Recording Hearing Loss on the OSHA 300 Log: The 25 dB STS Rule, Explained

Most cases land on your OSHA 300 Log the same way: a worker gets hurt, the treatment goes beyond first aid or they miss work, and you record it. Hearing loss does not work like that. It has its own recording rule with its own numeric thresholds, its own retest window, its own baseline logic, and its own column on the log. If you run a hearing conservation program — and you must if your workers are exposed to noise at or above an eight-hour time-weighted average of 85 decibels — this is the recordability question you are most likely to get wrong, because the answer comes out of arithmetic rather than a treatment decision.

The rule lives at 29 CFR 1904.10. Its heart is a two-part test built around a 25-decibel threshold, and the mechanics are precise in ways that matter: which baseline you compare against, which half of the test you may age-correct, what a retest can and cannot undo. Everything below is taken from the text of 1904.10 and the noise standard at 29 CFR 1910.95 (including its Appendix F age-correction tables), plus three OSHA letters of interpretation that settle the questions the rule text leaves open. Where a number appears, it is the number in the source.

Why Hearing Loss Has Its Own Rule

Occupational hearing loss is gradual. A worker does not lose their hearing in a single instantaneous event the way they break a finger. The damage accumulates over months and years of noise exposure, and the only way to detect it is to compare audiograms over time. The general recording criteria in 1904.7 — medical treatment beyond first aid, days away, restricted duty — do not map onto a condition you detect with a hearing test rather than a treatment decision, so OSHA wrote a dedicated rule with a measurable, audiometric trigger.

That trigger is not a single number. It is two numbers that both have to be true, in the same ear.

The Two-Part Test

Under 1904.10(a), you must record a hearing loss case on the OSHA 300 Log when an employee's audiogram reveals both of the following:

1. The employee has experienced a work-related Standard Threshold Shift (STS) in hearing in one or both ears, and
2. The employee's total hearing level is 25 decibels or more above audiometric zero, averaged at 2000, 3000, and 4000 Hz, in the same ear(s) as the STS.

Both conditions have to be met, in the same ear, before the case is recordable. An STS alone is not enough. A 25 dB hearing level alone is not enough. You need the shift and the level together, on the same side.

This is the single most important thing to understand about hearing loss recordkeeping, because the two conditions are measured against different reference points — and one of them allows an adjustment the other does not.

Where the Audiograms Come From: The Noise Standard's Testing Program

The recordkeeping rule assumes you already have audiograms to compare. They come from the hearing conservation program that 1910.95 requires once employee noise exposure reaches the action level. The recordkeeping mechanics only make sense against that program, so here is the part of it that feeds 1904.10, paragraph by paragraph:

• Action level. Under 1910.95(c)(1), a "continuing, effective hearing conservation program" is required whenever exposures equal or exceed an 8-hour time-weighted average of 85 dBA — computed without regard to any attenuation from hearing protectors.
• Baseline audiogram. Under (g)(5)(i), you must establish a valid baseline within 6 months of an employee's first exposure at or above the action level (within 1 year where a mobile test van is used, per (g)(5)(ii), with hearing protectors worn in the gap beyond six months). Under (g)(5)(iii), the baseline test must be preceded by at least 14 hours without exposure to workplace noise; hearing protectors may substitute for that quiet period. Under (g)(5)(iv), you must tell employees to avoid high non-occupational noise during those 14 hours.
• Annual audiogram. Under (g)(6), every employee exposed at or above the action level gets a new audiogram at least annually after the baseline.
• Evaluation. Under (g)(7)(i), each annual audiogram is compared to the baseline to determine whether it is valid and whether an STS has occurred — and the rule says plainly that this comparison "may be done by a technician." Under (g)(7)(ii), if the annual audiogram shows an STS, the employer may obtain a retest within 30 days and treat the retest as the annual audiogram. Under (g)(7)(iii), an audiologist, otolaryngologist, or physician reviews problem audiograms.
• Follow-up. Under (g)(8)(i), an employee whose comparison shows an STS must be informed in writing within 21 days of the determination. Under (g)(8)(ii), unless a physician determines the shift is not work-related or aggravated by occupational noise, the employer must fit or refit hearing protectors, retrain the employee, refer for clinical audiological or otological evaluation where appropriate, and inform the employee of the need for an otological exam if a pathology unrelated to hearing protectors is suspected.
• Revised baseline. Under (g)(9), an annual audiogram may be substituted for the baseline when, in the judgment of the reviewing audiologist, otolaryngologist, or physician, the STS is persistent, or the annual audiogram shows significant improvement over the baseline.
• Records. Under (m)(2), you must retain all audiometric test records obtained under paragraph (g) — name and job classification, test date, examiner, calibration date, and the employee's most recent noise exposure assessment.

