Hearing Conservation Program: OSHA Requirements and Compliance Guide (2026)

Occupational noise-induced hearing loss is permanent. There’s no treatment that restores it. OSHA’s 1910.95 noise standard exists because once the damage happens, you can’t undo it, and the damage accumulates silently over years before a worker notices.

Most hearing conservation program failures aren’t about malicious cost-cutting. They’re about incomplete implementation. Employers hand out earplugs, log a training session, and assume they’ve met the standard. They haven’t. A compliant hearing conservation program has five required elements, and most partial programs are missing at least two of them.

The Two Noise Thresholds You Need to Know

The action level is 85 dBA as an 8-hour time-weighted average (TWA). Once any employee is exposed at or above this level, OSHA requires you to implement a full hearing conservation program. That means all five elements: monitoring, audiometric testing, hearing protectors, training, and recordkeeping.

The permissible exposure limit (PEL) is 90 dBA TWA. At the PEL and above, hearing protection is mandatory, not just available. You’re also required to implement engineering and administrative controls to the extent feasible, even if that doesn’t reduce exposure below the PEL.

The gap between 85 and 90 dBA matters. At 85 dBA, employees must be included in the hearing conservation program and hearing protection must be available, but they can choose whether to use it. At 90 dBA, that choice goes away.

NIOSH actually recommends a lower recommended exposure limit (REL) of 85 dBA, with a 3-dB exchange rate rather than OSHA’s 5-dB. OSHA’s standard is the legal floor. NIOSH’s guidance is where the science points.

Noise Monitoring: What Triggers It and How It Works

You’re required to conduct noise monitoring when there’s reason to believe employees may be exposed at or above the 85 dBA action level. “Reason to believe” isn’t a high bar. If you work in manufacturing, construction, woodworking, metalworking, or any operation with loud machinery, you need monitoring data.

Two approaches work. Area monitoring measures noise levels at specific locations throughout the facility using a sound level meter. It works well for fixed workstations where workers stay in one place. Personal noise dosimetry is more accurate for workers who move through multiple noise environments during a shift. A dosimeter clips to the worker’s collar and measures actual exposure throughout the day.

Integrate both 80 dBA and 90 dBA into your dosimeter settings. The 80 dBA threshold matters because it’s the low end of what feeds into the TWA calculation under OSHA’s methodology.

When monitoring results change, because of new equipment, new processes, or facility changes, you repeat the monitoring. You also repeat it if you have reason to believe exposures have increased.

Monitoring results must be shared with employees. Any employee exposed at or above the action level has the right to observe monitoring, see the results, and know what the measurements mean.

The Five Required Program Elements

Monitoring is the foundation. Without measured data, everything else in the program is guesswork.

Audiometric testing is what most programs either skip or set up incorrectly. Every employee exposed at or above the action level needs a baseline audiogram within 6 months of first exposure (or within 1 year if a mobile test van is used). Annual audiograms follow every year after that.

The baseline audiogram is the reference point for everything. If an employee’s hearing shifts significantly compared to their baseline, that’s a standard threshold shift (STS). An STS is a shift of 10 dB or more averaged across 2,000, 3,000, and 4,000 Hz in either ear. When you confirm an STS, you notify the employee within 21 days, refit them for hearing protection, retrain them, and refer for clinical evaluation if indicated.

Hearing protectors must be made available at no cost to all employees exposed at or above the action level. At the PEL, use is mandatory. At the action level, it’s available and encouraged, but employees can decline. That changes the moment exposure hits 90 dBA.

Training happens annually. It must cover the effects of noise on hearing, the purpose and use of hearing protectors, and the purpose and procedures of audiometric testing.

Recordkeeping closes the loop. Every component requires documentation.

Hearing Protector Selection and the NRR

Selecting hearing protection based on the noise reduction rating (NRR) label alone will get you in trouble. The NRR is measured in lab conditions. Real-world attenuation is consistently lower.

OSHA uses a derating method. For earmuffs, the real-world protection is estimated at 75% of the labeled NRR. For formable earplugs (foam plugs), it’s 50%. For all other earplugs (premolded, banded), it’s 30%. After applying the derating, subtract 7 from the result and divide by 2 to account for the conversion between C-weighting and A-weighting scales.

