Accident Frequency Rate Calculator
Five accidents in a year sounds small until you ask the only question that matters: small compared with what? A site with 50 workers and a site with 5,000 workers cannot compare raw accident counts, and that is exactly the problem the accident frequency rate (AFR) solves. It divides your accident count by the hours your people actually worked, then scales the result to a standard base, so a small plant and a national contractor end up speaking the same language.
The Accident Frequency Rate Calculator below gives you every standard view of that number at once: the international rate per 1,000,000 hours, the OSHA-style rate per 200,000 hours, the incident rate per 100 employees, and a safety rating against benchmark bands for 13 industries. Tell it what counts as an accident, pick your industry, and it also writes out practical next steps based on where your rate lands.
Safety recommendations
Quick Answer
Key takeaways
- AFR = accidents x 1,000,000 / hours worked. The OSHA-style rate uses 200,000 instead, so it is always the million-hour rate divided by 5.
- The 200,000 base is not arbitrary: it is 100 employees working 40 hours a week for 50 weeks, the convention the Bureau of Labor Statistics uses for incidence rates per 100 full-time workers.
- Definition beats arithmetic. A rate that includes first-aid cases will always run higher than one built on OSHA recordables only, so state your definition next to every AFR you publish.
- There is no universal good AFR. Judge the number inside your industry band and against your own trend, and never celebrate a fall that came from under-reporting.
- Quick Answer
- What the accident frequency rate actually measures
- The formula, step by step
- Worked example: 5 accidents in 800,000 hours
- AFR vs LTIFR, TRIFR, TRIR and DART: which rate is which
- Industry benchmarks: what counts as a good AFR
- What competing calculators miss
- Frequently asked questions
- Add This Calculator to Your Website
- References
What the accident frequency rate actually measures
AFR is a lagging indicator: it counts harm that has already happened, scaled by exposure. The scaling is the whole point. Raw counts reward small sites and punish large ones, while a rate per million hours puts both on the same footing, provided they count the same kinds of events over the same period. Safety teams use it in board reports, contractor prequalification, and insurance discussions, and they track it month by month to see whether interventions are working or whether a good quarter was just luck.
The Occupational Safety and Health Administration (OSHA) frames the US version through its recordkeeping rules: a case is recordable when it involves death, days away from work, restricted work or job transfer, medical treatment beyond first aid, loss of consciousness, or a significant injury or illness diagnosed by a clinician. If your AFR definition follows those criteria, your million-hour figure is the same family as the total recordable incident rate (TRIR) that US clients ask for, just on a different base. If your definition also sweeps in first-aid cases, your AFR will sit noticeably higher, and comparing it with a recordables-only benchmark is a category error, not a safety failure.
One caution belongs in every safety meeting: AFR says nothing about severity. Five fractures and five bandaged cuts produce the same rate if both sets meet your definition. That is why mature programs pair AFR with a severity measure and with leading indicators such as near-miss reports and inspection findings, instead of managing to a single number.
The formula, step by step
Three short calculations sit behind the result box, and you can check each one on paper. The calculator runs all three from the same inputs so the views can never drift apart.
Fix the definition and the period first. Decide which cases count (all accidents, recordables only, lost-time only, fatalities only, or first-aid cases) and lock the period, usually a month, a quarter, or a rolling year. Count accidents and hours for exactly the same people and the same dates.
Total the hours actually worked. Use payroll or timesheet hours, including overtime, and leave out vacation, holidays, and sick leave, because nobody can be injured at work while away from it. If you only know headcount, the estimator in the calculator uses the OSHA convention of 2,000 hours per full-time employee per year. An overtime hours calculator helps when heavy overtime makes the headcount estimate unreliable.
Scale the count to the base you need. Multiply accidents by 1,000,000 for the international rate or by 200,000 for the OSHA-style rate, then divide by hours worked. For the headcount view, divide accidents by employees and multiply by 100.
Worked example: 5 accidents in 800,000 hours
A manufacturing site records 5 accidents over 800,000 hours worked by 100 employees, counting all accidents. Here is exactly what the calculator returns for those pre-filled defaults, verified against its code:
| Figure | Calculation | Value |
|---|---|---|
| Accident frequency rate | 5 x 1,000,000 / 800,000 | 6.25 |
| OSHA recordable rate | 5 x 200,000 / 800,000 | 1.25 |
| Incident rate | 5 / 100 x 100 | 5.00 |
| Safety rating (manufacturing) | 6.25 vs bands: excellent 0.5, good 2.0, average 4.0, poor 8.0 | Below average: needs attention |
Read the three rates as one story told at three zoom levels. The site has an accident for every 160,000 hours worked, which the million-hour base expresses as 6.25. On the OSHA base the same record reads 1.25 recordables per 100 full-time workers, comfortably below the Bureau of Labor Statistics private-industry average of 2.3 total recordable cases per 100 FTE in 2024, a reminder that US benchmark tables and international AFR bands come from different reporting cultures. The headcount view is the bluntest: 5 accidents among 100 employees means 1 worker in 20 was involved in a counted accident during the period.
