Hearing aids: selection and fitting

In short

Hearing aid selection and fitting is the clinical process of choosing an appropriate amplification device for a person with hearing loss, programming it to a prescriptive target, verifying the output objectively in the patient's own ear, and validating the benefit in real life. It is the single largest service area in Indian audiology practice, because most sensorineural losses have no medical cure and amplification is the first-line management. The topic is examined in BASLP and MSc (Audiology) papers (AIISH Mysuru, AYJNISHD Mumbai, NISH, JIPMER, SRM, Manipal and the RCI-affiliated colleges), in RCI licensure-level viva, and in technical rounds for jobs with hearing aid manufacturers and retail chains (Signia/WS Audiology, Phonak/Sonova, Oticon/Demant, Starkey, ReSound, Widex, Alps, Hearzap, Hearing Solutions, Aural Clinics). Government work under the ADIP scheme and ALIMCO supply, NPPCD district programmes and ESIC/CGHS panels all require exactly this skill set. Expect numerical questions on gain prescription, SII and real-ear measures alongside conceptual ones on compression and verification.

Explained

Candidacy, case history and pre-fitting assessment

• Candidacy is based on functional difficulty, not on the audiogram alone: a person with a 30 dB HL flat loss who teaches may need aids more than a 45 dB HL loss in a retired person.
• Conventional guidance: PTA above roughly 25-30 dB HL (adults) with communication difficulty; for children, any confirmed permanent loss is aided as early as possible, ideally within 1 month of confirmation and before 6 months of age.
• Medical clearance / red flags before fitting: sudden or rapidly progressive loss, unilateral loss, asymmetry, active discharge, pain, conductive loss with a large air-bone gap, tinnitus with vertigo, visible canal pathology - refer to ENT first.
• Pre-fitting test battery: otoscopy, pure-tone air and bone thresholds with masking, UCL/LDL, SRT and speech identification scores in quiet and noise, tympanometry and reflexes, and where needed tone decay/ABR to rule out retrocochlear pathology.
• Key derived numbers: PTA (0.5, 1, 2 kHz; four-frequency PTA adds 4 kHz), dynamic range = UCL − threshold, and the aidable region. A dynamic range under about 30 dB is a "small-dynamic-range" ear needing strong compression.
• Unaided speech measures matter: very poor speech identification (say under 40%) predicts limited hearing aid benefit and starts the cochlear implant discussion.
• Ear selection: bilateral fitting is the default (binaural summation ~3 dB, better localisation, squelch, avoids late-onset auditory deprivation). Monaural is chosen for cost, dexterity, dead ear, or asymmetric UCLs. For a dead ear, consider CROS; for a dead ear with an aidable better ear, BiCROS.

Types and classification of hearing aids

• By style: body-worn (now mainly ALIMCO/ADIP legacy and profound loss), BTE (behind-the-ear), RIC/RITE (receiver-in-canal), ITE (in-the-ear), ITC (in-the-canal), CIC (completely-in-canal), IIC (invisible-in-canal), eyeglass (obsolete).
• By signal processing: analogue (class A/B/D amplifiers, trimmer-adjusted), digitally programmable analogue (transition), and digital signal processing (DSP) aids - the current standard, with multiple channels, bands and memories.
• By conduction route: air-conduction aids; bone-conduction aids (spectacle, steel headband, softband) for atresia, chronic discharging ears and bilateral conductive loss; bone-anchored implants (BAHA/Ponto/Bonebridge) and middle-ear implants - surgical, so beyond routine fitting.
• Special configurations: CROS/BiCROS for unilateral deafness; frequency-lowering (frequency compression/transposition) aids for severe high-frequency dead regions; extended-bandwidth aids.
• By level of technology: essential/basic (few channels, limited noise management) to premium (many channels, adaptive directionality, machine-learning scene classifiers, Bluetooth LE Audio streaming, rechargeable Li-ion). Indian retail price bands run roughly ₹8,000-25,000 (basic), ₹25,000-75,000 (mid) and ₹75,000-2,50,000+ per pair (premium); ALIMCO digital BTEs supplied under ADIP are far cheaper. Prices change, so treat these as indicative.
• Over-the-counter (OTC) self-fitting aids exist as a US regulatory category; in India they are not a separate licensed class, and online "sound amplifiers" are commonly sold without audiological fitting - a known quality problem.