An example: a foam earplug with a labeled NRR of 30. Derated at 50%, the effective NRR is 15. Subtract 7: 8. Divide by 2: 4 dB of expected attenuation in A-weighted noise. That’s far less than the 30 dB on the label.

This means many single-protection solutions are inadequate in high-noise environments. At noise levels above 100 dBA, OSHA recommends dual protection (earplugs plus earmuffs) because the attenuation available from either alone may not bring the effective exposure below the PEL.

Fit matters as much as the protector itself. A foam earplug properly inserted provides dramatically more attenuation than the same earplug loosely placed in the ear canal. Fit-testing programs for hearing protectors (using ANSI S12.71 protocols with individual fit measurement) show workers their actual attenuation, not just the label. These aren’t required by 1910.95, but they’re worth doing.

Engineering and Administrative Controls First

OSHA’s hierarchy puts hearing protection devices last, not first. Before you reach for earplugs, the standard requires you to implement feasible engineering and administrative controls.

Engineering controls reduce noise at the source. Enclosing loud machinery, adding vibration isolation mounts, installing acoustic baffles, replacing worn or unbalanced components that generate excess noise. These are often more cost-effective than they look when you factor in long-term hearing protection compliance costs.

Administrative controls reduce exposure by limiting time. Rotating workers through high-noise areas, scheduling noisy operations for times when fewer employees are present, or shifting break patterns to reduce cumulative exposure. These work but require consistent management.

The standard says “feasible” controls, not all controls. If engineering controls can reduce exposure but not below the PEL, you still implement them and supplement with hearing protection. You don’t skip engineering controls just because they won’t solve the problem alone.

This connects to broader hazard control thinking. Your job hazard analysis should capture noise hazards and document what controls are in place or why specific controls weren’t feasible.

Training Requirements Under 1910.95

Annual training is required for all employees in the hearing conservation program. The training must cover three areas: the effects of noise on hearing, the purpose of hearing protectors and instructions on fitting and use, and the purpose and procedures of audiometric testing.

The training doesn’t need to be elaborate. A 15-minute session with demonstrated fit testing, a quick explanation of how hearing loss accumulates, and a walkthrough of the audiogram process covers the requirement. Document it with dates, employee names, and trainer name.

New employees get trained before or at the time they’re enrolled in the program. Don’t wait until the annual cycle catches them.

If you’re consolidating your employer safety training requirements into an annual calendar, hearing conservation training fits cleanly alongside other OSHA-required annual training.

Recordkeeping Under 1910.95

Audiogram records stay on file for the duration of employment plus 30 years. That’s a long retention window. Make sure your records system can handle long-term retention before an employee leaves and you archive their file.

Noise exposure records (dosimetry results, area monitoring data) need to be kept for 2 years. Background data used to support the measurements also needs 2 years.

Employees have the right to access their own audiogram records. Physicians, audiologists, and OSHA can also request access during inspections or clinical evaluations.

If you’re tracking OSHA recordable injuries and illnesses, STS cases may be recordable under 29 CFR 1904. An STS confirmed at the annual audiogram, where work-relatedness is established, goes on the OSHA 300 log. The specifics of what makes an STS recordable vs. non-recordable involve age correction and work-relatedness determination.

The Most Common Program Failure

The most common hearing conservation program failure is distributing hearing protection without completing the other four program elements. A box of foam earplugs in a cabinet next to a loud machine does not constitute a hearing conservation program.

Programs collapse at audiometric testing more than anywhere else. Setting up the testing relationship with a qualified audiologist or occupational health clinic takes time and costs money. Employers defer it. A year passes, then two, and now employees who were exposed at the action level have never had a baseline audiogram. Without a baseline, you can’t identify an STS. Without STS identification, you can’t notify affected employees. The program is essentially running blind.

Get the baseline audiograms done first. Before you print training certificates, before you fit-test earplugs, get your exposed workforce into audiometric testing. That baseline is the only thing standing between your program and being legally unable to demonstrate whether your controls are working.

An OSHA 30 general industry certification covers noise standard requirements as part of the broader 29 CFR 1910 framework. If you’re building this program from scratch, that training gives supervisors the context to run it correctly and catch the gaps before an inspector does.