The bars use the manufacturing poor threshold of 8.0 as the full scale. Your example site sits between the average band and the poor line, which is why the calculator pairs the rating with a concrete instruction for manufacturing above 2.0: review machine guarding, lockout and tagout, and near-miss reporting before the rate climbs further.
AFR vs LTIFR, TRIFR, TRIR and DART: which rate is which
The safety alphabet is simpler than it looks once you separate two choices: which cases are counted, and which base the rate sits on. Change either one and the number moves, even when nothing on the shop floor has changed.
| Metric | Cases counted | Base | Notes |
|---|---|---|---|
| AFR | Whatever your written definition says | 1,000,000 hours | International default. May include first aid, so it often runs highest. |
| LTIFR | Lost-time injuries only | 1,000,000 hours | The narrowest frequency rate. Misses recordables that caused no lost time. |
| TRIFR | All OSHA recordables | 1,000,000 hours | Matches AFR when the AFR definition is recordables only. |
| TRIR / incident rate | All OSHA recordables | 200,000 hours | The US standard. Equals TRIFR divided by 5. |
| DART rate | Recordables with days away, restriction, or transfer | 200,000 hours | A severity-flavoured subset of TRIR. Always at or below TRIR. |
| Incident rate (this tool) | Your counted accidents | Per 100 employees | Headcount view. Simple, but distorted by overtime and part-time mixes. |
Two conversion habits prevent most reporting errors. First, never place a million-hour figure next to a 200,000-hour benchmark without dividing or multiplying by 5. Second, never compare two rates built on different case definitions: a first-aid-inclusive AFR against a recordables-only TRIR will manufacture a gap that no safety program caused. OSHA's general recording criteria in 29 CFR 1904.7 are the reference point when a client or auditor asks which cases belong in the recordable count.
Industry benchmarks: what counts as a good AFR
There is no universal good AFR, because exposure and reporting culture differ by sector. The calculator rates your result against the bands below, the same bands the live tool has always used. Use them as a first read, then check your sector against the Bureau of Labor Statistics employer-reported injury and illness release for US recordable rates by industry.
| Industry | Excellent | Good | Average | Poor (above) |
|---|---|---|---|---|
| Oil and gas | 0.5 or less | 1.0 | 2.5 | 5.0 |
| Office / administrative | 0.1 or less | 0.5 | 1.0 | 2.0 |
| Utilities | 0.5 or less | 1.5 | 3.0 | 6.0 |
| Manufacturing | 0.5 or less | 2.0 | 4.0 | 8.0 |
| Mining | 0.8 or less | 2.5 | 5.0 | 10.0 |
| Retail | 1.0 or less | 2.5 | 5.0 | 10.0 |
| Construction | 1.0 or less | 3.0 | 6.0 | 12.0 |
| Food service | 1.5 or less | 3.0 | 6.0 | 12.0 |
| Transportation | 1.5 or less | 3.5 | 7.0 | 14.0 |
| Warehousing | 1.5 or less | 4.0 | 8.0 | 16.0 |
| Healthcare | 2.0 or less | 4.0 | 8.0 | 16.0 |
| Agriculture | 2.0 or less | 5.0 | 10.0 | 20.0 |
| Other | 1.0 or less | 3.0 | 6.0 | 12.0 |
Read the table with one eye on your own trend. A warehousing operation that moves from 14 to 9 has made real progress even though 9 still sits above the average band, while an office drifting from 0.4 to 1.8 has a problem the absolute number hides. Accident patterns also feed the absence figures managers watch: tracking your rate alongside an absence rate calculator shows whether injuries are turning into lost shifts, and an absentee payroll calculator puts a cost on those lost hours when you build the case for prevention spending.
Where an hours-based AFR helps
- Fair across sizes: a 40-person workshop and a 4,000-person group compare on one scale.
- Handles overtime and part-time work honestly, because exposure is hours, not heads.
- One input set produces the international, OSHA, and headcount views, so reports stay consistent.
- Trend-friendly: monthly rates expose drift long before annual counts do.
Where AFR can mislead
- It is a lagging indicator: it only moves after someone has been hurt.
- It ignores severity, so minor and life-changing injuries weigh the same.
