A plant manager sees TRIR at 2.1 and DART at 1.8 for the same quarter and asks which number actually matters to the corporate safety team. The honest answer is both, but they're not measuring the same thing. Your DART rate strips out every recordable case where someone just got treatment and kept working their normal job, and keeps only the cases severe enough to cost days away, restricted duty, or a job transfer. That makes DART the number that tracks how badly people actually got hurt, not just how many times a form got filled out. Here's the formula, a full worked example, and the restricted-duty judgement calls that cause most of the arguments.
What does DART rate actually measure?
DART stands for Days Away, Restricted, or Transferred. It's a rate, not a raw count, so it can be compared across sites and years even when headcount changes. The formula is:
DART = (Number of DART cases × 200,000) ÷ Total hours worked by all employees
The 200,000 is a normalising constant — it represents 100 employees working 40 hours a week for 50 weeks a year. Without it, a 40-person site and a 4,000-person site couldn't be compared on the same scale. A DART case is any recordable injury or illness that involved at least one of three things: days away from work, restricted work activity, or job transfer to a different role. Deaths count too, though they're rare enough that most sites never see one in their DART numerator.
What DART deliberately leaves out is the "other recordable case" category — injuries that needed medical treatment beyond first aid (think stitches, prescription-strength medication, or an X-ray that finds a fracture) but where the employee never lost a day and never had their job duties changed. Those cases are still recordable under 1904.7 and still belong on the OSHA 300 Log. They just don't belong in DART's numerator, because DART exists to isolate severity, not to count every recordable event.
How is DART different from TRIR?
TRIR, or Total Recordable Incident Rate, uses the same formula and the same 200,000-hour base, but the numerator is every recordable case, full stop. That means TRIR always includes everything DART includes, plus the milder medical-treatment-only cases DART excludes. The result: TRIR is always equal to or higher than DART for the same period, and if the two numbers are identical, either the site had zero medical-treatment-only cases, or something in the counting is worth a second look.
Here's the practical difference in one table.
| Case type | Counts toward TRIR | Counts toward DART |
|---|---|---|
| Death | Yes | Yes |
| Days away from work (1+ calendar day) | Yes | Yes |
| Restricted work or job transfer, any part of a day | Yes | Yes |
| Medical treatment beyond first aid, no days away, no restriction | Yes | No |
| First aid only (per the first aid list) | No | No |
TRIR counts every recordable case; DART counts only the subset severe enough to involve days away, restricted duty, or transfer.
If your team tracks TRIR alone, a site can look flat or improving while its injuries are quietly getting more severe — more days away, more people on light duty — because a drop in minor stitches-and-back-to-work cases can offset a rise in serious ones. DART catches that shift because it's built to ignore the noise from minor cases. We cover the TRIR side of the formula in full in our TRIR calculation guide, and if you want the underlying recordability logic that feeds both rates, start with what makes an injury OSHA recordable.
A worked DART calculation
Take a 220-employee distribution centre, one calendar year, 431,200 total hours worked across all employees. Over the year, the site logs eight recordable cases:
| Case | Description | Recordable? | Counts toward DART? |
|---|---|---|---|
| 1 | Forklift operator, crushed toe, 6 days away | Yes | Yes |
| 2 | Warehouse associate, back strain, moved to light duty for 4 days | Yes | Yes |
| 3 | Picker, laceration, 3 stitches, back on shift same day | Yes | No — medical treatment only |
| 4 | Dock worker, shoulder strain, restricted from lifting for 1 partial day | Yes | Yes |
| 5 | Associate, chemical splash, prescribed eye ointment, no days off | Yes | No — medical treatment only |
| 6 | Supervisor, fractured wrist, 22 days away | Yes | Yes |
| 7 | Associate, minor burn, aloe vera and a bandage from the first aid kit | No | No — first aid |
| 8 | Forklift operator, knee injury, transferred to a desk role for 9 days | Yes | Yes |
Five of the eight cases are DART cases (1, 2, 4, 6, 8). Case 7 never makes the 300 Log at all, because first aid isn't recordable. Cases 3 and 5 are recordable and belong on the log, but they don't touch DART's numerator.
DART = (5 × 200,000) ÷ 431,200 = 1,000,000 ÷ 431,200 = 2.32
TRIR for the same site, using all seven recordable cases, comes out at (7 × 200,000) ÷ 431,200 = 3.25. Same site, same year, two different stories — TRIR says seven incidents happened; DART says five of them were serious enough to change what someone could do at work.
What counts as "restricted duty" in the DART numerator?
This is where most of the disagreement lives, and it's worth slowing down on. Under 1904.7(b)(4), a case counts as restricted work if the employee's normal job functions were curtailed for any part of a day because of the injury. That includes being moved to a different role, being given lighter tasks, or simply being told not to complete part of their usual duties — even for a single partial shift, and even with no doctor's order behind it.