How a modern hearing aid works - the signal chain

1. Microphone(s) convert sound to an electrical signal; two or three mics allow directionality.
2. A/D conversion (typically 16-24 bit, 16-32 kHz sampling) then DSP: splits the signal into channels by a filter bank.
3. Gain and compression per channel: WDRC applies frequency-shaped, level-dependent gain.
4. Output limiting: peak clipping (distorting) or compression limiting / AGCo (preferred) keeps output below UCL; MPO/OSPL90 is the ceiling.
5. Feature block: adaptive noise reduction, impulse-noise and wind-noise control, feedback cancellation (phase inversion/notch filtering), frequency lowering, tinnitus noise generator.
6. D/A and receiver deliver sound to the ear; telecoil gives an alternative magnetic input; DAI/FM, Bluetooth, 2.4 GHz and LE Audio/Auracast provide wireless input.
7. Compression terms: compression threshold (kneepoint), compression ratio (CR), attack and release time, channels (where gain is computed) versus bands (where frequency response is shaped).
8. Compression types: linear (CR 1:1 with limiting), output compression (AGCo) - high kneepoint, protects against loudness discomfort, input compression (AGCi)/WDRC - low kneepoint (~40-50 dB SPL), low CR (1.5:1 to 3:1), restores audibility of soft speech and normalises loudness. Syllabic compression = fast release (<100 ms); dual/adaptive compression varies release with the input.

Prescriptive selection: formulas and rationale

• Two philosophies: loudness normalisation (restore normal loudness at each frequency - DSL) and loudness equalisation (make all speech bands equally loud and maximise intelligibility - NAL family).
• Linear-era threshold formulas: half-gain rule (Lybarger), POGO and POGO II, Berger, NAL-R and NAL-RP (for severe/profound), Libby one-third/two-thirds.
• Suprathreshold/comfort formulas: MCL-based procedures, Shapiro, Bragg; now mostly historical.
• Non-linear current formulas: NAL-NL1 → NAL-NL2 (National Acoustic Laboratories, Australia; maximises speech intelligibility at normal loudness; accounts for age, gender, experience, language - tonal versus non-tonal, bilateral/unilateral, compression speed) and DSL m[i/o] v5.0 (University of Western Ontario; audibility and loudness normalisation across input levels; separate adult and child targets - the paediatric default).
• NAL-NL2 generally prescribes less overall gain (especially low frequencies and for severe loss) than DSL v5 child targets; DSL gives more audibility, important for language learning.
• Each manufacturer also has a proprietary first-fit (Phonak Adaptive Digital, Oticon VAC+, Signia, ReSound Audiogram+). First-fit is *not* the target - manufacturer first-fits routinely under-deliver gain versus NAL-NL2/DSL by 5-15 dB, which is why REM is needed.
• Earmould/coupling selection is part of prescription: acoustic coupling changes the real-ear response as much as software gain.

