Counting days away and restricted work days on the OSHA 300 Log (1904.7)

You have already done the hard part. A case is work-related, it meets a recording criterion, and you have entered it on the 300 Log. But if the injury sent the employee home for a while or put them on restricted duty, you are not finished — you have to count the days, and the day count is where careful recordkeepers and sloppy ones part ways. Columns K and L on the 300 Log are not decorative. Those numbers roll up into your annual summary, feed your DART rate, and are the first thing an inspector cross-checks. Miscount them and you have a quiet, recurring error that distorts your safety metrics and invites a closer look at the rest of your log.

The rules for counting live in 29 CFR 1904.7(b)(3) for days away from work and 1904.7(b)(4) for restricted work and job transfer. They are specific, they are frequently misunderstood, and most of the confusion comes from treating the day count like a payroll calculation. It is not. Here is how OSHA actually wants it done.

Two columns, two separate day counts

The 300 Log gives you two places to record days, and they capture different things:

• Column K — days away from work. The number of calendar days the employee was kept away from work entirely because of the injury or illness.
• Column L — days on job transfer or restriction. The number of calendar days the employee was at work but either restricted from one or more routine job functions, kept from working a full day, or transferred to another job.

A single case can land in both columns. If a worker breaks an arm in a fall, is sent home for 7 days, then returns on light duty for 10 days, you enter 7 in Column K and 10 in Column L. The same case produces two independent day counts, and you tally each one separately at the bottom of the log.

The classification check marks in columns G through J are different from the day counts. There you check only one box — the most serious outcome. For a case with both days away and restricted days, days away is the more serious outcome, so you check Column H (days away) and leave Column I (job transfer/restriction) blank, even though you are still filling in both Column K and Column L. We walk through the column-by-column mechanics in detail in our guide to filling out the OSHA 300 Log; this post is about the counting itself.

When the clock starts: the day after, never the day of

The most common counting error is including the day of the injury. Don't. Under 1904.7(b)(3), you begin counting on the day after the injury occurred or the illness began. The day of the incident itself is never counted as a day away or a restricted day, no matter what happened that day.

So if an employee is hurt on Tuesday and stays home Wednesday, Thursday, and Friday, returning the following Monday, that is not "Tuesday through Sunday." You start counting Wednesday and count every calendar day the employee remained unable to work. The injury day is excluded because it is captured by the date-of-injury field (Column D), not the day count.

Count calendar days, not workdays

This is the rule that trips up employers who think in terms of scheduled shifts. You count calendar days the employee was unable to work — not the number of shifts missed. Weekends, holidays, vacation days, scheduled days off, plant-shutdown days: if the employee would have been unable to work because of the injury, those days all count.

An employee on a Monday-through-Friday schedule who is injured and out for two full weeks has missed 14 calendar days, not 10 workdays. The Saturdays and Sundays in that stretch count because the employee was still unable to work on those days. OSHA reaffirmed this calendar-day reading in a 2019 standard interpretation: the count is calendar days the employee could not work, regardless of whether the employee was scheduled to work on a given day.

The same counting method applies to restricted days under Column L. 1904.7(b)(4) says you count days of job transfer or restriction the same way you count days away — calendar days, beginning the day after the injury.

The 180-day cap (and it's a combined cap)

You do not have to count days forever. Under 1904.7(b)(3)(viii), you may cap the count at 180 calendar days. Once a case has resulted in more than 180 calendar days away from work and/or job transfer or restriction, you are not required to keep tracking the exact number — entering 180 is considered adequate.

The detail employers miss is that the cap is combined, not 180 per column. The regulation caps days away from work and/or days of job transfer or restriction at 180 together. You do not get 180 days away plus another 180 restricted days before you can stop counting; once the combined total reaches 180, you can cap.

In practice, for a case that is purely days-away, you stop at 180 in Column K. For a case that is purely restricted, you stop at 180 in Column L. For a case that mixes both, OSHA's interpretation (FAQ 7-18) is that when an employee on restricted work hits the 180-day cap and is then moved to days away, you reclassify the case to the more serious outcome — remove the restriction check, check days away — and you may enter 180 in the restricted column and 1 day in the days-away column. You are never required to keep tracking beyond the combined 180.