Two things follow for recordkeeping. First, an STS triggers the 21-day notice and the follow-up steps under the noise standard whether or not the case ever becomes recordable — the noise standard and the recordkeeping rule are separate obligations with separate triggers. Second, the retained audiometric records are what let you defend, or later correct, a 300 Log entry, which is why the line-out provision discussed below works.

What a Standard Threshold Shift Actually Is

1904.10(b)(1) does not define the STS itself; it points to the noise standard. Under 1910.95(g)(10)(i), a standard threshold shift is "a change in hearing threshold relative to the baseline audiogram of an average of 10 dB or more at 2000, 3000, and 4000 Hz in either ear." The recordkeeping rule restates it as a 10 dB average change "in one or both ears."

The reference point is the employee's own baseline — the audiogram established when they entered the hearing conservation program, or the revised baseline discussed next. You are measuring change over time against that personal starting point, not against any fixed standard. Take the three thresholds at 2000, 3000, and 4000 Hz on the current audiogram, average them, do the same for the baseline, and compare. A difference of 10 dB or more is an STS.

Which Baseline? The Revised-Baseline Rule

This is the part of 1904.10 the short version of this post skipped, and it decides whether a worker with an old recorded case gets recorded again.

Under 1904.10(b)(2)(i):

• If the employee has never previously experienced a recordable hearing loss, you compare the current audiogram with the employee's baseline audiogram.
• If the employee has previously experienced a recordable hearing loss, you compare the current audiogram with the employee's revised baseline — which the rule defines as "the audiogram reflecting the employee's previous recordable hearing loss case."

The practical effect is that a worker is not re-recorded every year for the same loss. Once a case has been recorded, the audiogram that produced it becomes the new zero point for that worker's future STS determinations. A further 10 dB average shift from that audiogram is what it takes to produce a second recordable case.

Notice how this interlocks with 1910.95(g)(9). The noise standard lets the reviewing professional substitute an annual audiogram for the baseline when a shift is persistent. The recordkeeping rule ties its revised baseline specifically to a previous recordable case. In the common situation — a persistent, work-related STS that met the 25 dB level and was recorded — both rules point at the same audiogram. Have the reviewing audiologist or physician document which audiogram is serving as the baseline for each worker and why; that note is what makes next year's determination defensible.

The 25 dB-From-Zero Check

The second condition uses a different reference point entirely. Under 1904.10(b)(2)(ii), "audiometric test results reflect the employee's overall hearing ability in comparison to audiometric zero," so you use the current audiogram's average hearing level at 2000, 3000, and 4000 Hz to determine whether the employee's total hearing level is 25 dB or more.

The STS measures how much the employee's hearing has changed from their own baseline. The 25 dB check measures how much total hearing loss they have accumulated against the universal zero point. A worker can have a recordable-sized shift without yet reaching 25 dB total, and the case is not recordable until both are true. A worker can also sit well above 25 dB for years without a new recordable case, as long as no further STS occurs.

Age Adjustment: Allowed for One Half, Not the Other

Here is the rule that catches even experienced safety managers. Hearing naturally declines with age, and OSHA lets you account for that — but only for one half of the test.

