-22 and the next release has been deferred to 2026-27.
• The Australian Work Exposures Study ultraviolet exposure figures were collected in 2011 and 2012. They remain the most recent national measurement.
• Contractors, labour hire workers, gig-model delivery riders and seasonal harvest labour are inconsistently captured across both employment and compensation datasets.
• AgHealth Australia farm incident data is derived from media reporting and includes non-workers such as children and bystanders. It should not be read as equivalent to Safe Work Australia's worker fatality series, and it is sensitive to changes in reporting intensity.
• Section 7 draws an inferential link between mental health claim drivers and outdoor working conditions. The published claims data is not split by indoor and outdoor work, and that connection should be read as reasoned interpretation rather than a measured finding.
• Claim counts and frequency rates are 2023-24 preliminary and will be revised upward as open claims finalise. Median time lost and median compensation are quoted from 2022-23, the most recent non-preliminary period.

Sources#

1. Safe Work Australia, Key Work Health and Safety Statistics Australia 2025, October 2025. Source for all fatality and serious claims figures unless otherwise noted. data.safeworkaustralia.gov.au/insights/key-whs-statistics-australia/latest-release
2. Australian Bureau of Statistics, Labour Account Australia, March quarter 2026. abs.gov.au/statistics/labour/labour-accounts/labour-account-australia/latest-release
3. Australian Bureau of Statistics, Work-related Injuries, 2021-22. abs.gov.au/statistics/labour/earnings-and-working-conditions/work-related-injuries/latest-release
4. AgHealth Australia and AgriFutures Australia, Non-intentional farm-related incidents in Australia, annual statistics 2024. aghealth.sydney.edu.au/projects-and-reports/research-reports
5. Mesa-Castrillon C, Peachey K-L, Lower T. Farm injury deaths and workers' compensation claims in Australia and their economic costs. Australian Journal of Rural Health 2025;33(5):e70087. doi.org/10.1111/ajr.70087
6. Cancer Council Australia, National Cancer Prevention Policy: Skin Cancer, Settings: Workplace. cancer.org.au
7. Carey RN, Glass DC, Peters S, Reid A, Benke G, Driscoll TR, Fritschi L. Occupational exposure to solar radiation in Australia: who is exposed and what protection do they use? Australian and New Zealand Journal of Public Health 2014;38(1):54-9. doi.org/10.1111/1753-6405.12174
8. Schmitt J, Seidler A, Diepgen TL, Bauer A. Occupational ultraviolet light exposure increases the risk for the development of cutaneous squamous cell carcinoma. British Journal of Dermatology 2011;164(2):291-307. doi.org/10.1111/j.1365-2133.2010.10118.x
9. Bauer A, Diepgen TL, Schmitt J. Is occupational solar ultraviolet irradiation a relevant risk factor for basal cell carcinoma? British Journal of Dermatology 2011;165(3):612-25. doi.org/10.1111/j.1365-2133.2011.10425.x
10. Australian Government Department of Climate Change, Energy, the Environment and Water, National Climate Risk Assessment, September 2025. dcceew.gov.au/climate-change/policy/adaptation/ncra
11. WorkSafe ACT, Data snapshot: Working in heat. worksafe.act.gov.au
12. Infrastructure Australia, 2025 Infrastructure Market Capacity Report, November 2025. infrastructureaustralia.gov.au
13. Jobs and Skills Australia, 2025 Employment Projections and 2025 Occupation Shortage List. jobsandskills.gov.au
14. National Centre for Vocational Education Research, Apprentices and trainees 2025: December quarter and Apprentice and trainee completion rates 2025. ncver.edu.au
15. Safe Work Australia, Crystalline silica substances and Engineered stone ban. safeworkaustralia.gov.au/safety-topic/hazards/silica
16. Hoy RF, Dimitriadis C, Abramson M, Glass DC, Gwini S, Hore-Lacy F, Jimenez-Martin J, Walker-Bone K, Sim MR. Prevalence and risk factors for silicosis among a large cohort of stone benchtop industry workers. Occupational and Environmental Medicine 2023;80(8):439-446. doi.org/10.1136/oemed-2023-108892
17. Safe Work Australia, Australian Work Health and Safety Strategy 2023-2033. safeworkaustralia.gov.au/awhs-strategy_23-33

This report synthesises publicly available Australian government and research data. It is general information about workplace safety trends and is not legal advice. Penalty figures, regulatory obligations and code of practice requirements vary by jurisdiction and change frequently; verify current requirements with the work health and safety regulator in your jurisdiction before relying on anything in this document. Prepared August 2026.

How this report was made#

This report is a synthesis of published national datasets, not an original survey. Every number in it is traceable to a named public source listed in section 11, and every one of those sources is free to access.

Where a figure carries a caveat, that caveat is printed alongside it rather than buried in a footnote. Outdoor work is not a category the Australian statistical system measures directly, so section 2 explains exactly how the workforce was approximated and where that approximation breaks down.

Statistical vintages are stated wherever two datasets are placed side by side, because mixing them is the most common way a report of this kind goes wrong. Fatality data is calendar year 2024. Claim counts and frequency rates are financial year 2023-24 preliminary. Median time lost and median compensation are financial year 2022-23, the most recent period for which those medians are published.

If you find an error, we would like to know, and it will be fixed.