- Small sites swing wildly: one accident in 40,000 hours is an AFR of 25.
- It collapses if reporting is punished: a falling AFR built on silence is worse than a high honest one.
What competing calculators miss
We checked the top-ranking tools for this search before rebuilding this page: dedicated AFR pages, TRIR calculators, and safety software guides. The formulas agree everywhere. The gaps are in everything around the formula.
Winning move 1: every base on one screen, not one base per tool. The best-known AFR competitor returns only the two hour-based figures from two inputs, the leading TRIR tools return only the 200,000-hour figure, and one AFR specialist makes you rerun the calculation to switch views. None of them adds the per-100-employee incident rate. This calculator returns all four views from a single entry, so the number you send to a US client, an international parent company, and your own workforce meeting is the same record told three ways, with no conversion errors in between.
Winning move 2: a severity selector that travels with the result. Most tools silently assume OSHA recordables, then let you benchmark a first-aid-inclusive count against recordable rates. Here the severity field (all accidents, recordables, lost-time, fatalities, first aid) is part of the input, is named in the copied results, and triggers its own guidance: a lost-time AFR above 1.0 gets flagged, and any fatality count above zero gets an investigation instruction instead of a bland rating.
Winning move 3: an industry rating and next steps, not a naked number. One major calculator grades TRIR with three universal bands (low under 3, high over 8) that ignore industry entirely, which rates a strong construction firm as mediocre and a weak office as safe. This tool rates your AFR against bands for 13 industries and prints tailored recommendations, such as the manufacturing prompt to review machine guarding and lockout/tagout. It also includes the headcount hours estimator (employees x 2,000) inside the card, which none of the AFR competitors offer, for the many small firms that know their headcount but not their exact hours.
Frequently asked questions
It depends on industry and definition: strong oil and gas programs run well below 1 per million hours, manufacturing averages around 4, and construction around 6. Compare against your industry band and your own trend, not an absolute number.
Use 1,000,000 for international reporting and 200,000 for OSHA-style US reporting. The calculator shows both, so you do not have to choose. Never mix bases without converting (divide or multiply by 5).
Whatever your written definition says: typically recordable injuries, sometimes including first aid or near misses. Write the definition next to every AFR you publish and keep it consistent year to year.
Multiply full-time headcount by 2,000 hours per year (40 hours x 50 weeks), the OSHA convention. Prefer timesheet or payroll hours when you have them; the estimator button above does the headcount math for you.
AFR scales accidents by hours worked (per million or per 200,000); incident rate scales by headcount (per 100 employees). Hours-based rates are fairer when overtime or part-time work varies.
Include them if contractors work under your supervision and your reporting scope covers them, and keep that choice consistent every period. Many clients require contractor accidents and hours in project statistics. State the scope on every report so the rate stays comparable.
Monthly for internal trending and annually for official reporting and benchmarking. Small sites should use a rolling 12-month window, because a single accident in a short period swings a million-hour rate sharply: one accident in 100,000 hours alone produces an AFR of 10.
Add This Calculator to Your Website
References
- Occupational Safety and Health Administration, "Recordkeeping" - OSHA guidance on which injuries and illnesses employers must record, including the recordable criteria used for incidence rates. OSHA: Recordkeeping. Accessed October 2026.
- Occupational Safety and Health Administration, 29 CFR 1904.7, "General recording criteria" - the regulation defining recordable cases: death, days away, restricted work, medical treatment beyond first aid, and loss of consciousness. OSHA: 29 CFR 1904.7 General Recording Criteria. Accessed October 2026.
- U.S. Bureau of Labor Statistics, "How to Compute Your Firm's Incidence Rate for Safety Management" - the BLS method behind the 200,000-hour formula and rates per 100 full-time workers. BLS: How to Compute Your Firm's Incidence Rate. Accessed October 2026.
- U.S. Bureau of Labor Statistics, "Employer-Reported Workplace Injuries and Illnesses" news release - annual national incidence rates, including the 2024 private-industry total recordable rate of 2.3 per 100 full-time workers. BLS: Employer-Reported Workplace Injuries and Illnesses. Accessed October 2026.
- U.S. Bureau of Labor Statistics, "Handbook of Methods: Survey of Occupational Injuries and Illnesses" - how the national injury and illness rates behind industry benchmarks are collected and calculated. BLS: Handbook of Methods, SOII. Accessed October 2026.
- Occupational Safety and Health Administration, "Report a Fatality or Severe Injury" - employer duties to report fatalities within 8 hours and severe injuries within 24 hours. OSHA: Report a Fatality or Severe Injury. Accessed October 2026.