Paediatric audiology and newborn hearing screening

In short

Paediatric audiology is the identification, measurement and management of hearing loss in infants and children, where the patient cannot give a reliable voluntary response and the goal is to put sound into a developing auditory system before the critical period for speech closes. Newborn hearing screening (UNHS) is its front end: a physiological test on every baby, ideally before discharge, because permanent bilateral hearing loss affects roughly 1-2 per 1,000 live births (4-6 per 1,000 for any degree, including unilateral) and 2-4% of NICU graduates — estimates vary by study and by screening coverage. In India it is delivered through the National Programme for Prevention and Control of Deafness (NPPCD) and Rashtriya Bal Swasthya Karyakram (RBSK), with District Early Intervention Centres (DEICs) for follow-up, and habilitation funded under ADIP. It is examined in every BASLP/MASLP audiology paper, in RCI licence-level vivas, and asked in interviews at AIISH Mysuru, AYJNISHD Mumbai, NISH Thiruvananthapuram, medical-college ENT departments, and by hearing-aid and cochlear-implant companies.

Explained

Why infants need a separate audiology

• The auditory system is plastic but time-limited: unaided severe-profound loss in the first 2-3 years costs speech, language, literacy and later earnings; outcomes fall sharply with each month of delay in amplification.
• A baby cannot raise a hand to a tone, so testing moves from voluntary to physiological (OAE, ABR, ASSR, immittance) and observational/conditioned (BOA, VRA, play audiometry) methods.
• Anatomy and physiology differ: small, highly compliant ear canals, immature myelination (longer ABR latencies), middle-ear effusion and mesenchyme in the first weeks, vernix and debris causing screening refers.
• Cross-check principle: no management decision rests on one test; threshold estimates from ABR/ASSR must agree with OAE, immittance, behavioural responses and parental report.
• Developmental milestones as a screening net: startle/auropalpebral reflex and arousal to loud sound at birth; quietening to mother's voice by 2-3 months; head turn/localisation by 4-6 months; canonical babbling 6-9 months; first words by 12 months; about 50 words and two-word phrases by 24 months. Absent babbling or regression of babbling is a red flag.

The screening programme: targets, tools and protocols

• JCIH 1-3-6 benchmark: screen by 1 month of age, diagnostic confirmation by 3 months, intervention (amplification/early intervention) by 6 months. JCIH 2019 sets a more ambitious 1-2-3 goal and adds cochlear implant candidacy evaluation by 9-12 months where indicated.
• Targeted condition: permanent congenital hearing loss, conventionally bilateral loss ≥30-35 dB HL; modern programmes also aim to catch unilateral and mild losses, and neural (ANSD) losses.
• TEOAE (transient evoked OAE): clicks at about 80-83 dB peSPL; pass when the reproducibility and signal-to-noise ratio criteria are met (commonly SNR ≥6 dB in at least 3 of 5 frequency bands). Fast, cheap, no electrodes, done in natural sleep.
• DPOAE: two tones f1, f2 with f2/f1 ≈ 1.22, levels typically L1/L2 = 65/55 dB SPL; the 2f1−f2 distortion product is measured. Slightly better in noise and at higher frequencies.
• AABR (automated ABR): clicks at 35 dB nHL, three surface electrodes, automated statistical pass/refer. Detects auditory neuropathy spectrum disorder and retrocochlear loss, which OAE misses.
• Standard protocols:
  1. Well babies: OAE → if refer, repeat OAE (same visit/before discharge) → if still refer, AABR or repeat screen within 2-4 weeks; a second refer means diagnostic referral, not a diagnosis.
  2. NICU stay ≥5 days: AABR is mandatory as the screening test (ANSD risk from hyperbilirubinaemia, hypoxia, prematurity); if a NICU baby refers, go straight to diagnostic assessment, not to rescreening.
• Programme quality indicators: coverage ≥95% of live births; refer rate ≤4% (≤2-3% in mature programmes); loss to follow-up <5%; false-positive rate low enough to keep parental anxiety and cost acceptable.
• Screening in India: NPPCD (under the National Health Mission) places audiology equipment and manpower at district hospitals and CHCs; RBSK screens the "4 Ds" — defects at birth, deficiencies, diseases, developmental delays including disability — from birth to 18 years through mobile health teams, with referral to DEICs at district level. Coverage is uneven; most institutional screening is in tertiary and private hospitals, and the main failure is loss to follow-up, not a lack of screening tools.