Earmoulds, acoustics and the occlusion effect

• Impression taking: otoscopy → place an otoblock/cotton dam past the second bend → inject silicone/addition-cure material from the deepest point out → cure → remove with gentle rotation. Contraindications: perforation without care, active discharge, grommet, recent surgery, excessive wax, exostoses.
• Mould styles: full/shell, skeleton, half-skeleton, canal, canal-lock, shell with CROS tubing; materials: hard acrylic, soft vinyl/silicone (children, profound loss, better seal), semi-soft.
• Acoustic modifications: vent (parallel/diagonal/trench; cuts low-frequency gain, relieves occlusion, increases feedback risk), damper (750-3,000 Ω, smooths the 1 kHz tubing peak), horn/Libby horn (boosts 2-4 kHz), reverse horn (boosts lows), tube length/diameter (changes resonance).
• Occlusion effect: own-voice boominess from bone-conducted low-frequency energy trapped in an occluded canal; measured as the objective occlusion effect (ear-canal SPL with /i/ vocalisation, occluded minus open). Remedies: larger vent, open fitting, deeper canal seal into the bony portion, low-frequency gain reduction.
• Open-fit/RIC with a thin tube and open dome suits mild-moderate high-frequency loss; it needs effective feedback cancellation and loses usable low-frequency gain.
• Feedback occurs when amplified sound leaks back to the microphone; maximum stable gain limits how much gain is usable, especially in open and vented fittings.

Fitting, verification and validation - the clinical sequence

1. Select style, technology level, coupling and features against the audiogram, dexterity, ear anatomy, lifestyle, cognition, vision and budget.
2. Pre-programming: enter audiogram, UCL, age, experience and acoustic coupling into the manufacturer software; choose the prescriptive rule (NAL-NL2 for adults, DSL v5 child for children).
3. Electroacoustic check: verify the aid against specifications in a 2 cc coupler test box (ANSI/ASA S3.22 and IEC 60118-0/60118-7 measures: OSPL90, full-on gain, frequency response, THD, equivalent input noise, battery current, attack/release times, telecoil SPLITS). IS standards for hearing aids exist historically but are largely aligned to IEC - quote IEC 60118 series unless sure of a current IS number.
4. Verification by real-ear measurement (REM/probe-microphone): measure REUR/REUG, REOR/REOG, REAR/REIG, RECD and REDD; match REAR for speech inputs (55/65/80 dB SPL) to the prescriptive target and REAR90 to the UCL/MPO limit. Speech mapping displays these against the speech spectrum.
5. Alternative verification where REM is not available: functional gain (aided minus unaided sound-field thresholds, warble tones at 45° azimuth) - limited by floor effects, noise, non-linear processing and single-input level; aided speech testing; coupler-plus-RECD simulation (essential in infants).
6. Fine-tuning to complaints: too sharp/tinny → reduce 2-4 kHz; boomy/hollow → reduce low-frequency gain or vent more; own voice too loud → occlusion management; loud sounds uncomfortable → lower MPO/compression kneepoint or raise CR at high levels; speech unclear in noise → directionality, noise reduction, channel-specific gain.
7. Orientation and counselling (HIO basics): insertion and removal, battery/charging and battery safety, on-off and programme/volume controls, telecoil and loop use, phone and TV streaming, wearing schedule (full-time use; graded acclimatisation for new users), cleaning, moisture and wax management, do's and don'ts, warranty and service interval, realistic expectations.
8. Validation of real-world benefit: self-report tools - COSI (Client Oriented Scale of Improvement), APHAB, IOI-HA (translated into several Indian languages), HHIE/HHIA, SADL; for children PEACH, TEACH, LittlEARS, ELF and parent questionnaires; datalogging gives objective use hours.
9. Follow-up: typically 1-2 weeks, 1 month, 3 months, then annually; re-verify after repair, remake or audiogram change. Report hours/day of use from datalogging - under about 4 hours/day signals a problem.
10. Auditory training and ALDs complete the plan: FM/remote microphone and induction loops in classrooms, amplified telephones, TV streamers, alerting devices.