The phrase that trips people up is "any part of a day." A supervisor doesn't need to sign paperwork, and a physician doesn't need to be involved. If a warehouse employee strains their back lifting a pallet, sees the on-site nurse, and is told to skip the loading dock and stay on light picking for the rest of the shift, that's a restricted-duty case the moment the tasks change — not a maybe. It goes on the 300 Log under the restricted/job transfer column, and it counts in the DART numerator for that period, even though the employee never left the building and never missed a single day of pay.
One EHS manager we spoke with put it this way: "The mistake I see most is treating restricted duty like it needs a doctor's note to count. It doesn't. If I tell someone 'don't do the heavy stuff today,' I've just created a recordable restriction, whether I meant to or not." That's not a loophole in the rule — it's the rule working as written. The bar for restricted work is deliberately low because OSHA wants employers counting the real scope of an injury's impact, not just the days someone physically stayed home.
For the full walkthrough of a light-duty case start to finish, see our worked restricted-duty scenario. And if you're not sure whether a specific case even clears the recordability bar before you get to counting DART, run it through the free recordability check first.
Where teams get DART rate wrong
A handful of mistakes show up again and again when we look at how sites actually track DART. Watch for these.
- Not counting a light-duty day because "they were still at work." Being present doesn't matter. What matters is whether normal duties were curtailed. An employee at their desk all day on modified tasks is still a DART case.
- Missing partial-day restrictions. A restriction that lasts two hours on one shift still counts as a full restricted day under 1904.7(b)(4). Teams sometimes wait for a restriction to last a full day or more before logging it — that's not how the rule reads.
- Confusing DART with TRIR and reporting the wrong one. Corporate scorecards sometimes ask for TRIR and get handed DART, or vice versa, because the two numbers look similar and use the same formula shape. Label every rate you report with which cases are in the numerator.
- Forgetting job transfer counts even without any lost capability. Moving someone to a different position because of their injury is a DART trigger on its own, even if the new job has identical pay and hours. The transfer itself is the trigger, not a demotion or a pay cut.
- Letting the days-away count run past 180 or start on the wrong day. Days away counting starts the calendar day after the injury, not the day it happened, and caps at 180 days for DART purposes exactly as it does for the 300 Log. Miscounting the start date shifts cases between reporting periods.
How does DART rate compare to industry benchmarks?
OSHA and the Bureau of Labor Statistics publish industry injury and illness data that many sites use as a rough yardstick, and BLS breaks its published rates out by NAICS industry code, so a warehouse operation and a manufacturing plant aren't judged against the same baseline. We won't put a specific number on "good" here — benchmarks shift year to year and vary a lot by industry, and quoting one without checking the current release does more harm than good. The BLS Injuries, Illnesses, and Fatalities program publishes the current data if you want a real comparison point for your NAICS code. What matters more day to day is your own trend line: is DART falling relative to your own prior years, and is the gap between your DART and TRIR narrowing or widening.
Get your DART rate right, then track it
DART rate is only as accurate as the recordability calls and the restricted-duty judgement calls feeding it. Get a case wrong at the source — miss a restriction, misjudge a first-aid call — and both DART and TRIR are wrong downstream. Run your cases through the TRIR/DART rate calculator to get both numbers from the same hours and case data, side by side, so you can see exactly where the gap between them comes from. For teams that need the underlying 300 Log, 300A summary, and ITA export kept current as cases close out, see how Job13 handles the full recordkeeping cycle.
This article is for general information about federal 29 CFR Part 1904 and is not legal advice.
Frequently asked questions
Is a lower DART rate always better than a lower TRIR?
Not necessarily in isolation, but DART is the better severity signal. A site can lower TRIR by reducing minor medical-treatment cases while DART stays flat or climbs, which usually means the injuries that are happening are getting more serious even as the total count drops.
Does DART rate include first aid cases?
No. First aid cases aren't recordable at all under 1904.7, so they never reach the 300 Log and never touch DART or TRIR. Check the first aid list if you're unsure whether a specific treatment counts as first aid or crosses into medical treatment.
Do partial-day restrictions count toward DART even if the employee wasn't told by a doctor?
Yes. Under 1904.7(b)(4), a restriction counts the moment normal job functions are curtailed for any part of a day, regardless of who made that call. A supervisor telling someone to skip a task is enough to trigger it.
What's a good DART rate for my industry?
There's no single universal target — BLS publishes industry-specific rates by NAICS code that make a far better comparison point than a generic number. Track your own DART trend year over year alongside the current BLS data for your industry rather than chasing a fixed benchmark.
Can a case count toward DART without counting toward TRIR?
No. Every DART case is also a TRIR case, because DART cases are a subset of all recordable cases. TRIR can never be lower than DART for the same period; if you see the reverse, recheck the case counts.