Risk indicators and the high-risk register

A baby who passes screening can still develop late-onset, progressive or acquired loss, so risk-based surveillance continues (audiological review at least once by 9-12 months, earlier and repeatedly for CMV, ECMO or syndromic cases).

• Family history of permanent childhood hearing loss (first-degree).
• NICU >5 days; assisted ventilation; ECMO; hypoxic-ischaemic encephalopathy.
• Hyperbilirubinaemia needing exchange transfusion.
• Ototoxic drugs — aminoglycosides (gentamicin, amikacin), loop diuretics, platinum chemotherapy.
• In-utero infections — congenital CMV (the commonest non-genetic cause, often progressive), rubella, syphilis, toxoplasmosis, herpes, Zika.
• Craniofacial anomalies — microtia, atresia, ear pits/tags, cleft palate, dysmorphic syndromes.
• Syndromes with hearing loss — Waardenburg, Usher, Pendred, Jervell and Lange-Nielsen, Alport, Treacher Collins, branchio-oto-renal, Down syndrome, CHARGE; non-syndromic GJB2 (connexin 26) is the commonest genetic cause.
• Postnatal bacterial meningitis (urgent — risk of cochlear ossification), head trauma, measles, mumps, chemotherapy, neurodegenerative disorders.
• Caregiver concern about hearing, speech, language or developmental delay — a risk indicator in its own right and never to be dismissed.
• Prevention is part of the job: rubella (MR) immunisation, avoiding consanguinity-related risk counselling, safe delivery, rational antibiotic use, neonatal jaundice management, and otitis media control.

Diagnostic assessment of the infant who refers

• Order of testing (sleeping infant): immittance → OAE → ABR threshold search (click and tone-burst at 500, 1000, 2000, 4000 Hz, air and bone) → ASSR if needed → behavioural confirmation when age permits.
• Tone-burst ABR gives frequency-specific estimates; results are in dB nHL, converted to estimated behavioural level (dB eHL) by subtracting published correction factors before any hearing-aid fitting.
• Bone-conduction ABR separates conductive from sensorineural loss when air-conduction thresholds are raised.
• ASSR (auditory steady-state response) gives objective, statistically detected, frequency-specific thresholds and is useful in severe-profound loss where ABR has no residual response; it tends to overestimate thresholds in normal ears.
• Immittance in infants under about 6-9 months must use a 1000 Hz probe tone — a 226 Hz probe gives falsely normal tympanograms because the highly compliant neonatal canal wall dominates the measurement. Wideband acoustic immittance/absorbance is increasingly used.
• Acoustic reflex and wideband absorbance help confirm middle-ear status; ANSD typically shows present OAE, absent/grossly abnormal ABR with cochlear microphonic, and absent reflexes.
• Medical and genetic workup: ENT examination, paediatric/developmental assessment, ophthalmology (Usher, CMV), CMV PCR on saliva/urine in the first 3 weeks, ECG if Jervell and Lange-Nielsen suspected, renal imaging for BOR/Alport, genetic counselling, and CT/MRI of temporal bones when implant or anomaly is in question.
• Behavioural audiometry by age:
  • BOA (behavioural observation audiometry), 0-5 months — unconditioned responses (startle, eye-widening, cessation of sucking, arousal); gives *response levels*, not thresholds; high examiner bias, so use two observers.
  • VRA (visual reinforcement audiometry), 6 months-2.5 years — conditioned head-turn rewarded by a lighted/animated toy; insert phones preferred; gives near-adult thresholds for a cooperative child.
  • Conditioned play audiometry, ~2.5-5 years — put a peg in a board when you hear the sound.
  • Speech measures — speech awareness/detection threshold, speech recognition threshold with picture or body-part pointing, phonetically balanced kindergarten word lists; Ling six sounds (/m, ɑ, u, i, ʃ, s/) for quick functional checks of aided audibility.

