TRIR & LTIR Calculator: Formulas, Benchmarks, Improvement

Total Recordable Incident Rate (TRIR) is a widely reported safety metric in EHS. Boards ask about it, clients require it on prequalification questionnaires, and insurers take it into account. It is also routinely calculated wrong, and compared against numbers that do not mean what people think they mean.

This guide gives you the formulas, the actual published national rates, and the levers worth pulling — with the evidence base for each stated plainly.

Try our free TRIR and LTIFR calculator to compute your rate, then compare it with the published rate for your industry.

What Is TRIR?

Total Recordable Incident Rate (TRIR) is a normalised measure of work-related injuries and illnesses per 100 full-time equivalent workers per year. It lets you compare safety performance between organisations of different sizes.

The formula:

TRIR = (Number of OSHA-recordable incidents × 200,000) ÷ Total hours worked

The 200,000 represents 100 employees working 40 hours a week for 50 weeks (BLS, how to compute your firm's incidence rate). It is the normalisation constant that makes companies of different sizes comparable.

Worked Example

For example, suppose your company logged 12 OSHA-recordable incidents over the past year. Your 250 employees worked a combined 520,000 hours.

TRIR = (12 × 200,000) ÷ 520,000 = 4.62

That number alone tells you little. Whether it is high or low depends on your industry — and, as the section below explains, on what you are comparing it against.

Two further points about this example. First, with 520,000 hours each recordable case adds 200,000 ÷ 520,000, or about 0.38, to the rate, so 11 cases instead of 12 would give 4.23. A single classification decision can therefore move a small employer's rate by more than the year-on-year change a board is told about. Second, if the same company is in health care and social assistance (published BLS rate 3.4 for 2024, see below), the comparison is 4.62 against 3.4, which is about 1.4 times the published average; with 12 cases in one year that is a prompt to look at the case mix and the hours count, not a verdict on the programme.

What Counts as "Recordable"?

This is where calculations often go wrong. Not every incident is OSHA-recordable. Under the general recording criteria in 29 CFR 1904.7 (part of OSHA's recordkeeping rule, Part 1904), an incident is recordable if it results in any of:

• Death
• Days away from work
• Restricted work or transfer to another job
• Medical treatment beyond first aid
• Loss of consciousness
• A significant injury or illness diagnosed by a licensed healthcare professional

First aid does not count (see the glossary entries on first aid cases and recordable incidents). OSHA defines first aid treatment as a specific list in 1904.7(b)(5)(ii), and the distinction between first aid and medical treatment beyond first aid is what decides recordability — not how serious the event felt at the time. Check the definition rather than relying on judgement.

Recordability decisions: first aid or medical treatment?

The list in 1904.7(b)(5)(ii) is complete: OSHA says no other treatment counts as first aid. The pairs below, taken from that list and its notes, show how a small change in treatment changes the answer. They concern the treatment only; a case must also be work-related and a new case to be recorded, and you should read the full text before classifying a real case.

The rule also does not depend on who gives the treatment: a first-aid-list treatment given by a physician is still first aid.

What Is LTIR — and how it differs from DART

Lost Time Incident Rate (LTIR) counts only cases where the worker missed at least one full day of scheduled work after the day of the incident:

LTIR = (Lost-time incidents × 200,000) ÷ Total hours worked

LTIR and DART are not the same metric, and an earlier version of this guide wrongly said they were. Note also that "lost time incident" is not a term in OSHA's recordkeeping rule, which counts cases with days away from work. Under 29 CFR 1904.7(b)(3), days away are calendar days counted from the day after the injury, including weekends and days off the worker would not have worked, and the count may be capped at 180. Many companies and clients define lost time as scheduled working days instead, so when a questionnaire asks for an LTIR, ask which definition it uses and say which one you used. DART stands for Days Away, Restricted or Transferred, and it counts restricted-duty and job-transfer cases as well as days-away cases. DART is therefore always equal to or higher than a days-away-only rate. BLS publishes a DART rate; if you are comparing against a published figure, check which one it is before drawing a conclusion.

