Share article
OSHA 1910.95 requires employers to retain audiometric records for the duration of employment. The regulation specifies what must be kept — but not how. Across occupational hearing conservation programs, that format choice ranges from fully digital cloud databases with automatic threshold-shift calculations to PDF exports stored in shared drives, to manila folders in a filing cabinet. The format distinction is not administrative trivia. It determines whether a worker’s progressive hearing loss gets detected in year three or year nine. It determines whether an STS gets flagged or missed. It determines whether a program survives a litigation review or falls apart under discovery. And it shapes whether the surveillance function that NIOSH argues is essential to preventing occupational hearing loss can actually be performed.
Soundtrace’s cloud portal maintains complete digital audiometric histories per worker, with automated STS calculations, trend visualization, and per-audiogram communication between client staff and the Soundtrace audiology team — eliminating every failure mode of paper-based programs.
Under 29 CFR 1910.95(m), employers must retain audiometric test records for the duration of each affected employee’s employment. The standard requires that these records be available to employees, former employees, and their representatives on request. The regulation specifies the minimum contents of what must be recorded — name, job classification, date, examiner identity, most recent noise exposure measurement, and the audiogram itself — but says nothing about format.
That format silence has produced a wide spectrum of actual practice. Some employers maintain digital databases with longitudinal trend views and automated STS calculations. Many more maintain folders of paper printouts, stacks of PDF exports from a testing vendor’s system, or spreadsheets manually maintained by safety staff who were hired after the testing records they are responsible for were created. OSHA does not prescribe which of these approaches is acceptable — but the practical consequences of the choice become apparent when a record is requested, when an STS is reviewed, or when an audiometric history must be reconstructed under litigation.
In 2009, NIOSH established the Occupational Hearing Loss Surveillance Program to collect de-identified worker audiograms from audiometric service providers and analyze occupational hearing loss trends by industry. As of its most recent reports (2024), the program has collected millions of audiograms from thousands of U.S. workplaces. The program’s published statistics — including the finding that approximately 18% of all manufacturing workers have hearing difficulty and about 20% of noise-exposed tested workers have material hearing impairment — are derived entirely from this structured digital audiometric database.
The surveillance program explicitly requires structured digital data contributions from participating providers. A manila folder of PDFs cannot feed a longitudinal database. A spreadsheet of manually transcribed threshold values cannot generate the population-level trend analyses that NIOSH uses to identify high-risk industries and evaluate the effectiveness of hearing conservation interventions. The research infrastructure that guides OSHA’s enforcement priorities and NIOSH’s intervention recommendations is built on the premise that audiometric data is machine-readable, longitudinally linked, and structured — not on the premise that it is printed, filed, and retrievable only by a human flipping through folders.
Paper and PDF programs fail not because the data does not exist, but because the data exists in a form that cannot be processed. The AIHA has catalogued the most common failure modes in hearing conservation data management, several of which have been observed repeatedly across industrial HCP programs:
1. The retest record that never makes it back into the file. When a worker is referred for a retest at an outside clinic, they often return with a paper audiogram. That paper audiogram must be manually filed, scanned, and linked to the worker’s history. It frequently is not — or it is filed in the wrong location, attached to the wrong worker’s record, or simply lost.
2. Stale baselines that no one updated. OSHA requires that a worker’s baseline audiogram be revised when a revised baseline is more protective (i.e., reflects better hearing). In a paper system, baseline revision requires manually identifying that a revision is warranted, pulling the original baseline, annotating it, and creating a new reference record. In programs with high employee volume, this rarely happens consistently.
3. The provider-transition data loss event. When an employer changes audiometric testing vendors, data migration is rarely clean. AIHA’s documentation of this problem is explicit: data settings may be lost, records incorrectly uploaded, and baseline indicators disrupted. For employers who have changed providers multiple times over a 20-year program, the audiometric history may have gaps, duplications, or miscoded records that are invisible until someone tries to reconstruct the longitudinal picture under legal scrutiny.