Under 1904.10(b)(3), when you are determining whether an STS has occurred, you may age-adjust the employee's current audiogram using Tables F-1 (males) or F-2 (females) in Appendix F of 1910.95. When you are determining whether the employee's total hearing level is 25 dB or more above audiometric zero, you may not use an age adjustment. The 25 dB check is taken straight from the current audiogram, unadjusted.

Mixing these up is the most common technical error in hearing loss recordkeeping. Age adjustment can shrink a shift below the 10 dB STS threshold and legitimately keep a case off the log. It cannot be used to shrink the 25 dB total hearing level. Apply it to the first condition only.

How the Appendix F correction actually works

Appendix F is non-mandatory — its opening line says allowance "may be made" for aging — but if you choose to age-correct, you must follow its procedure. The procedure is per frequency, not on the average:

1. Find the employee's age at the most recent audiogram in Table F-1 or F-2 and write down the correction values at each test frequency.
2. Find the employee's age at the baseline audiogram and write down those correction values.
3. Subtract the baseline-age values from the current-age values. The differences are the portion of the change in hearing that may be attributed to aging.
4. Subtract each difference from the corresponding threshold on the current audiogram. Then compare the corrected current thresholds to the baseline.

OSHA's own example in Appendix F is a 32-year-old male whose baseline was taken at 27. At 4000 Hz his threshold went from 5 dB to 25 dB — a raw shift of 20 dB. Table F-1 gives a 4000 Hz correction of 10 at age 32 and 7 at age 27, a difference of 3 dB. Subtracting 3 from the current 25 yields 22; comparing 22 to the baseline 5 gives an age-corrected shift of 17 dB instead of 20. For an STS determination you do that at 2000, 3000, and 4000 Hz and average the three corrected shifts.

Two things the tables do not do. They stop at "60 or older," and they were built from a 1970s NIOSH dataset. OSHA's February 9, 2024 letter of interpretation on this point allows an employer to use age-correction tables derived from the newer NHANES data instead of Appendix F, under conditions: every value must come from one dataset (a hybrid that splices Appendix F values under 60 with NHANES values over 60 is expressly impermissible); the decision to age-correct and which table to use must be made by a certified audiologist, otolaryngologist, or other physician, taking into account the worker's exposure history; OSHA recommends one set of tables be selected and applied to all workers from their baseline onward; and the employer should keep a copy of the published tables with the audiometric records. The same letter notes that NIOSH's 1998 criteria document recommends not age-correcting individual audiograms at all — which the rules still permit you to do. The rule says "may," not "must." Whatever you decide, decide it once, write it down, and apply it the same way to every worker.

Five Worked Determinations

Real table values, real arithmetic. Each example averages the thresholds at 2000, 3000, and 4000 Hz; the age corrections come from Tables F-1 and F-2 as published.

1. A recordable case, start to finish

A male head-rig operator's baseline was taken at age 35; his current annual audiogram is at 47. Left ear thresholds at 2000/3000/4000 Hz:

Step one — STS, with age correction. Table F-1 at age 47 reads 8 / 14 / 19 at 2000/3000/4000 Hz; at age 35 it reads 5 / 8 / 11. The differences are 3 / 6 / 8. Subtract those from the current thresholds: 17 / 24 / 27, average 22.7 dB. Corrected shift: 22.7 − 8.3 = 14.3 dB. That is 10 dB or more, so an STS has occurred.

Step two — 25 dB level, no age correction. The unadjusted current average is 28.3 dB, which is 25 or more. Met.

Step three — retest. The employer retests on day 12 and the retest confirms the STS. Under 1904.10(b)(4) the case must be recorded within seven calendar days of the retest.

Step four — work-relatedness. No physician or other licensed health care professional has found a non-occupational cause; the worker's exposure is at the head rig. Work-related under 1904.5.

Both thresholds met, same ear, work-related, confirmed: check column M(5), Hearing loss. With no days away and no restriction, the case is classified in column J, "Other recordable cases." (This is the shape of the recordable hearing loss case in our demo dataset — a head rig operator at a fictional sawmill, left ear, confirmed by retest, refitted with dual hearing protection.)