Paediatric and Indian regulatory context

• Children need RECD-based fitting: measure the child's RECD, fit to DSL v5 child targets, use the SII to document audibility, re-earmould often as the ear grows, and avoid relying on behavioural aided thresholds alone.
• Rehabilitation Council of India (RCI) Act, 1992: only RCI-registered audiologists/speech-language pathologists may practise; registration is renewable and continuing rehabilitation education is required.
• Rights of Persons with Disabilities Act, 2016 with its Rules: "hearing impairment" (deaf and hard of hearing) is a specified disability; benchmark disability is 40% or more. Disability certification uses the Ministry of Social Justice and Empowerment gazette guidelines (percentage computed from better-ear and speech discrimination data) and the UDID card.
• ADIP scheme (Assistance to Disabled Persons for Purchase/Fitting of Aids and Appliances, MSJE) funds aids through ALIMCO and implementing agencies, with income-based eligibility slabs; NPPCD (National Programme for Prevention and Control of Deafness) runs district-level screening and aid provision. Scheme ceilings and income limits are revised periodically - confirm the current circular before quoting figures.
• Hearing aids are medical devices notified under the Medical Devices Rules, 2017 (CDSCO) - manufacture and import need licences; Class A/B risk classification applies. GST on hearing aids and parts has been nil/exempt for years; after the 22 September 2025 rate restructuring the slabs are 5%, 18% and 40%, with hearing aids and specified assistive devices continuing in the concessional/exempt treatment - verify the current notification before quoting a rate.
• Ethics: no over-selling of premium technology, informed written consent, data privacy (currently the IT Act s.43A and SPDI Rules; DPDP Rules, 2025 Data Fiduciary duties apply from 13 May 2027), and referral when a cochlear implant is the better option.

Key points

• Candidacy - functional difficulty plus audiometric loss, not PTA alone; red flags go to ENT first.
• Dynamic range - UCL minus threshold; it sets how much compression the ear needs.
• Bilateral default - gives binaural summation, squelch, localisation and prevents late-onset deprivation.
• WDRC - low kneepoint, low compression ratio; makes soft speech audible and loud speech comfortable.
• Channels vs bands - channels compute gain independently; bands only shape the frequency response.
• OSPL90/MPO - maximum output with a 90 dB SPL input; must stay below the patient's UCL.
• Half-gain rule - insertion gain ≈ 0.5 × threshold (HL); the historical linear starting point.
• NAL-NL2 - intelligibility-maximising non-linear target, adult default; less gain than DSL, especially in lows.
• DSL v5 child - loudness-normalisation target, paediatric default; maximises audibility for language learning.
• REUG - unaided ear-canal gain, open-ear resonance peak about 2.7 kHz, roughly 17-20 dB.
• REIG = REAR − REUR - the gain the aid actually adds in that ear; the quantity matched to target.
• RECD - real-ear-to-coupler difference; converts coupler data to real ear, essential for infants.
• Functional gain - aided minus unaided sound-field thresholds; crude, floor-limited, unreliable with non-linear aids.
• Occlusion effect - own-voice boominess; fixed by venting, open fitting or deep seal.
• Maximum stable gain - the feedback ceiling; venting and open domes lower it.
• Verification ≠ validation - verification is electroacoustic (REM/coupler); validation is real-world benefit (COSI, APHAB, IOI-HA, datalogging).
• Datalogging - objective daily use hours; the honest check on reported benefit.
• Indian framework - RCI Act 1992 for practice, RPwD Act 2016 for rights and certification, ADIP/ALIMCO and NPPCD for provision, CDSCO Medical Devices Rules 2017 for the device.

Formulas and examples

Example 1 - Dynamic range, half-gain and MPO. Right ear: 0.5 kHz 45, 1 kHz 55, 2 kHz 65, 4 kHz 70 dB HL; UCL 100 dB HL at 1 and 2 kHz.

• PTA (3-frequency) = (45 + 55 + 65)/3 = 165/3 = 55 dB HL → moderate loss.
• DR at 1 kHz = 100 − 55 = 45 dB; at 2 kHz = 100 − 65 = 35 dB → moderate compression needed, more at 2 kHz.
• Half-gain targets: 0.5 kHz = 22.5, 1 kHz = 27.5, 2 kHz = 32.5, 4 kHz = 35 dB insertion gain.
• MPO: UCL 100 dB HL at 2 kHz; with RETSPL ≈ 9 dB at 2 kHz (TDH-39, 6 cc coupler), that is about 109 dB SPL in the ear, so set OSPL90 at 2 kHz a few dB below - say 103-106 dB SPL to keep a safety margin.