Cases that cross into the next calendar year

Injuries don't respect the December 31 boundary, but your 300 Log does. If a case occurs in one year and the days away extend into the next, you do not record it on both years' logs. Under 1904.7(b)(3), you record the case only once — on the log for the year the injury or illness occurred — and enter all the day counts there, even the days that fall in the following year.

That raises an obvious problem: when you complete your 300A annual summary in January or February, the employee may still be out. The rule is to estimate. Estimate the total number of calendar days you expect the employee to be away (or restricted), use that estimate for your annual summary totals, and then go back and update the log entry later when the actual count is known — or when it reaches the 180-day cap, whichever comes first. The log is a living document for five years; updating a day count as a case resolves is exactly the kind of maintenance the retention-and-update rule under 1904.33 requires.

When the employee leaves before the case resolves

What if the employee separates from the company while still recovering? It depends on why they left. Under 1904.7(b)(3)(vii):

• If the employee leaves for a reason unrelated to the injury or illness — they retire, the plant closes, they take another job — you may stop counting days as of the separation date.
• If the employee leaves because of the injury or illness, you must estimate the total number of days away (or restricted) the case would have produced and enter that estimate on the log.

The distinction matters because it prevents two opposite abuses: stopping the clock can't be used to understate a serious injury that drove someone out of the workforce, and you are not forced to count phantom days for a worker who left for reasons that had nothing to do with their case.

When the physician's note and reality disagree

Sometimes a doctor recommends time off and the employee works anyway, or is cleared but the employer keeps them out. The recordkeeping answer is driven by the medical recommendation, not by attendance. Under 1904.7(b)(3)(vi), if a physician or other licensed health care professional recommends that the employee stay away from work — or perform only restricted work — you record the days based on that recommendation, whether or not the employee actually follows it.

So if a doctor says "stay home for five days" and your employee shows up on day three because they feel fine, you still record five days away. The recommendation is the trigger. (OSHA also encourages employers to follow the medical recommendation — having an employee work against medical advice is its own liability problem, separate from recordkeeping.)

Why the day counts matter beyond the log

The numbers in Columns K and L are not trapped on the 300 Log. They flow directly into your DART rate — the Days Away, Restricted, or Transferred rate that contractors use as a prequalification filter and that OSHA uses to help target inspections. A case with 40 days away counts the same toward your DART case rate as a case with 4 days; DART counts cases, not days, so a single miscounted case won't move that rate. But the raw day counts still drive how serious your log looks to an inspector, and they are the easiest line item to cross-check against your 300A summary and your 301 incident reports.

Getting the count wrong in either direction has consequences. Undercount — by including the injury day, counting workdays instead of calendar days, or dropping a cross-year case — and you understate the severity of your record, which OSHA treats as a recordkeeping violation that can run up to $16,550 per case. Overcount and you inflate your own severity metrics, scaring off insurers and bid sponsors with numbers worse than reality. The count is supposed to be a faithful measurement, not a guess.

Let the software do the counting

Day counting is exactly the kind of rule-bound, easy-to-fumble task that should not depend on someone remembering whether weekends count. LogStead starts the clock on the day after the incident, counts calendar days automatically, applies the combined 180-day cap, keeps days-away and restricted-day totals in their own columns, and rolls them straight into your 300A summary and DART rate — so the totals reconcile by construction instead of by January scramble. When a case crosses into a new year or an employee is still out at summary time, it flags the entry to revisit rather than letting it go stale. If you are still deciding whether a borderline case is even recordable before you get to the counting, our free recordability checker walks the 1904.7 criteria with you first.

Counting days correctly is unglamorous, but it is where a clean log is either earned or quietly lost. Start on the right day, count calendar days, cap at 180 combined, record cross-year cases once, and update estimates as cases close — and Columns K and L will hold up to the only audit that matters.