When an Audiogram Looks Wrong: Validity Checks Before You Call It a Shift

An annual audiogram that shows a sudden 20 dB drop at 4000 Hz in one ear can mean two very different things. It can be real noise-induced hearing loss, or it can be a test that went wrong: an earphone sitting off-center on the ear canal, a forklift idling outside the van, an employee who pressed the button out of habit. OSHA expects the employer to sort those apart. Under 29 CFR 1910.95(g)(7)(i), each annual audiogram is compared to the baseline "to determine if the audiogram is valid and if a standard threshold shift ... has occurred." Validity comes first in that sentence for a reason. This post walks through the checks that separate a bad test from a real shift, and what the retest rule allows.

What makes an audiogram invalid in the first place?

An audiogram is invalid when something other than the employee's hearing moved the thresholds. The NIOSH-published chapter on hearing measurement lists the known error sources in occupational audiometry: audiometer calibration error or drift, excessive background noise in the test room, earphone placement, tester and procedure bias, residual temporary threshold shift from recent noise, a blocked ear canal, tinnitus, and functional (non-organic) hearing loss. Most of these push thresholds in one direction, worse, which is why an invalid test so often looks like a threshold shift. A loose earphone lets low and mid tones leak away. Ambient noise masks the quietest tones at the frequencies where it is loudest. Earwax or a collapsing ear canal under the cushion pressure blocks high frequencies. None of these is hearing loss, and all of them can push a three-frequency average past 10 dB. The technician's job before anyone writes "STS" in a report is to rule each one out.

Which quick checks catch a bad test before the employee leaves?

The fastest check happens during the test itself. The NIOSH chapter describes a standard sequence: test 1000 Hz in the better ear, then 500 Hz, then retest 1000 Hz. If the two 1000 Hz thresholds agree within 5 dB, the test continues. If they do not, the tester stops, refits the earphones, re-instructs the listener, and starts over. A listener who still cannot repeat within 5 dB should be referred for clinical testing rather than recorded as a valid audiogram.

Does OSHA allow a retest when the annual audiogram shows a shift?

Yes. 29 CFR 1910.95(g)(7)(ii) says that if the annual audiogram shows a standard threshold shift, "the employer may obtain a retest within 30 days and consider the results of the retest as the annual audiogram." The word is "may," so the retest is optional, but it is the only built-in way to replace a questionable annual result. OSHA has also said in a 2012 interpretation letter that a retest counts only when both ears are retested at the same time and place within the 30 days; a one-ear retest does not qualify. That detail matters in the field. A technician who sees a shift in the left ear and retests only the left ear has not produced a usable annual audiogram. Retest both ears in the same session, record it as the retest, and let the retest stand as the annual record. If the retest still shows the shift, the 21-day written notice clock in (g)(8)(i) starts from that determination.

Does an annual audiogram need 14 quiet hours like a baseline?

No. The 14-hour quiet period in 1910.95(g)(5)(iii) applies to the baseline audiogram only. Annual tests can be run mid-shift, and NIOSH actually recommends testing annuals during or at the end of the work shift so that any temporary threshold shift shows up. The NIOSH chapter then recommends an immediate retest when a shift appears, and if the shift holds, a confirmation test preceded by quiet. That sequence uses the temporary shift as an early warning instead of treating it as noise in the data. A shift that disappears after a quiet period still says something: the employee's hearing protection or exposure is not adequate on a normal workday. A shift that persists after quiet is the one that moves toward follow-up under (g)(8), including refitting protectors and possible referral. Record the time of the test relative to the shift so the reviewer can tell which kind of result it is.

Who decides whether a problem audiogram is real?

A technician may do the baseline comparison, per (g)(7)(i), but 1910.95(g)(7)(iii) gives "problem audiograms" to an audiologist, otolaryngologist, or physician. The employer must hand that reviewer the baseline and most recent audiograms, the background sound pressure levels in the test room under Appendix D, and the audiometer calibration records. That list is a clue to what the reviewer needs to judge validity: if ambient noise or calibration records are missing, the reviewer cannot rule those causes out, and a disputed shift becomes hard to defend either way. The reviewer can also compare the problem audiogram against earlier annuals, since a single outlier year that returns to the prior pattern looks very different from a steady decline. Foster's STS calculation guide covers the arithmetic, and the STS follow-up post covers what happens once a shift is confirmed.

How Foster handles a questionable audiogram on a testing day

On Foster's hearing testing visits the audiometric software compares each annual to the stored baseline as the test finishes, so a possible shift is visible while the employee is still in the booth. The technician then works the checklist above: look in the ear canal, refit the earphones, re-instruct, and retest both ears in the same session. The octave-band ambient reading taken at that parking spot before testing, and the daily functional check result, are already in the file. When a result still looks wrong after the retest, it goes to the reviewing professional with those records attached instead of being labeled an STS on the spot. Employers who run their own booth can lean on Foster's audiometer calibration service to keep the calibration side of the record clean, or request a quote for on-site testing.

Frequently asked questions

Can the retest happen on a different day?

Yes, as long as it is within 30 days of the annual audiogram and both ears are retested in the same session. Same-day retests are common because the employee is already there.

If the retest is normal, which audiogram goes in the file?

Both stay in the record, but under 1910.95(g)(7)(ii) the retest is treated as the annual audiogram for STS purposes.

Can bone conduction testing replace a doubtful air conduction test?

No. 1910.95(h)(1) requires pure tone, air conduction threshold testing, and OSHA has said bone conduction cannot substitute for it in the hearing conservation program.

Does a technician need certification to flag an invalid test?

A technician can do the comparison and flag it. The final call on a problem audiogram belongs to the audiologist, otolaryngologist, or physician named in (g)(7)(iii).

Sources

• 29 CFR 1910.95, Occupational noise exposure (eCFR)
• Franks, J.R., Hearing Measurement, NIOSH/CDC Stacks
• OSHA interpretation letter, Audiograms to retest hearing for threshold shifts (2012-10-12)