Workplace Injury Statistics 2026: 2.5M Injuries, 5,070 Deaths, 1 Every 104 Minutes

Workplace injury figures get quoted constantly, and they are quoted wrong almost as often. The two US government series that most people mean — the Survey of Occupational Injuries and Illnesses (SOII) and the Census of Fatal Occupational Injuries (CFOI) — cover different workers, use different units and are released a month apart. A rate per 100 workers sits next to a rate per 100,000, and private-industry counts are mixed with all-worker counts. This page puts the latest release of each side by side, with the population and period of every number, so you can cite the right one.

Every figure below was checked against the primary publication: BLS for US injuries and deaths, OSHA for penalties and incentive-programme rules, the National Academy of Social Insurance (NASI) for workers' compensation, and the UK Health and Safety Executive (HSE) and International Labour Organization (ILO) for the wider view. The latest complete US data year is 2024; BLS publishes 2025 data from late 2026 into early 2027. Our own derived metrics are labelled and their arithmetic shown.

Workplace injury benchmark cheat sheet

The numbers most often needed for a safety report, a board paper or an article. Full sources are in the methodology section.

1. How many workplace injuries happen each year

US private employers reported 2,488,400 recordable injuries and illnesses in 2024, according to the BLS Survey of Occupational Injuries and Illnesses. That is 3.1% fewer than in 2023 and the lowest count in the series since 2003. A case is "recordable" under OSHA rules when it needs more than first aid: medical treatment, days away, restricted work, loss of consciousness or a significant diagnosis.

Source: BLS SOII, Table 1, private industry, published in thousands and shown here in full. Excludes farms with fewer than 11 employees, the self-employed and federal workers.

Three readings of the table matter. Over ten years, total cases fell 14.4% and injuries alone fell 15.4% (our calculation, 2015 to 2024). The 2020–2022 bulge is illnesses, not injuries: respiratory cases jumped from 10,800 in 2019 to 428,700 in 2020 as COVID-19 cases were recorded. And illnesses have not returned to baseline — respiratory cases in 2024 were still 5.0× the 2019 level (54,000 vs 10,800), even after a 46.1% fall in one year.

Days away from work

The most serious nonfatal cases are those involving days away from work, job transfer or restriction (DART). For 2023–2024 combined, BLS counted 1.8 million days-away-from-work cases (61.5% of DART cases, an annualised 86.6 per 10,000 full-time workers) and 1.1 million job transfer or restriction cases (38.5%, 54.2 per 10,000). The median days-away case kept the worker off for 8 days; the median restriction lasted 15 days.

By cause, overexertion, repetitive motion and bodily conditions led with 946,290 DART cases over the two years, followed by contact incidents (being struck by or caught in objects and equipment) with 860,050. That is why the cheapest safety wins are often ergonomic: lifting aids, rotation and rest breaks prevent the single largest group of serious injuries. For how injuries feed into wider absence, see our employee absenteeism statistics.

2. Injury rates by industry

The private-industry average of 2.3 cases per 100 full-time workers hides a tenfold spread between desk sectors and physical ones. Rates below are total recordable cases and days-away cases per 100 full-time equivalent workers in 2024.

Source: BLS SOII 2024, Table 1 (incidence rates by industry, private industry). Multiples are ours (rate ÷ 2.3). A dash means BLS did not publish the figure.

Health care is the uncomfortable headline. Hospitals (5.1) and nursing homes (5.5) report more injuries per worker than construction (2.2), largely from patient handling and workplace violence. The health care and social assistance sector improved from 3.6 to 3.4 in 2024, and BLS notes that no sector's rate rose that year. The public sector runs hotter than private: including state and local government lifts the all-industry rate to 2.6. If you run reward programmes for clinical staff, our guide to employee rewards in healthcare covers what works on shift-based teams.

3. How many workers die on the job

BLS counted 5,070 fatal work injuries in 2024, down 4.0% from 5,283 in 2023 and 7.6% below the 2022 figure of 5,486 (our calculation). The fatal injury rate fell to 3.3 per 100,000 full-time equivalent workers. Unlike SOII, the census covers every US worker, including the self-employed and government employees, and counts injuries only — deaths from occupational illness are not included.

Sources: BLS CFOI news releases of February 19, 2026 (2024 data) and December 19, 2024 (2023 data, with 2019–2022 totals in Table 1). Dashes mark values not taken from those two releases.

How workers die: fatal injuries by event, 2024

Source: BLS CFOI 2024 release. The 38.2% transportation share is stated by BLS; the other shares are ours (deaths ÷ 5,070). Contact with objects and equipment, and fires and explosions, make up most of the remainder.

Two movements stand out. Overdose deaths at work fell 19.9% in one year (512 to 410, our calculation), the main reason harmful-exposure deaths dropped 16.2%. And pedestrian deaths rose 19.0% while roadway crashes fell 8.5% — workers struck on foot, often in roadside, warehouse-yard and delivery settings, are the fastest-growing transport risk in the data. Women accounted for 8.1% of work deaths (413) but 15.3% of homicide victims (72). Hispanic or Latino workers had 1,229 deaths and a rate of 4.3 per 100,000, above the 3.3 average; 68.5% of those deaths were among foreign-born workers.

Falls caused 370 deaths among construction and extraction workers in 2024 — down 7.5% in a year, but still more than one a day.