2. An STS that is not recordable — yet

A 24-year-old female sorter; baseline at 21. Right ear:

Table F-2 at 24 reads 5 / 4 / 4; at 21 it reads 4 / 4 / 3; differences 1 / 0 / 1. Corrected current thresholds 9 / 15 / 19, average 14.3 dB; corrected shift 14.3 − 3.3 = 11.0 dB. An STS.

But the total hearing level — the unadjusted current average — is 15.0 dB, below 25. Not recordable. The noise-standard obligations still fire in full: written notice within 21 days, refit and retraining, referral if appropriate. And the STS matters for next year, because if her hearing keeps falling she will cross 25 dB, and at that point the question is whether a new 10 dB shift exists against whatever baseline is then in force. Document what happened this year.

3. Age correction defeats the STS

A 58-year-old male millwright; baseline at 45. Right ear:

On raw numbers this looks like an STS, and the total hearing level of 36.7 dB is comfortably over 25. But Table F-1 at 58 reads 12 / 22 / 31 and at 45 reads 7 / 13 / 18 — differences of 5 / 9 / 13. Corrected current thresholds: 20 / 26 / 37, average 27.7 dB. Corrected shift: 27.7 − 25.0 = 2.7 dB. No STS. Not recordable, despite the total hearing level.

Note what did the work: the employer's standing decision to age-correct. Had this employer's policy been not to age-correct (as NIOSH recommends), the 11.7 dB raw shift is an STS, the 25 dB level is met, and — if work-related and confirmed — the case is recordable. Both outcomes are lawful. What is not defensible is age-correcting for the workers where it helps and skipping it where it doesn't.

4. Right shift, wrong ear

A worker's right ear shows a corrected shift of 12 dB, but the right-ear total hearing level is 18 dB. The left ear sits at a total hearing level of 31 dB but shows a shift of only 3 dB. Each ear satisfies one half of the test; neither satisfies both. Under 1904.10(a) the 25 dB level must be "in the same ear(s) as the STS." Not recordable.

5. A second case for a previously recorded worker

A worker was recorded in 2023 when his left-ear average reached 30 dB. Under 1904.10(b)(2)(i), that 2023 audiogram is now his revised baseline. His 2026 audiogram averages 36 dB: a shift of 6 dB from the revised baseline. No new STS, no new case — even though he is far past 25 dB. Had the 2026 audiogram averaged 41 dB, the shift from the revised baseline would be 11 dB, the 25 dB level is already met, and (subject to retest and work-relatedness) a second case goes on the 2026 log.

Work-Relatedness for Hearing Loss

Under 1904.10(b)(5) there are no special work-relatedness rules for hearing loss. You use 1904.5: if an event or exposure in the work environment caused or contributed to the hearing loss, or significantly aggravated a pre-existing hearing loss, the case is work-related. Our decision tree for OSHA recordability walks through the general analysis, including the nine exceptions in 1904.5(b)(2).

The rule also gives you a specific off-ramp. Under 1904.10(b)(6), if a physician or other licensed health care professional determines, following the rules in 1904.5, that the hearing loss is not work-related or that occupational noise exposure did not significantly aggravate it, you do not have to consider the case work-related or record it. OSHA's August 29, 2007 letter gives the canonical example: a worker in a high-noise environment meets the recording criteria, but a physician finds an inner ear infection that is entirely responsible for the loss — not work-related under 1904.5(b)(2)(ii).

That same 2007 letter resolves a mismatch worth knowing. Under the noise standard, 1910.95(g)(8)(ii), only a physician can determine that an STS is not work-related for purposes of excusing the follow-up steps. Under the recordkeeping rule, 1904.10(b)(6), a physician or other licensed health care professional — an audiologist operating within the scope of their state license — can make the determination that keeps a case off the log. So an audiologist's finding can settle recordability while the noise standard's follow-up obligations continue unless a physician signs off. OSHA also reminds employers in that letter that the work-relatedness decision is ultimately the employer's responsibility, made case by case; the professional's opinion informs it.