From diagnosis to intervention

• Amplification within weeks of confirmation: behind-the-ear aids with soft paediatric earmoulds remade frequently, DSL v5 (child) prescription, RECD measured (not averaged) because small ear canals raise real-ear level, verified by probe-microphone measurement to targets, output limited by measured thresholds, and tamper-resistant battery doors.
• Bone-conduction devices on a softband for bilateral atresia/microtia or persistent conductive loss; FM/remote-microphone systems for classroom listening.
• Cochlear implantation is considered when well-fitted aids do not give adequate access, usually from 6-12 months in severe-profound bilateral loss, urgently after meningitis.
• Early intervention: auditory-verbal therapy, parent-infant guidance, Indian sign language where chosen, and a family-centred, informed-choice approach; early intervention and school support run through DEICs and Samagra Shiksha inclusive education.
• Indian legal and scheme framework:
  • RPwD Act, 2016 — hearing impairment is a specified disability; "deaf" is defined as 70 dB and above hearing loss in speech frequencies in both ears, "hard of hearing" as 60-70 dB in speech frequencies in both ears; 40% or more is a *benchmark disability* that unlocks reservation, concessions and scheme benefits.
  • Disability certification follows the Ministry of Social Justice and Empowerment's 2018 Gazette guidelines, which compute percentage hearing disability from the pure-tone average of 500, 1000, 2000 and 4000 Hz with a tabulated scale and separate provisions for children and for speech-language disability; the exact percentage bands should be read from the current notification rather than quoted from memory.
  • RCI Act, 1992 — an audiologist/speech-language pathologist must hold a valid Rehabilitation Council of India registration (CRR number) with periodic renewal through continuing rehabilitation education; practising without it is an offence.
  • ADIP scheme (Ministry of Social Justice and Empowerment) funds hearing aids and paediatric cochlear implants for eligible low-income children, with an age limit and a per-implant cost ceiling (commonly cited as about ₹6 lakh); eligibility, age and ceiling are revised from time to time and must be checked against the current scheme guidelines. State schemes (for example in Kerala, Tamil Nadu and Maharashtra) and Ayushman Bharat packages also fund implants.
  • Equipment calibration follows the IEC 60645 series for audiometers and the ISO 389 series for reference equivalent threshold levels, with permissible ambient noise levels per ISO 8253; OAE/ABR probe and transducer calibration must be current.

Advantages, limitations and comparisons

• OAE screening: cheap, quick (1-2 minutes), no electrodes, high sensitivity to cochlear loss; but high refer rate in the first 24-48 hours from vernix/debris and transient middle-ear fluid, and it cannot detect ANSD or mild loss.
• AABR screening: lower false-positive rate after the first day, detects neural loss; but costlier (disposable electrodes), slower, needs a quiet, settled baby, and misses purely cochlear high-frequency losses outside the click band.
• Two-stage (OAE then AABR) protocols cut false positives and overall cost while keeping sensitivity — the usual compromise for well babies.
• Screening is not diagnosis: "refer" means "not yet cleared". "Pass" does not rule out late-onset, progressive or mild loss, nor mild conductive loss.
• Universal vs high-risk (targeted) screening: risk-factor-only screening misses about half of congenitally hearing-impaired babies, since many have no identifiable risk indicator — this is the standard exam justification for *universal* screening.

Key points

• UNHS - screening every newborn with a physiological test, not just high-risk babies; targeted screening alone misses roughly 50% of cases.
• 1-3-6 - screen by 1 month, diagnose by 3 months, intervene by 6 months (JCIH 2019 prefers 1-2-3).
• Prevalence - about 1-2/1,000 permanent bilateral congenital loss; 4-6/1,000 for any degree; 2-4% in NICU graduates.
• NICU ≥5 days - screen with AABR, never OAE alone, because of ANSD risk.
• ANSD - present OAE with absent/abnormal ABR and a cochlear microphonic; OAE-only screening misses it.
• Refer, not fail - screening outcome language matters for parental counselling; a refer rate above about 4% signals a protocol or technique problem.
• 1000 Hz probe tone - mandatory for tympanometry under about 6-9 months; 226 Hz gives false normals.
• Tone-burst ABR in dB nHL - convert to dB eHL with correction factors before fitting hearing aids.
• ASSR - objective frequency-specific thresholds, especially useful in severe-profound loss; overestimates in normal ears.
• BOA / VRA / CPA - 0-5 months / 6-30 months / 2.5-5 years; BOA gives response levels, not thresholds.
• RECD must be measured in children - small canals raise real-ear SPL and change prescribed gain and output limits.
• Congenital CMV - commonest non-genetic cause; progressive, so repeated audiological surveillance is required even after a pass.
• GJB2 (connexin 26) - commonest genetic non-syndromic cause; genetic counselling is part of management.
• Meningitis - urgent audiological assessment; risk of cochlear ossification narrows the implant window.
• RPwD Act 2016 - "deaf" ≥70 dB both ears, "hard of hearing" 60-70 dB; 40% is benchmark disability.
• RCI registration - legally required to practise as an audiologist/SLP in India.
• Loss to follow-up - the dominant weakness of Indian screening; target <5%, achieved by appointment-before-discharge, phone tracking and DEIC linkage.
• Milestones screen - no babbling by 9 months, no words by 15 months, or caregiver concern at any age warrants full assessment.