4. The most dangerous jobs in America

Occupational fatality rates show where risk concentrates far better than industry averages do. The table lists the highest-rate occupations BLS publishes for 2024, with our multiple of the all-worker rate.

Source: BLS CFOI chart "Rate and number of fatal work injuries in selected occupations, 2024". Multiples are ours.

Rate and volume tell different stories. Logging is the most dangerous job per worker, but driving kills the most people: 950 truck and driver/sales workers died in 2024, 18.7% of all US work deaths (our share). The transportation and material moving group as a whole lost 1,391 workers (12.5 per 100,000), and construction and extraction 1,032. Building and grounds cleaning and maintenance (356) and protective service occupations (281) were the two groups where deaths rose. Much of this risk sits with the frontline and deskless workforce — see our frontline and deskless workforce statistics.

5. Original metric: the severity index — why low injury rates can mislead

Comparing the two BLS series by industry exposes a pattern that neither shows alone. Some industries report few injuries but a lot of deaths; others report many injuries that are rarely fatal. We call the ratio the severity index: the fatal injury rate per 100,000 full-time workers divided by the total recordable case rate per 100. A higher number means a larger share of the harm in that industry is lethal.

Inputs: BLS CFOI "Number and rate of fatal work injuries, by selected private industries, 2024" and BLS SOII 2024 Table 1. Index and ranking are ours.

Caveats, because this is a comparison across two surveys: CFOI includes the self-employed (heavily present in agriculture and construction) while SOII does not, and CFOI publishes education and health services as one sector. The index is therefore a ranking tool, not a probability. The pattern it reveals — extraction, agriculture and construction at the top, care and retail at the bottom — holds under either caveat.

6. What workplace injuries cost

NASI notes that benefits and employer costs, adjusted for covered payroll, have continued to fall, and that 43 jurisdictions saw employer costs per $100 of payroll decline between 2022 and 2023. Benefits paid are only the visible part of the bill: they exclude overtime to cover absent colleagues, hiring and training replacements, damaged equipment and lost output, all of which land on the employer directly. For the retention side of that cost, see our employee turnover statistics.

Regulatory exposure: OSHA penalties in 2026

For penalties assessed after January 15, 2026, OSHA's inflation-adjusted maximums are $16,550 per serious or other-than-serious violation, $16,550 per day for failure to abate, and $165,514 per willful or repeated violation — exactly 10× the serious maximum (our ratio).

7. Workplace injuries outside the US

The ILO figure is the one to remember when injury data is used to argue that workplaces are now safe: nine in ten work-related deaths worldwide come from disease, not accidents (2.6 million of nearly 3 million), with circulatory diseases, cancers and respiratory diseases the top three causes. The British figures make the same point from the other side — stress, depression or anxiety now account for about half of all work-related ill health, which is why injury statistics belong next to our employee wellbeing statistics and burnout statistics, not apart from them.

8. Safety incentive programmes: what the rules allow

Rewards and safety meet in one specific, regulated place. OSHA's recordkeeping rule (29 CFR 1904.35(b)(1)(iv)) prohibits retaliation against workers for reporting an injury, and incentive programmes that discourage reporting can breach it. OSHA's October 11, 2018 enforcement memo clarified what is allowed:

• Rewarding workers for reporting near misses or hazards is always permissible. So are rewards for identifying unsafe conditions in the workplace.
• Rate-based programmes — rewards for injury-free periods — are permissible only if they are not run in a way that discourages reporting. A non-retaliation statement on its own may not be enough; OSHA points to reporting-rights training and a way to check whether employees actually feel free to report.

For what incentive spending looks like more broadly, see our employee incentive program statistics.

Methodology and sources

Every figure on this page was checked against the primary publication on 2 November 2026. US nonfatal data come from the BLS Survey of Occupational Injuries and Illnesses, an employer survey based on OSHA recordkeeping that covers private industry and state and local government but excludes the self-employed, farms with fewer than 11 employees and federal agencies. US fatal data come from the BLS Census of Fatal Occupational Injuries, a count compiled from death certificates, workers' compensation and other records that covers all workers but only fatal injuries, not deaths from illness. We did not use vendor or insurer estimates whose methods we could not open and check.

Derived metrics (ours, recomputable from the tables above): the injury clock (2,488,400 cases ÷ 31,536,000 seconds); ten-year change in cases and injuries (2015 vs 2024); the 2024-to-2019 respiratory multiple; occupational multiples of the all-worker fatal rate; the truck driver share of deaths (950 ÷ 5,070); event shares; the severity index (CFOI rate per 100,000 ÷ SOII rate per 100); benefits per covered job; and the OSHA willful-to-serious ratio. Inputs are published rounded, so derived ratios carry rounding error of a few percent.

Key takeaways

• US private employers recorded 2.5 million injuries and illnesses in 2024, 2.3 per 100 workers — the lowest count since 2003 and 14.4% below 2015.
• 5,070 workers died from work injuries, one every 104 minutes; transportation incidents caused 38.2% of the deaths.
• Hospitals and nursing homes injure more workers per head than construction, while logging, fishing and roofing are 15–34× deadlier than the average job.
• Low injury rates are not the same as safety: mining's severity index of 11.5 is eight times the all-industry reference.
• Rewards should target reporting, not silence — OSHA always permits rewarding near-miss and hazard reports, and scrutinises injury-free bonuses.