The relationship between your TRIR and your lost-time rate is worth watching, but be careful about the story you read into it. A high proportion of lost-time cases can indicate severity, weak early intervention or a limited return-to-work programme — or simply a small number of cases in a small workforce, where one event moves the rate substantially.

Published incidence rates by industry (BLS, 2024)

The US Bureau of Labor Statistics publishes incidence rates annually through its Survey of Occupational Injuries and Illnesses. These are the total recordable case rates for 2024, per 100 full-time equivalent workers, as published:

Source: BLS, Incidence rates of nonfatal occupational injuries and illnesses by industry, 2024. The all-private-industry rate of 2.3 was down from 2.4 in 2023 and is the lowest in this series going back to 2003 (BLS news release). Oil and gas extraction is not shown because the published cell was suppressed.

A correction, and why it matters. An earlier version of this guide published a different table with a second column headed "Top-Quartile TRIR". BLS does not publish quartiles in this series, and those figures had no source. Nine of the ten rates in that table were also wrong — healthcare was given as 5.5 against a published 3.4. If you benchmarked against that table, re-check your comparison against the BLS figures above.

Two cautions on using any of these numbers:

• These are national averages for broad industry groups, not targets and not performance thresholds. Use the most specific NAICS code you can find rather than a sector headline.
• An average is not a quartile. Being below the sector average does not make you a top performer, and we do not publish a top-quartile figure because we have not found a published one to cite.

Levers worth pulling

These are commonly pursued interventions. We have not measured the effect of any of them and are not aware of controlled studies isolating them from the wider programme, so treat each as a hypothesis to test against your own baseline rather than a proven reduction.

1. Leading indicators

Lagging indicators tell you what already broke; leading indicators — observations, near misses, audit findings, overdue actions — tell you where to look next. See our leading indicators guide for choosing them and the QHSE KPIs and metrics page for the wider set.

2. Near-miss reporting

The logic is straightforward: an event you can investigate before anyone is hurt is the cheapest learning opportunity you have. Be careful with ratio targets. The often-quoted accident-triangle ratios come from H.W. Heinrich's 1931 work, and a study of MSHA data from more than 25,000 US mining establishments (Yorio and Moore, 2017) found that establishments with more lower-severity incidents were more likely to have a fatality, but that the shape of the triangle depended on how severity was classified; it did not establish any fixed ratio. Set a target from your own measured baseline instead. See our near-miss reporting guide and the near-miss reporting software page.

3. Investigation quality

A weak investigation produces a weak corrective action. The measurable thing here is not the method you train but whether a second person reviews closure evidence and can reject it — track that, and corrective action close-out, as KPIs in their own right.

4. Observation programmes

Structured peer observation aims to surface at-risk conditions before they cause injury. Whether it works depends heavily on execution: volume, blameless feedback and whether anything changes as a result. See our behaviour-based safety guide.

5. High-hazard task control

Permit systems target high-hazard tasks such as confined space entry, hot work, work at height, energy isolation and lifting. A permit system with real pre-task checks is a control on those specific tasks. See our permit-to-work guide.

6. Contractor management

Where contractors do a large share of the hazardous work, their incidents belong in your picture. Note the calculation trap: if you include contractor incidents you must include contractor hours, or your rate is overstated. See our contractor safety guide.

7. Frontline ownership

Whether supervisors own safety in their crew, or treat it as the EHS team's job, is hard to measure and slow to change, and it is not a software feature.

Common calculation mistakes

1. Including first-aid cases — inflates the rate. Check the OSHA definition rather than judging severity.
2. Excluding contractor hours while including contractor incidents — inflates the rate, because the incidents are counted without the hours that produced them.
3. Using calendar hours instead of actual hours worked — overtime, sick leave and holiday all skew the denominator.
4. Annualising partial-year data by multiplying — report TRIR over a full 12 months (the calendar year for comparison with OSHA and BLS figures, or a rolling 12 months), not a quarter multiplied by four.
5. Comparing against the wrong population — a broad sector average is not your benchmark if a specific NAICS rate exists.
6. Comparing a days-away rate against a published DART rate — different metrics, and the DART figure will be higher.