Example 2 - POGO versus Berger at 1 kHz. HL at 1 kHz = 60 dB.

• Half-gain: 0.5 × 60 = 30 dB. POGO above 0.5 kHz = same as half-gain = 30 dB.
• Berger at 1 kHz: 60/1.6 + 10 = 37.5 + 10 = 47.5 dB - markedly more gain, as Berger targets speech-frequency emphasis.
• At 0.25 kHz with HL 40 dB: POGO = 0.5 × 40 − 10 = 20 − 10 = 10 dB, versus half-gain 20 dB - POGO cuts the lows to reduce upward spread of masking.

Example 3 - REIG from probe-microphone data. At 2 kHz, REUR = 17 dB, REAR (65 dB SPL speech input) = 92 dB SPL in the canal; the free-field input at 2 kHz is 62 dB SPL for that speech band.

• Aided ear-canal gain = 92 − 62 = 30 dB. REIG = 30 − 17 = 13 dB.
• If the NAL-NL2 target REIG at 2 kHz is 20 dB, the fitting is 7 dB under target → raise 2 kHz channel gain and re-measure. This is the classic manufacturer first-fit shortfall.

Example 4 - Compression arithmetic. WDRC channel: kneepoint 45 dB SPL, CR 2.5:1, gain below kneepoint 25 dB.

• Input 45 dB SPL: output = 45 + 25 = 70 dB SPL (gain 25 dB).
• Input 65 dB SPL: output = 70 + (65 − 45)/2.5 = 70 + 8 = 78 dB SPL → gain = 78 − 65 = 13 dB.
• Input 85 dB SPL: output = 70 + (85 − 45)/2.5 = 70 + 16 = 86 dB SPL → gain = 1 dB.
• So a 40 dB input range is squeezed into 16 dB of output range, and gain falls 25 → 1 dB: exactly what a reduced dynamic range requires.

Example 5 - Paediatric coupler-plus-RECD prediction. An 8-month-old infant; measured RECD at 2 kHz = 14 dB (adult average about 5 dB). The chosen aid in a 2 cc coupler gives 88 dB SPL at 2 kHz for a 65 dB SPL speech input.

• Predicted REAR = 88 + 14 = 102 dB SPL in the infant's ear.
• If the DSL v5 child target REAR at 2 kHz is 95 dB SPL, the aid is 7 dB over target → reduce gain, else risk loudness discomfort and over-amplification. The small ear canal is why infant fitting must never use adult average RECDs.

Where it is used

• Hospital and clinic audiology: ENT departments of AIIMS, PGIMER, CMC Vellore, JIPMER, Apollo, Fortis; district hospitals under NPPCD. Role titles: Audiologist, Clinical Audiologist, Audiologist cum SLP.
• National institutes and training: AIISH Mysuru, AYJNISHD Mumbai, NISH Thiruvananthapuram, PDUNIPPD Delhi - clinical fitting, RECD/REM labs, research on Indian-language validation tools.
• Manufacturers and distributors: product/clinical trainer, audiology specialist and field support roles with WS Audiology (Signia), Sonova (Phonak, Unitron), Demant (Oticon), Starkey, GN ReSound, Widex, Alps Hearing (Indian manufacturer) and ALIMCO Kanpur.
• Retail hearing aid chains: Hearzap, Hearing Solutions, Aural Clinics, Sound Life, Sanjeevni Hearing - selection, trial, fitting, REM verification, after-sales and counselling, plus GST/warranty paperwork.
• Government schemes and camps: ADIP/ALIMCO fitment camps, NPPCD screening, UDID disability certification support, ESIC/CGHS/Railways panel fittings, state school-health programmes.
• Industrial and occupational: hearing conservation in noisy plants (audiometry, hearing protector selection, aid fitting for noise-induced loss), now administered under the OSH&WC Code, 2020 framework.

