Calculate accident frequency rates from cases and employee hours, or estimate hours from staffing to report AFR, TRIR, LTIFR, or DART per work hours.

Use matching case counts and employee hours from the same reporting period and population.

The same formula does not make different case definitions interchangeable.

Standard basis uses 200,000 for TRIR/DART and 1,000,000 for AFR/LTIFR. Check your reporting standard.

Optional settings

Tiny nonzero rates use scientific notation instead of rounding to zero.


Related Calculators

Accident Frequency Rate Formula

The calculator uses a standard frequency rate formula. The base value changes the scale of the result but not the underlying ratio.

AFR = (Number of Accidents / Total Hours Worked) ร— Base
  • AFR = Accident Frequency Rate; the same rate structure is used with the case definitions for TRIR, LTIFR, or DART
  • Number of Accidents = count of cases meeting the selected metricโ€™s definition in the period
  • Total Hours Worked = sum of hours worked by all employees in the period
  • Base = 1,000,000 hours (common AFR / LTIFR), 200,000 hours (OSHA TRIR / DART), or 100,000 hours

If you do not know the exact hours worked, the calculator estimates them from staffing:

Total Hours = Employees ร— Hours per Person ร— Period Length
  • Employees = average headcount or full-time equivalents (FTE)
  • Hours per Person = average hours per week (x52) or per year
  • Period Length = years, months (/12), or weeks (/52)

Assumptions: hours include only time actually worked, not vacation or sick leave. The illustrative 40-hour weekly default equals 2,080 hours per 52-week year; adjust it to actual exposure. Choosing FTE does not overwrite the entered schedule. Use the same case definition (all accidents, recordables only, lost-time only, or DART) consistently across periods you compare.

Calculator modes:

  • Known hours applies the first formula directly. Enter cases and total hours, pick a base, and get the rate.
  • Estimate hours runs the second formula first to build total hours, then feeds that into the first formula. Use this when payroll hours are not available.
  • Metric selector (AFR, TRIR, LTIFR, DART) identifies the case definition and chooses its usual base when the automatic base option is selected. Manual base choices rescale the same cases; they do not convert one case definition into another.

Reference Tables

The industry table reports 2023 private-sector total recordable case rates from BLS Table 1. Match the industry, year and case definition when comparing rates.

Metric Cases Counted Typical Base
AFRAll workplace accidents1,000,000 hours
TRIROSHA recordable injuries and illnesses200,000 hours
LTIFRInjuries causing lost work time1,000,000 hours
DARTDays Away, Restricted, or Transfer200,000 hours
Industry (US private sector, 2023) TRIR per 200,000 hrs
Construction2.3
Manufacturing2.8
Warehousing and storage4.7
Healthcare and social assistance3.6
Retail trade3.1
Professional and business services1.1

Worked Examples and FAQ

Example 1 โ€” Known hours. A site logged 6 accidents over 1,500,000 hours worked.

AFR = (6 / 1,500,000) * 1,000,000 = 4.00 per million hours.

Example 2 โ€” Estimating hours. 750 employees, 40 hours per week, over 1 year.

Total hours = 750 * 40 * 52 = 1,560,000. With 6 accidents: AFR = (6 / 1,560,000) * 1,000,000 = 3.85 per million hours.

Example 3 โ€” OSHA TRIR. 4 recordables, 410,000 hours.

TRIR = (4 / 410,000) * 200,000 = 1.95.

Why does the base matter? The base only rescales the rate. For the same recordable case count and hours, 2.0 per 200,000 hours equals 10.0 per 1,000,000 hours. This rescaling does not turn recordable cases into all accidents or lost-time cases.

Should I include contractors? Include contractor cases and contractor hours together, or exclude both. Mixing the two will distort the rate.

What hours count? Only hours actually worked. Do not include vacation, sick days, or holidays. If records are unavailable, estimate actual hours worked, excluding leave and including overtime, under the applicable reporting guidance.

Is a lower number always better? Yes, mathematically. But very low rates can also signal under-reporting. Pair frequency rates with leading indicators like near-miss reports and inspection findings.