4. The STS calculation that was never performed. In a paper system, STS calculation requires manually comparing the current audiogram to the baseline record at each of the relevant frequencies, summing the shift, and applying age correction if used. This calculation is not performed automatically by any paper or PDF system. It is performed only if someone performs it — and in programs where the Professional Supervisor reviews audiograms in batches rather than individually in real time, audiograms that should have triggered an STS review may pass through months before the shift is noticed.
5. The calibration and ambient noise record that was never linked. OSHA requires that audiometric test rooms meet maximum permissible ambient noise levels, and that audiometers be calibrated per Appendices C and D of 1910.95. In paper programs, these records exist as separate documents — often in a different file from the individual audiograms. When an OSHA inspector or plaintiffs’ attorney asks whether the ambient conditions at the time of a specific audiogram were within permissible limits, the paper program must manually correlate two separate document sets. A digital system that logs ambient conditions per test event answers the question instantly.
6. The 30-year retention problem. NIOSH recommends retaining audiometric records for 30 years. OSHA requires retention for the duration of employment. For a worker hired at age 25 who retires at 65 with a WC hearing loss claim, that is a 40-year record chain. Paper records degrade, file cabinets are cleared during office moves, and safety staff retire. In NIOSH’s history of occupational hearing loss research, the inability to establish and maintain surveillance data — explicitly noted by the National Academies in its 2005 evaluation of the NIOSH Hearing Loss Research Program — has been the single largest constraint on understanding the true burden of occupational hearing loss in the U.S.
The value of a structured digital audiometric record is not merely that it is harder to lose. It is that it is computable. A frequency-specific threshold value stored as a number in a database can be subtracted from a baseline value, compared against the Appendix F age-correction tables, averaged across frequencies, and compared against a 10 dB STS criterion — all automatically, at the moment the test is completed, with no human calculation required and no risk of arithmetic error. A PDF cannot do this. A paper audiogram cannot do this. Neither can a scanned image of an audiogram stored in a document management system.
The 2024 editorial on digital hearing health in Frontiers in Neuroscience describes the field’s shift toward computational audiology — the integration of automated audiometry, machine learning, and digital data infrastructure to improve both access and analytical depth. That shift is happening at the clinical and research level precisely because structured digital audiometric data enables analyses that paper-based data simply cannot support.
One of the most consequential and least-discussed risks in occupational hearing conservation is what happens to audiometric records when an employer changes testing vendors. The AIHA has documented this pattern directly: when clients change service providers, data settings may be lost in the process, records may be incorrectly uploaded into the new database, and out-of-date baseline indicators are a common result. For employers who have changed providers multiple times, these errors compound.
The specific risk of a stale or incorrect baseline indicator is particularly serious. OSHA’s STS calculation requires comparing the current audiogram to the established baseline. If the baseline record is lost during a vendor transition, or if it was never migrated correctly, the STS calculation cannot be performed accurately. The employer may continue testing workers annually and generating audiograms that appear compliant — while the cumulative shift from the actual original baseline goes undetected because the reference point no longer exists.
Before engaging any audiometric testing service, employers should ask for a written data portability commitment: in what format will records be exported, at what level of granularity (raw threshold values, not just PDFs), and who owns the data if the relationship ends. An employer who cannot export their own audiometric data in a structured, machine-readable format does not truly own their compliance records — they own a vendor relationship.
When an OSHA compliance officer arrives for an inspection of a hearing conservation program, or when a workers’ compensation plaintiff’s attorney requests audiometric records in discovery, the employer faces a document production requirement with real legal consequences. A digital system produces a complete, timestamped, searchable record export in minutes. A paper system requires manual retrieval, physical review, and a chain-of-custody argument for every individual document.
More significantly, under litigation, the employer who cannot produce a complete longitudinal audiometric record — particularly the original baseline and the series of annual audiograms showing progressive threshold change — faces the inference that the records are incomplete, that the program was not operating as required, or that records have been selectively withheld. A gap in the record chain caused by a provider transition is functionally indistinguishable from a gap caused by deliberate suppression, from the perspective of a reviewing court.
Soundtrace maintains complete, structured digital audiometric histories for every enrolled worker — with automated STS calculations, trend visualization, per-test ambient noise validation, and data portability that survives any provider transition.
Get a Free Quote