The 30-Day Retest Window Most Employers Skip

This provision can save you from recording temporary or erroneous shifts, and many employers never use it.

Under 1904.10(b)(4), if you retest the employee's hearing within 30 days of the first test and the retest does not confirm the recordable STS, you are not required to record the case. Hearing thresholds fluctuate — a worker who had a noisy weekend, an ear infection, or a bad test day can show an apparent shift that a clean retest does not confirm. If the retest does confirm the recordable STS, you must record the case within seven calendar days of the retest.

OSHA's March 4, 2004 letter settles two timing questions the text leaves open. The 30 days run from the date of the first test, not from the date the results reach the employer — a point that matters when an outside audiometric provider reports weeks later. And a retest obtained more than 30 calendar days after the initial audiogram cannot be substituted for it. If you do not retest at all, the general clock in 1904.29(b)(3) applies: seven calendar days from receiving information that a recordable case has occurred.

There is also a back-end correction, and it is in the rule text itself. Under 1904.10(b)(4), if subsequent audiometric testing performed under the noise standard indicates that an STS is not persistent, you may erase or line out the recorded entry. The 2004 letter adds the record-keeping consequence: Part 1904 does not require you to document the removal, but 1910.95(m)(2) requires you to keep all audiometric test records, so the evidence for the line-out is preserved regardless.

Filling Out the 300 Log, the 301, and the 300A

Under 1904.10(b)(7), when you enter a recordable hearing loss case on the 300 Log you must check the log's hearing loss column — column M(5) on the current form, one of the six "injury or illness type" checkboxes (the note in the rule says this requirement took effect January 1, 2004). Classify the outcome in columns G through J like any other case; most recordable hearing loss cases involve no days away or restriction and land in column J, "Other recordable cases," but you classify on the actual outcome for that worker.

Every recordable case also needs a 301 Incident Report or equivalent under 1904.29(b)(2), within the same seven days. And because hearing loss is tracked in its own column, it rolls up on its own line of the 300A annual summary. Our step-by-step 300 Log guide covers each column, and the 300A annual summary guide covers how the totals carry forward.

Privacy. A hearing loss case is not automatically a privacy concern case. The list in 1904.29(b)(7) is closed — (b)(8) says so — and hearing loss is not on it by name. It can become one under (b)(7)(vi), "other illnesses, if the employee voluntarily requests that his or her name not be entered on the log." If the worker asks, enter "privacy case" instead of the name and keep the confidential case-number list required by (b)(6). If they do not ask, the name goes on the log. The mechanics are in our privacy concern cases guide.

Construction Employers: You Record Too

The noise standard at 1910.95 does not apply to construction — OSHA's 2007 letter recites that it excludes construction, agriculture, and gas well drilling and servicing. Construction noise is governed by 29 CFR 1926.52, which requires a "continuing, effective hearing conservation program" when exposures exceed its Table D-2 levels but does not spell out an audiometric testing schedule the way 1910.95(g) does.

The recordkeeping rule has no such carve-out. OSHA's March 4, 2004 letter answers the question directly: yes, employers in the construction industry must follow Part 1904, and "hearing losses of employees that meet the recording criteria set forth in 29 CFR 1904.10 must be recorded." If you are a construction employer and you have audiometric data showing a recordable, work-related STS, you record it.

Why This Matters Beyond the Paperwork

The numbers are from NIOSH's occupational hearing loss surveillance pages (last updated April 13, 2026): about 27 million U.S. workers are exposed to hazardous noise at work each year; roughly 11 percent of all workers have hearing difficulty; about 20 percent of noise-exposed workers who have been tested have a material hearing impairment; and 53 percent of noise-exposed workers report not wearing hearing protection. Occupational hearing loss is common, permanent, and preventable — which is why OSHA built a measurable, audiogram-based recording rule around it rather than folding it into the general criteria.

