The formula
OSHA's recordkeeping forms package gives it: the total recordable case rate is the total number of injuries and illnesses times 200,000, divided by the number of hours worked by all employees; the DART rate does the same with the cases in columns H and I (OSHA forms package). BLS explains the 200,000: it is the equivalent of 100 employees working 40 hours a week, 50 weeks a year, so the rate reads as cases per 100 full-time workers (bls.gov).
TRIR = (G + H + I + J) × 200,000 ÷ hours worked
DART = (H + I) × 200,000 ÷ hours worked
Lost-time rate, on forms that ask for one = H × 200,000 ÷ hours worked
G is deaths, H is cases with days away from work, I is cases with job transfer or restriction, J is other recordable cases. Hours worked is the total on the same 300A.
Every number in the formula is already on your 300A, which totals the log's columns and adds the total hours worked by all employees covered by the log (1904.32(b)(2)). If you don't have a 300A for a year, see the 300A for prequalification.
Worked examples
BLS's own example: a company with 7 recordable cases and 400,000 hours worked has a rate of 7 × 200,000 ÷ 400,000 = 3.5; with 3 of those cases involving days away, job transfer or restriction, the DART rate is 1.5 (bls.gov).
A 10-person crew working a full year logs about 20,000 hours. Then:
- No recordable cases: TRIR 0.0.
- One case that needed stitches, no time lost: 1 × 200,000 ÷ 20,000 = TRIR 10.0, DART 0.0.
- The same case with a day away from work: TRIR 10.0, DART 10.0.
- One case over three years, 60,000 hours: 1 × 200,000 ÷ 60,000 = 3.3.
BLS put 2024 rates at 2.2 total recordable cases and 1.3 DART cases per 100 full-time workers for construction, 2.3 and 1.4 for specialty trade contractors, and 2.7 and 1.7 for manufacturing (bls.gov). With few hours, one case moves your rate far past those averages. That is why reviewers look at three years, and why the honest answer for a bad year is the 301 form, the investigation and the fix. BLS also suggests comparing your rate with employers of your own industry and size rather than with one number.
When a form defines it differently
Forms don't all use the same words or columns. Howard University's questionnaire defines TRIR and DART with the 200,000 formula and asks for the counts and the employee hours behind them, excluding vacations and holidays (ehs.howard.edu). McGough's sample defines its own recordable and lost-time rates from 300A columns (mcgough.com). Torcon asks for raw counts by 300-log column and man-hours, and does the math itself (torcon.com). Highwire calculates three-year average TRIR and DART rates from the 300A data you enter and compares them with the BLS industry average for your trade (help.highwire.com).
The rule of thumb: follow the definition printed on the form in front of you, show the counts and the hours next to every rate, and keep a worksheet so you can explain any number a reviewer asks about.
Mistakes that make numbers disagree
- Hours with vacation and holidays in them. OSHA's instructions say to leave out vacation, sick leave, holidays and any other non-work time.
- Missing workers. Hours include salaried, hourly, part-time, seasonal and temporary workers you supervise day to day.
- First aid counted as recordable. First-aid-only cases don't go on the log. Our OSHA 300 log guide has the first-aid list.
- A case counted twice. Each case goes in one column, its most serious outcome; the forms package ranks column G as the most serious and J as the least.
- Mixed years. Calendar years from the 300A, not your fiscal year or your policy year.
- Different rounding on each form. Pick one decimal place and use it everywhere.
- A rate with no hours behind it. A reviewer can't check it, so it reads as a guess.
Rather have the 29Safety safety team write it? The Bid and prequal packet, $795:
- Your company safety manual
- Up to 8 written programs from your GC's list
- A site-specific safety plan for one job
- English and Spanish, as separate copies
Written from the list your general contractor or platform gives you, filled in with your company. What passes is still the reviewer's decision. See what's in the Bid and prequal packet.
More on prequalification
- Written programs for ISNetworld, Avetta, Veriforce and GC prequalification
- ISNetworld RAVS written programs: what the review is and how to prepare
- Avetta safety manual audit: what it reviews and why documents come back
- Highwire safety assessment: the documents, the score and how to clear a discrepancy
- Subcontractor prequalification packet: what general contractors ask for
- OSHA 300A for prequalification: three years, exempt companies and numbers that match
- EMR for prequalification: what the experience mod means and where the letter comes from
Every rate printed with its formula and its counts
In 29Safety your 300, 300A and 301 data build themselves from the incidents you record, with TRIR and DART. The prequal packet prints every rate with its formula and the counts behind it, the way OSHA's forms package figures it, for the three calendar years most questionnaires ask for, so a reviewer can check the math any way a form asks.
See exactly what you'd get
The JHA plus a toolbox talk record, an inspection and more, in one PDF: five documents for a jobsite crew, or three for a plant. No watermark.
Plans from $199 a month. No per-person fees. 14 days free.
Rather have us write it? Two written programs ($295)
Need the whole packet? Bid and prequal packet ($795)
Questions we get
How do I calculate TRIR?
Add columns G, H, I and J on your OSHA 300A, multiply by 200,000, and divide by the total hours worked by all employees on the same 300A. For example, 3 cases in 150,000 hours: 3 × 200,000 ÷ 150,000 = 4.0.
What is the difference between TRIR and DART?
TRIR counts every recordable case. DART counts only the cases with days away from work or job transfer or restriction, columns H and I. Same formula, same hours.
What hours do we use?
The total hours worked by all employees covered by your log, from payroll, without vacation, sick leave, holidays or other non-work time, including temporary workers you supervise day to day. That is the number on your 300A.
What is a good TRIR?
Lower than your industry, by the client's measure. BLS put 2024 construction at 2.2 and specialty trade contractors at 2.3, and Highwire scores your three-year average against the BLS average for your trade, with full points below 60 percent of it. A company with few hours should expect one case to swing the rate, and be ready to explain it.
Does 29Safety calculate these for us?
Yes, from the incidents you record: 29Safety builds your 300, 300A and 301 data with TRIR and DART, and the prequal packet prints every rate with its formula and the counts behind it, so a reviewer can check it any way a form asks.
Not sure how this applies to your work? in the chat, or email safetyteam@29safety.com.
Sources
Checked November 5, 2026. Platforms, owners and general contractors set their own requirements and change them; the list in your account or on your form is the one that counts.
- OSHA recordkeeping forms package (300, 300A, 301 and instructions), OSHA
- How To Compute Nonfatal Incidence Rates, U.S. Bureau of Labor Statistics
- Table 1. Incidence rates of nonfatal occupational injuries and illnesses by industry and case types, 2024, U.S. Bureau of Labor Statistics
- Subcontractor health and safety questionnaire, Howard University
- Sample prequalification application questionnaire (May 2025), McGough
- Subcontractor prequalification form (2025), Torcon
- How is a Safety Score calculated?, Highwire help center
- 29 CFR sections linked in the text: eCFR, the official text