Formulas and examples

Example 1 - two-stage programme yield, refer rate and PPV. A district hospital has 6,000 live births a year. Stage-1 TEOAE before discharge refers 8% → 0.08 × 6,000 = 480 babies. Repeat OAE/AABR at 3-4 weeks clears 75% of them, so 0.25 × 480 = 120 go for diagnostic ABR. Diagnosis confirms permanent bilateral loss in 24.

• Prevalence (yield) = 24 / 6,000 = 0.4% = 4 per 1,000.
• Final refer rate after stage 2 = 120 / 6,000 = 2.0% (acceptable; stage-1 8% is high and points to early testing/vernix).
• PPV of a stage-2 refer = 24 / 120 = 20% — four out of five referred babies turn out normal, which is why counselling must stress "needs a second look", not "has hearing loss".

Example 2 - sensitivity and specificity of AABR in a NICU cohort. 1,000 NICU graduates screened with AABR and all followed to diagnostic ABR: TP = 38, FN = 2, FP = 30, TN = 930 (total 1,000; true cases 40, i.e. 4%).

• Sensitivity = 38/40 = 95.0%; Specificity = 930/960 = 96.9%.
• PPV = 38/(38+30) = 38/68 = 55.9%; NPV = 930/(930+2) = 930/932 = 99.8%.
• Reading: a pass is very reassuring (NPV 99.8%), a refer is barely better than a coin toss — so a refer triggers assessment, never a diagnosis.

Example 3 - corrected age before interpreting ABR. A baby born at 30 weeks gestation is brought at 8 weeks of life. Corrected age = 8 − (40 − 30) = 8 − 10 = −2 weeks, i.e. 38 weeks post-conceptional age — still "pre-term equivalent". Wave V latency must be read against neonatal, not infant, norms (wave V is longer and amplitude smaller), and a "delayed" latency here may be pure immaturity. Rescreen/retest nearer term-equivalent rather than diagnosing conductive delay.

Example 4 - DPOAE stimulus arithmetic. For f2 = 4,000 Hz with f2/f1 = 1.22: f1 = 4,000 / 1.22 = 3,278.7 Hz. The distortion product frequency 2f1 − f2 = (2 × 3,278.7) − 4,000 = 6,557.4 − 4,000 = 2,557.4 Hz. The response is *recorded* at ≈2,557 Hz but *reflects* cochlear function near f2 = 4,000 Hz — a standard viva trap.

Example 5 - cost per case detected (rupees). Using Example 1: 6,000 OAE screens at ₹120 each (consumables, probe tips, staff time) = ₹7,20,000; 120 diagnostic ABR sessions at ₹1,500 = ₹1,80,000; total ₹9,00,000 for 24 confirmed cases → ₹37,500 per case detected. Against a lifetime of special education and lost productivity, this is the standard cost-effectiveness argument for UNHS in Indian public hospitals.

Example 6 - disability percentage input. Right ear 70/75/85/95 dB at 0.5/1/2/4 kHz → PTA = 325/4 = 81.25 dB. Left ear 60/65/70/80 → PTA = 275/4 = 68.75 dB. Better ear = 68.75 dB, which lies within the 60-70 dB range the RPwD Act uses for "hard of hearing" — but note the Act requires that band in both ears, and here the right ear is 81.25 dB, so the categorical label does not follow from the better ear alone; the exact disability percentage is read off the 2018 Gazette table by the certifying medical board, not computed by the audiologist alone.

Where it is used

• Hospital newborn screening services — maternity and NICU units of medical colleges, district hospitals under NPPCD, and private chains (Apollo, Fortis, Cloudnine) where OAE before discharge is routine.
• RBSK mobile health teams and District Early Intervention Centres — audiologists, audiometric assistants and early-intervention staff doing screening, diagnosis, hearing-aid fitting and therapy for 0-18 years.
• National institutes and tertiary centres — AIISH Mysuru, AYJNISHD Mumbai, NISH Thiruvananthapuram, PGIMER, AIIMS, CMC Vellore: paediatric diagnostics, implant programmes, teaching and RCI-recognised training.
• Cochlear implant programmes — candidacy evaluation, ADIP/state-scheme paperwork, switch-on and mapping support, outcome tracking of implanted children.
• Hearing-aid and equipment industry — paediatric fitting support, probe-microphone verification, and sales/application roles with Otometrics/Natus, Interacoustics, Maico, Phonak, Oticon, Widex, Signia, Cochlear, MED-EL, Advanced Bionics and Indian distributors.
• Schools, NGOs and medico-legal work — special and inclusive schools, early-intervention NGOs, disability certification boards, and insurance/forensic assessments of childhood hearing loss.