Accurate recording does two things for a small employer. It keeps you compliant with 1904.10, and it gives you the data to see whether your hearing conservation program is actually working. A pattern of recordable shifts in one work area is a signal worth acting on before it becomes a pattern of permanent loss.

There is a third consequence that surprises people: a recordable hearing loss case is a recordable case like any other, so it lands in your year-end rates. A shift recorded with no days away and no work restriction moves your TRIR but not your DART, since DART counts only cases involving days away, restricted work, or transfer — the free DART and TRIR calculator shows the size of that effect at your headcount. Employers who treat hearing loss as a side file separate from "real" injuries are often the ones surprised by their own TRIR at renewal.

Frequently Asked Questions

Is a standard threshold shift automatically recordable? No. An STS is the noise-standard trigger for notice and follow-up. It becomes recordable only when the same ear also shows a total hearing level of 25 dB or more above audiometric zero and the loss is work-related (1904.10(a)).

Does it matter that the employee wore hearing protection? Nothing in 1904.10 makes hearing-protector use a factor. The test is the audiogram plus work-relatedness under 1904.5. (Under the noise standard, exposure for program purposes is computed without regard to protector attenuation, 1910.95(c)(1).) A physician or other licensed health care professional may still conclude on the facts that occupational noise did not cause or significantly aggravate the loss.

Can I age-correct the 25 dB level? No. 1904.10(b)(3) allows age correction only for the STS determination.

Which baseline do I use for a worker I recorded before? The revised baseline — the audiogram that reflected the previous recordable case (1904.10(b)(2)(i)).

When does the 30-day retest clock start? On the date of the first test, not the date you received the results (March 4, 2004 letter of interpretation).

We recorded a case and a later audiogram shows the hearing recovered. Can we remove it? Yes. If subsequent testing under the noise standard shows the STS is not persistent, you may erase or line out the entry (1904.10(b)(4)). Keep the audiograms.

Is a hearing loss case a privacy concern case? Only if the employee voluntarily requests that their name not be entered (1904.29(b)(7)(vi)). It is not on the automatic list.

Do I need a 301 for a hearing loss case? Yes. Every recordable case entered on the 300 Log needs a 301 or equivalent within seven calendar days (1904.29(b)(2)–(3)).

Our workers are over 60. The tables stop there. Appendix F ends at "60 or older." OSHA's February 9, 2024 letter allows tables derived from NHANES data instead, provided every value comes from a single dataset, a certified audiologist, otolaryngologist, or physician makes the call, and you apply the chosen tables consistently from the baseline onward and keep a copy with the records.

We are in a state-plan state. State plans must be at least as effective as federal OSHA, and their interpretations must be at least as effective as federal interpretations (noted in the 2024 letter). Check your state's recordkeeping rule; the federal text above is the floor.

The Bottom Line

When an audiogram comes back showing a possible shift, work through the questions in order. Which baseline applies — original, or the audiogram from a previous recordable case? Is there a 10 dB or greater average shift from that baseline at 2000, 3000, and 4000 Hz, after any age correction your written policy applies? Is the total hearing level 25 dB or more above audiometric zero on the current, unadjusted audiogram — in the same ear? Is the loss work-related under 1904.5, with no physician or licensed health care professional finding otherwise? If you retest within 30 days, did the retest confirm? If every answer is yes, check column M(5) within seven days of the confirming retest, complete the 301, and let it roll into the 300A.

Get the age-adjustment rule right, use the 30-day retest window when a shift looks questionable, track the revised baseline for workers already recorded, and keep the two thresholds straight. Those habits resolve nearly every hearing loss recordkeeping question a small employer will face.

LogStead carries hearing loss as its own illness type, so when you classify a case that way the M(5) checkbox on the 300 Log and the hearing-loss line on the 300A are filled from the same entry, and the 301 is generated with it. The audiogram arithmetic above stays where it belongs — with your audiometric provider and your written age-correction policy — and the log reflects the determination you made.