How software helps

Calculating this in a spreadsheet is slow and easy to get wrong. Some EHS platforms document support for the following, and each product's record on this site says what it does and does not establish:

• Recordability classification against the OSHA criteria, with the decision recorded
• Hours worked pulled from HRIS or timekeeping rather than estimated
• Rolling 12-month rates by site, region or business unit
• Comparison against your own historical baseline

The defensible benefit is fewer classification errors and less manual effort, both of which you can measure in your own organisation. We do not publish a payback figure or an employee threshold above which software is justified, because we have not measured one.

Checklist before you send a TRIR to a client or board

1. Which OSHA 300 logs and which year? A calendar-year rate matches the BLS series; a rolling 12-month rate does not, so label it.
2. Do the hours come from payroll or timekeeping, and do they include contractors if their cases are included?
3. Has someone other than the person who classified the case checked each recordable against 1904.7 and the first aid list?
4. Is the comparison industry the most specific NAICS code available, and is the BLS table year stated?
5. Is the definition of any lost-time figure written next to it?
6. Is the number accompanied by a count of cases, since a rate on a few cases moves sharply?

For the electronic submission side of recordkeeping, see the OSHA electronic recordkeeping article. The glossary defines TRIR and LTIFR, and the leading indicators guide covers the measures that move before the rate does.

FAQs

Is TRIR per 100 employees or per 200,000 hours?

Both describe the same calculation. The 200,000 multiplier represents 100 full-time employees working a year, so the rate expresses recordable cases per 100 full-time equivalent workers.

What is a "good" TRIR?

There is no universal answer, and we do not publish a threshold. Compare against the published rate for the most specific industry classification you can find, and against your own trend. Be wary of any single number presented as a pass mark — including "sub-1.5", which an earlier version of this guide gave without a source.

Should I use TRIR or a lost-time rate as my KPI?

Tracking both is common. A rate that counts only lost-time cases leaves out restricted-duty and medical-treatment cases, and TRIR alone says nothing about severity. Whichever you report, be explicit about which lost-time metric you mean — days-away only, or DART — and keep it consistent year to year.

How often should I calculate it?

Monthly, as a rolling 12-month rate. Single-month rates can be too volatile to interpret, especially for a small workforce: because the rate is normalised to 200,000 hours, one additional case can move a small employer's rate substantially.

Is TRIR the same as the OSHA incidence rate?

They refer to the same measure. "OSHA incidence rate" or "total recordable case rate" is the formal terminology used in the BLS series; TRIR is the common shorthand.

---

Want to compute your rate? Use our free TRIR and LTIFR calculator, then compare it against the BLS figure for your industry above.

Sources

• US Bureau of Labor Statistics, Table 1: incidence rates of nonfatal occupational injuries and illnesses by industry and case types, 2024
• US Bureau of Labor Statistics, Employer-reported workplace injuries and illnesses, 2023-2024 (news release, 22 January 2026)
• US Bureau of Labor Statistics, How to compute your firm's incidence rate
• eCFR, 29 CFR 1904.7, general recording criteria; OSHA, Recordkeeping
• Yorio PL, Moore SM, Examining factors that influence the existence of Heinrich's safety triangle using site-specific H&S data from more than 25,000 establishments, Risk Analysis, 2017

Rates in this guide were checked against BLS published tables on 22 September and again on 4 October 2026; the 2024 figures are still the latest BLS release. Figures in the levers section are not measured effects and are labelled where the evidence is absent.

Revised 4 October 2026: the BLS 2024 rates, the all-private-industry trend, 29 CFR 1904.7 criteria and the first aid list were re-checked against the sources and stand. The description of the mining study was corrected to match its publication (MSHA data, not a NIOSH study), the lost-time definition was clarified against 1904.7(b)(3), a generalisation about what platforms automate was reworded, and a recordability table, a sensitivity example and a checklist were added.

Corrected 23 September 2026: statistics without a primary source were sourced, labelled as examples or removed, and the regulations cited now link to their official texts.