Know your DART rate.
Days Away, Restricted, or Transferred. The serious subset of your recordables, and the rate owners and insurers watch most closely.
A DART case is any row of your OSHA 300 log checked in column H (days away from work) or column I (job transfer or restriction). DART rate = (DART cases × 200,000) ÷ total hours worked. Because it drops medical-treatment-only cases, DART reads severity rather than raw frequency — US construction averaged 1.3 in 2024 against a 2.2 total-recordable rate (BLS SOII).
Your numbers
DART
Above the 1.3 construction average (BLS, 2024) — worth a closer look
Worked example: from a 300 log to a DART rate
A mid-size general contractor self-performing concrete logged 175,000 hours across its crews last year, pulled from certified payroll. The OSHA 300 log closed the year with five recordable rows: a laceration closed with sutures (medical treatment only, column J), a shoulder strain with 14 days away (column H), a hand fracture with 30 days away (column H), a back strain worked on restricted duty for 21 days (column I), and a foreign-body eye injury flushed at the clinic (column J).
- Count the DART cases. Two column-H rows plus one column-I row = 3. The two medical-treatment-only rows are recordable but not DART.
- Multiply by the 200,000-hour base. 3 × 200,000 = 600,000.
- Divide by hours worked. 600,000 ÷ 175,000 = 3.43.
- Put it next to TRIR. All five recordables give a TRIR of 5 × 200,000 ÷ 175,000 = 5.71. So 3 of 5 recordables, a 0.60 ratio, were serious enough to cost work time.
Both numbers are far above the 2024 construction means (DART 1.3, total recordable 2.2), and the ratio says the problem is severity, not paperwork volume: when this contractor gets hurt, people miss work. That points the fix at high-energy tasks like the concrete lifts and material handling behind the strains and the fracture, not at the clinic cases.
The number inside the number: DART ÷ TRIR
Because DART and TRIR share the same denominator and the same 200,000 base, dividing one by the other cancels everything except case mix. The quotient is the share of your recordables that were serious enough to take someone off full duty. US construction ran about 0.59 in 2024 (1.3 ÷ 2.2, BLS SOII), meaning roughly six of every ten recordables cost days or restrictions.
Watch the ratio quarter over quarter. A flat TRIR with a rising ratio means the same number of injuries is doing more damage — a severity drift that a TRIR-only scoreboard hides completely. The reverse, a rising TRIR with a falling ratio, often just means reporting culture improved and minor cases are finally making it onto the log, which is a good sign wearing a bad costume.
US construction benchmarks (BLS SOII, 2024)
Bureau of Labor Statistics, Survey of Occupational Injuries and Illnesses, 2024 survey year, construction sector (NAICS 23). Rates are per 100 full-time-equivalent workers. These are national means, not percentiles or per-trade targets.
| Measure | Rate (2024) |
|---|---|
| Construction DART (days away + restricted/transfer) | 1.3 |
| of which: days away from work | 0.9 |
| of which: job transfer or restriction | 0.4 |
| Construction total recordable cases (TRIR) | 2.2 |
| Implied construction DART-to-TRIR ratio | ≈0.59 |
| All private industry, total recordable cases | 2.3 (2.4 in 2023) |
Source: BLS Survey of Occupational Injuries & Illnesses, 2024survey year (bls.gov/iif). One construction detail worth noting: days-away cases (0.9) outnumber restricted-duty cases (0.4) roughly two to one, the opposite of many factory settings — on a jobsite there is often no light-duty bench to transfer someone to, so a serious injury becomes absence.
An annual DART is a postmortem. Make it a pulse.
Most contractors compute this rate once a year, when the 300A summary gets posted in February. POD recomputes it from the field reports your crews already file, as hours and cases land, so a severity drift shows up in the month it starts instead of in next year's filing.
Built from two columns
Your DART count is column H plus column I of the 300 log. The day totals in columns K and L grade how bad each case was, but the rate itself only counts cases.
Who reads it
Prequalification questionnaires and workers-comp underwriters usually ask for DART alongside TRIR and EMR, because a lost workday is where injury cost actually concentrates.
The 2024 yardstick
US construction posted a 1.3 DART in 2024 (BLS SOII, NAICS 23) — 0.9 days-away plus 0.4 restriction/transfer — against a 2.2 total-recordable rate.
Questions & answers
Each row of the 300 log classifies its case in exactly one column: G (death), H (days away from work), I (job transfer or restriction), or J (other recordable). Your DART count for the year is the number of rows checked in column H or column I. Columns K and L, which record how many days, measure severity but never change the case count.
Method & sources
DART rate equals the count of cases with Days Away, Restricted duty, or Transfer (columns H and I of the OSHA 300 log), multiplied by 200,000, divided by total hours worked — the OSHA recordkeeping convention. DART is a subset of your recordable cases, so it is always at or below your TRIR. The 1.3 construction reference is the BLS SOII national mean for the 2024 survey year (NAICS 23: 0.9 days-away + 0.4 restriction/transfer), a directional figure, not a per-trade target — BLS publishes an industry mean, not a contractor distribution.
Sources: BLS Survey of Occupational Injuries & Illnesses · OSHA 29 CFR 1904.7 (recording criteria)
Methodology reviewed by the Plan of Day construction team · July 2026
Your DART cases and hours, on the OSHA formula, nothing added.
Watch severity, not just the count.
POD recomputes your DART rate the moment hours and cases change and keeps it beside the near-misses and observations that predict where the next case comes from.
Last updated: September 2026