8.3 Verification protocol and clinical practice

The audiometric and prescriptive groundwork (Chs 2 and 7) feeds a clinical workflow that, in best-practice form, follows a six-step sequence. This lesson covers that sequence — the standard REM verification protocol, the variations for special populations, and the common deviations from target with their fixes.

The six-step verification

1. Otoscopic examination and probe-tube placement

Before any electrical measurement, inspect the ear canal. Cerumen (earwax) impaction in the canal will produce REM artefacts and is a contraindication to REM until cleared. Active otorrhea (drainage) is also a contraindication. Once the canal is clear, the probe tube is inserted to the appropriate depth (see Lesson 8.1): typically within 5 mm of the tympanic membrane in adults, less in children.

The probe tube’s acoustic placement is verified by sweeping a high-frequency tone (typically 4 kHz) and observing whether a notch appears in the response. A notch indicates a standing-wave between the tube tip and the drum; absence of notch indicates the tube is close enough to the drum that the wavelength-to-distance ratio prevents notches in the audiometric range. The placement is iterated until satisfactory.

2. REUR — real-ear unaided response

With the probe in place and the ear empty, a calibrated speaker (positioned 0.5–1 m from the patient, typically at 45° azimuth from the ear) delivers a swept reference tone (often warble tones at audiometric frequencies, or a continuous swept sine). The probe records the SPL at the eardrum. The REUR establishes the patient’s individual ear acoustics — the head/pinna/canal resonance — that the device will need to either preserve (in open-fit) or recreate (in closed-fit).

3. Insert device, perform REOR

The hearing aid is placed in the ear with the eartip seated, the device turned off. The probe tube runs alongside the eartip. The same calibrated signal is re-presented; the probe now records the canal’s response with the device occluding it but adding no gain. The REOR shows the insertion-loss that the device must first overcome before contributing positive gain.

4. Power up, perform REAR at three input levels

The device is turned on. The same calibrated speaker now delivers a speech-shaped signal — typically the ISTS (international speech test signal, Holube et al. 2010), a multi-talker pseudo-speech stimulus with the same long-term spectrum and modulation statistics as real speech. The ISTS is presented at three calibrated input levels (50, 65, 80 dB SPL), and the REAR is captured at each.

The three measurements together form the input/output trajectory of the device in the patient’s ear. The REAR-vs-input slope at any frequency reveals the effective compression ratio at that frequency in this ear — which can differ from the device’s programmed CR because of the canal acoustics and the WDRC time constants interacting with the modulated speech signal.

5. Compare to targets, adjust

The three REAR curves (or equivalently, the three REIG curves) are compared to the prescription targets (NAL-NL2 or DSL v5) at each audiometric frequency. Deviations exceeding ±5 dB are flagged. The audiologist adjusts the device’s per-channel gain values and re-measures, iterating until all frequencies at all input levels are within target.

In practice this takes 2–3 iterations for a typical fitting. The device fitting software supports a “match-to-target” mode in which the audiologist’s adjustments map directly to the relevant per-channel gain parameter; the REM system imports the prescription target from the audiogram and shows the deviation live.

6. Maximum output verification

Finally, a brief loud-input test confirms the device’s MPO doesn’t exceed the patient’s comfort. A short high-level signal (typically a 90 dB SPL pure tone at one or two frequencies, or a brief loud sweep) is presented and the REAR peak is verified to not exceed the patient’s UCL — typically 100–105 dB SPL in adults, lower in pediatrics. This guards against over-loud peaks that could cause discomfort or further cochlear damage.

The complete six-step verification, for both ears, takes 15–25 minutes of clinical time in a well-equipped office.

Special populations

Pediatric REM

The fundamental procedure is the same but the parameters shift. Pediatric ears have:

• Smaller canal volumes (0.2–0.5 mL for infants vs 1.5 mL for adults).
• Shorter canals (15–20 mm vs 25–30 mm).
• Higher RECD (small cavities produce more SPL at the drum for the same input).
• Different head transfer functions (the pinna and head are still developing).

