8.2 REIG and prescription targets

The audiologist’s verification question is: does the device, in this patient’s actual ear, produce the gain the prescription algorithm requires? The measurement that answers it is the real-ear insertion gain (REIG) — the dB-by-dB difference between the patient’s aided and unaided real-ear responses, plotted against the prescription target.

What REIG is

The REIG at each frequency is

$\text{REIG}(f) = \text{REAR}(f) - \text{REUR}(f),$

where REAR is the real-ear aided response (probe-tube SPL with the device on) and REUR is the real-ear unaided response (probe-tube SPL with the ear empty). REIG is therefore how much extra SPL the device delivers at this patient’s eardrum, beyond what the patient’s natural ear was doing on its own.

REIG is independent of:

• The patient’s ear-canal volume (both REAR and REUR are affected by the same canal acoustics, so the subtraction cancels canal effects).
• The position of the probe tube within the canal (within reason — both measurements use the same tube position).
• The presentation-level of the calibration speaker (the gain ratio is the same regardless).

REIG depends only on:

• The device’s electroacoustic processing.
• The eartip’s coupling to the canal (vent size, insertion depth, leakage).
• The interaction of the device’s output with the patient’s canal acoustics (especially at high frequencies where standing waves matter).

This is what makes REIG the clinically appropriate verification quantity: it measures what the device adds, normalised away from individual ear-canal differences that the manufacturer can’t predict from the spec sheet alone.

Prescription targets

The prescription target is the REIG (or, increasingly, the REAR at calibrated input levels) the algorithm computes for this patient at this audiometric configuration. NAL-NL2 and DSL v5 differ in their philosophies (see Lesson 7.2), but both produce per-frequency targets at three input levels:

• Soft input (50 dB SPL): the target REIG that brings soft speech to the patient’s perceived “soft but audible” range.
• Average input (65 dB SPL): the target REIG for conversational speech at the patient’s “comfortable” level.
• Loud input (80 dB SPL): the target REIG (typically reduced because compression has kicked in) for loud talkers and noisy environments.

A modern REM system measures REAR at the three input levels in quick succession and compares each to the corresponding target. A good fitting shows REIG within ±5 dB of target at every audiometric frequency (250, 500, 1000, 1500, 2000, 3000, 4000, 6000 Hz) at every input level.

NAL-NL2 vs DSL v5 in detail

NAL-NL2 prescribes less low-frequency gain than DSL v5 for the same audiogram, on the principle that low-frequency gain disproportionately raises perceived loudness without contributing to intelligibility (low-frequency speech components are not the most information-dense). NAL-NL2’s targets at 500 Hz might be 10 dB lower than DSL v5’s; at 4 kHz they might be similar.

DSL v5 prescribes more low-frequency gain on the principle of audibility — making the long-term average speech spectrum (LTASS) audible across the full audiometric range. DSL v5 also has explicit pediatric variants that prescribe more gain than the adult version, reflecting the evidence that audibility during language acquisition matters more than loudness normalisation.

The clinical default in US practice:

• Adult sensorineural fittings: NAL-NL2 by default. Audiologist may shift to DSL if patient reports needing more low-frequency information.
• Pediatric fittings: DSL v5 (pediatric variant). The audibility-driven philosophy fits the developmental requirement.
• Cochlear implant + hearing aid bimodal fitting: NAL-NL2 or DSL pediatric, depending on patient age and target.

Both algorithms are non-proprietary; both are implemented in essentially every hearing-aid fitting software. The choice between them is a clinical judgment, not a technological constraint.

Targets at non-standard frequencies

Modern REM systems target at audiometric frequencies (octave and inter-octave: 250, 500, 1000, 1500, 2000, 3000, 4000, 6000, 8000 Hz) because that’s what the audiogram is measured at. Hearing aids deliver gain at all frequencies (not just at the audiometric frequencies); the actual gain curve is a smooth function determined by the device’s filterbank. A clinically good fitting matches the target at every measured audiometric frequency and the in-between frequencies follow a smooth interpolation — meaning the audiologist should look at the shape of the REIG curve as well as the per-frequency match, ensuring no narrow notches or spikes.

The ±5 dB match criterion

Why ±5 dB? Three reasons:

1. Within-patient variability of REM measurements is about ±2 dB per frequency (test-retest, same audiologist, same patient, different sessions). Sub-5 dB precision is achievable but the practical precision of any hearing-aid fitting is limited at this level.
2. Behavioural noticeability of gain changes is about 5 dB across most of the audiometric range. A patient cannot reliably distinguish a 3 dB gain difference; they can distinguish a 5 dB difference; they readily notice 10 dB differences. Tolerating ±5 dB of fitting error keeps the patient’s perception within the just-noticeable region of optimal.
3. Clinical outcomes data (Aazh and Moore 2007, Valente et al. 2018) show that fittings within ±5 dB of NAL-NL2 across the audiometric range produce significantly better speech-in-noise outcomes and aided-listening satisfaction than fittings with larger deviations.

A fitting that is within ±5 dB at 6 of 8 frequencies but >10 dB off at one frequency is not a verified fitting — it’s a fitting with a localised deviation that needs adjustment. The ±5 dB criterion is per-frequency, not aggregate.

What deviation patterns mean

Deviations from target, viewed as a function of frequency, often have characteristic clinical interpretations:

• Uniform low-frequency excess (too much gain at 250–500 Hz): often a vent-leak issue (the device’s bass response is escaping through an oversized vent and being recaptured by the probe). Fix by reducing low-frequency gain on the device or by switching to a smaller vent.
• Uniform high-frequency deficit (too little gain at 4–6 kHz): often a poor coupler or feedback-cancellation engagement (the device may be backing off gain to prevent feedback). Fix by reseating the eartip, checking for cerumen blocking, or trying a smaller-vent custom mould.
• 3–4 kHz notch in REIG: usually a probe-tube placement issue — a standing-wave between the tube tip and the eardrum, more pronounced when the tube is far from the drum. Fix by advancing the tube and remeasuring.
• Negative REIG at high frequencies (the device is attenuating relative to the unaided ear): the device’s feedback canceller has disabled high-frequency gain entirely, or the eartip has occluded the ear without the device providing enough compensating gain. Fix by switching to a more aggressive prescription, a more powerful receiver, or a less-occluding fit.

The diagnostic skill of REM is reading these patterns and adjusting the device gain (per-channel, per-input-level) to bring the REIG into target. A practiced audiologist can do a full bilateral REM verification of soft, average, and loud inputs in 10–15 minutes.

Next lesson: the standard verification workflow, common deviations, and how REM integrates with the broader hearing-aid fitting process.