Interview questions

• Q: What is the difference between verification and validation of a hearing aid fitting? A: Verification is objective electroacoustic confirmation that the aid delivers the prescribed output in that ear (REM, coupler measures); validation measures the real-world benefit the patient experiences (COSI, APHAB, IOI-HA, datalogging).
• Q: Why is real-ear measurement preferred over functional gain? A: REM uses realistic speech inputs, works with non-linear processing, is not limited by ambient noise or threshold floor effects, and gives frequency-specific output in the patient's own ear canal.
• Q: State the half-gain rule and its limitation. A: Prescribed insertion gain is about half the hearing loss in dB; it is a linear threshold-based estimate, over-prescribes low-frequency gain and ignores suprathreshold loudness and recruitment.
• Q: When would you choose DSL v5 over NAL-NL2? A: For children and language-learning cases, where maximising audibility across input levels matters; NAL-NL2 is the adult default because it maximises intelligibility at a comfortable overall loudness.
• Q: What is RECD and why is it critical in infants? A: It is the difference in dB between the SPL produced in the real ear and in a 2 cc coupler by the same source; infant canals are small with high RECD, so using adult averages would over-amplify by 10 dB or more.
• Q: How do you manage a patient who says "my own voice sounds boomy"? A: That is the occlusion effect - enlarge the vent or move to an open dome, seal deeper into the bony canal, and reduce low-frequency gain; confirm with an objective occlusion-effect measurement.
• Q: What causes acoustic feedback and how is it controlled? A: Amplified sound leaking from the canal back to the microphone; controlled by a better-fitting mould, smaller vent, phase-cancellation/notch feedback managers and reducing high-frequency gain, accepting a lower maximum stable gain.
• Q: Why is OSPL90 set with reference to the UCL? A: Output must never exceed the patient's uncomfortable loudness level, otherwise the aid is rejected and could add further noise-induced damage; REAR90 is checked against UCL converted to ear-canal SPL.
• Q: What is the difference between channels and bands? A: Channels are independent compression/gain processing paths; bands are frequency regions where the response can be shaped - a 6-channel, 20-band aid computes compression in 6 paths but shapes gain in 20 regions.
• Q: When is CROS or BiCROS indicated? A: CROS for a single unaidable ear with a normal other ear; BiCROS when the better ear also needs amplification - the signal from the dead side is routed to the better ear.
• Q: A patient has 30% speech identification bilaterally with aids properly verified. What next? A: Document aided audibility and benefit, then counsel for cochlear implant evaluation, since amplification alone cannot restore speech understanding at that level.
• Q: Which Indian law governs your right to fit hearing aids, and which scheme funds them? A: Practice requires RCI registration under the Rehabilitation Council of India Act, 1992; funding for eligible persons comes mainly through the MSJE's ADIP scheme (ALIMCO supply) and NPPCD, under the rights framework of the RPwD Act, 2016.

Common mistakes

• Treating the manufacturer's first-fit as the prescriptive target and skipping REM - first-fits typically fall 5-15 dB short of NAL-NL2/DSL.
• Using aided sound-field (functional gain) thresholds to verify non-linear aids, where low-level test tones sit in the high-gain region and overstate benefit.
• Fitting an infant with adult average RECD, or with NAL-NL2 instead of DSL v5 child targets.
• Fitting gain from thresholds only and never measuring UCL, so MPO is left above the discomfort level.
• Forgetting the bilateral correction and prescribing full monaural gain to both ears, producing an over-loud fit.
• Confusing compression ratio with gain, or quoting a kneepoint in dB HL instead of dB SPL input.
• Over-venting a severe loss (loses low-frequency gain and causes feedback) or sealing a mild high-frequency loss (occlusion and rejection).
• Reporting a fitting as successful on patient report alone, without datalogging hours or a validation questionnaire.

Related topics

Diagnostic audiology: audiometry, immittance, OAE and ABR · Paediatric audiology and newborn hearing screening · Cochlear implants and aural rehabilitation · Psychoacoustics and speech perception in noise · Industrial audiology and hearing conservation · Earmould technology and assistive listening devices.