Pediatric REM uses measured RECD rather than the age-averaged estimate that suffices for routine adult fittings. The patient-specific RECD is measured once at fitting and re-measured every 3–6 months in growing children (canal grows; transfer function changes).

For non-cooperative pediatric patients (infants, toddlers, sleeping children), the REM workflow is compressed: the probe tube is inserted briefly while the child is calm, the necessary measurements taken quickly, and the device adjusted iteratively over multiple short sessions if needed.

Severe / profound loss

Patients with severe-to-profound losses (PTA > 70 dB HL) have devices that produce 50–70+ dB of real-ear gain at high frequencies. The verification protocol is the same but with several additions:

• Maximum output verification is critical — high-gain devices can produce peaks of 130+ dB SPL output, well above the patient’s UCL.
• Tight feedback margins mean the device may be working close to the AFC limit; small canal-acoustics changes (head turning, hand near ear) can disrupt the feedback canceller. REM is repeated with the patient in different head positions.
• Frequency lowering, if enabled, is verified separately — the audiologist confirms that the high-frequency input bands are being correctly remapped to the lower frequencies and that the remapped signal is audible.

Cochlear implant + hearing aid bimodal

In bimodal fittings (CI on one ear, HA on the other), the HA side is REM-verified normally. The CI side cannot be REM-verified directly (the implant’s electrical stimulation has no acoustic output at the eardrum), but its mapping is verified through the implant manufacturer’s own software using behavioural threshold and most-comfortable-level measurements. Coordinating the two devices for binaural cues — especially equalising loudness across ears — is part of the bimodal-fitting protocol that often takes 3–5 sessions to optimise.

Why REM uptake remains low

Surveys consistently find that fewer than 30% of US audiologists routinely use REM for verification — despite unambiguous best-practice guidelines and accumulating outcomes evidence that REM-verified fittings outperform non-verified ones. The reasons cluster:

1. Equipment cost. A clinical REM system runs $5,000–$15,000 USD, a non-trivial capital expense for solo and small-practice audiologists. Newer “REM-in-aid” technologies (where the hearing aid itself can perform a coarse self-verification using its own microphones) lower this barrier but don’t yet match clinical REM accuracy.
2. Time per patient. A REM-verified fitting takes 60–90 minutes of clinical time. Non-verified fittings take 30–45 minutes. In a high-volume practice, the additional time is real revenue cost.
3. “Sufficient” outcomes without verification. Most patients adapt to whatever fitting they receive, and a significant fraction report satisfaction even with non-verified fittings. The marginal benefit of verification is statistically significant but not always individually obvious.
4. Training gap. Many practicing audiologists graduated before REM was emphasised in the Au.D. curriculum. Continuing-education uptake on REM lags clinical practice norms; manufacturer-led training (which is the primary CE pathway for most clinicians) often de-emphasises REM in favour of manufacturer-specific fitting features.

The gap between recommended and actual practice is the single largest preventable source of underperforming hearing aids in the field. Outcomes research consistently shows that REM-verified fittings produce 5–15 percentage-point higher patient-reported benefit scores than non-verified fittings, and the data is strong enough that the major US payors (Medicare, VA) increasingly require REM documentation for reimbursement.

Closing the chapter

That closes Chapter 8. The arc: a hearing aid programmed by audiogram and manufacturer spec alone is a starting point — REM converts that starting point into a verified fitting by measuring what actually reaches the patient’s eardrum and adjusting the device’s per-channel gain to bring the real-ear response within ±5 dB of prescription target. The probe-tube microphone, the prescription algorithms (NAL-NL2 / DSL v5), and the ±5 dB match criterion together define the modern standard of care.

The next two chapters move beyond conventional hearing aids to the implanted devices that address losses too severe for amplification to help.

Next chapter: Ch 9 — Cochlear implants. The electrode array, the speech processor’s continuous-interleaved-sampling strategy, current spread along the array, and the pitch coding of electric hearing.