Interview questions

• Q: Why is universal screening preferred over high-risk screening? A: About half of babies with congenital permanent hearing loss have no identifiable risk indicator, so a risk-register-only approach misses roughly 50% of cases.
• Q: What does 1-3-6 mean? A: Screening by 1 month of age, diagnostic confirmation by 3 months, and intervention such as hearing aids and early intervention by 6 months; JCIH 2019 targets 1-2-3.
• Q: Why must NICU babies be screened with AABR rather than OAE? A: Risk factors like hyperbilirubinaemia, hypoxia and prematurity cause auditory neuropathy spectrum disorder, where OAEs are present but the ABR is absent — OAE screening would pass these babies.
• Q: A baby passes OAE but the parents say she does not respond to sound. What do you do? A: Take the caregiver concern as a risk indicator and do a full diagnostic workup including ABR, because OAE cannot rule out neural loss, mild loss or late-onset progressive loss.
• Q: Why use a 1000 Hz probe tone in infants? A: The neonatal ear canal wall is highly compliant and dominates the 226 Hz measurement, producing falsely normal tympanograms; 1000 Hz correctly identifies middle-ear effusion in babies under about 6-9 months.
• Q: Why is the refer rate high if you screen in the first few hours? A: Vernix and debris in the canal and residual middle-ear fluid block the OAE response, so screening before 24-48 hours, or before discharge with a repeat, inflates false positives.
• Q: Difference between dB nHL and dB eHL? A: nHL is the stimulus level referenced to normal-hearing adults' behavioural thresholds for that brief stimulus; eHL is the estimated behavioural threshold after applying stimulus- and age-specific correction factors, and only eHL is used for hearing-aid prescription.
• Q: Why is PPV of newborn hearing screening low even with good sensitivity and specificity? A: Because the condition is rare — at a prevalence of a few per thousand, the small number of true positives is swamped by false positives, so most refers are normal on diagnosis.
• Q: How do you estimate thresholds in a 4-month-old? A: Tone-burst ABR at 500-4000 Hz by air and bone in natural sleep, cross-checked with OAE, 1000 Hz tympanometry and observation; ASSR if residual hearing is minimal. BOA at this age gives response levels, not thresholds.
• Q: What is the significance of congenital CMV? A: It is the commonest non-genetic cause of congenital hearing loss, can be present with a passed newborn screen, and is often progressive — so these children need scheduled audiological surveillance.
• Q: What does the RPwD Act 2016 say about hearing disability? A: It lists hearing impairment as a specified disability, defines "deaf" as 70 dB and above loss in speech frequencies in both ears and "hard of hearing" as 60-70 dB, and sets 40% as the benchmark disability threshold for entitlements.
• Q: Which scheme funds paediatric cochlear implants for poor families in India? A: ADIP under the Ministry of Social Justice and Empowerment, with an age limit and a per-implant cost ceiling, supplemented by some state schemes; the figures are revised periodically and must be read from current guidelines.
• Q: Why should a child's RECD be measured and not taken from an age average? A: Individual ear canal volume varies widely; using an average can under- or over-estimate real-ear SPL by many decibels, causing under-amplification or over-exposure.

Common mistakes

• Treating a screening "refer" as a diagnosis of hearing loss, or a "pass" as lifelong clearance from late-onset and progressive loss.
• Using OAE alone for NICU babies, thereby missing auditory neuropathy spectrum disorder.
• Using a 226 Hz probe tone for tympanometry in neonates and reporting a normal middle ear.
• Fitting hearing aids straight from dB nHL ABR levels without applying nHL-to-eHL corrections, or using average rather than measured RECD.
• Reporting BOA "minimum response levels" as thresholds, and testing with one observer so examiner bias goes unchecked.
• Forgetting to correct for prematurity before calling an ABR latency delayed or a response absent.
• Not fixing the next appointment before discharge — the single biggest cause of loss to follow-up, which makes the whole screen worthless.
• Quoting disability percentages, scheme ceilings or age limits from memory instead of the current Gazette notification or scheme guideline.

Related topics

Diagnostic audiology: audiometry, immittance, OAE and ABR · Hearing aids: selection and fitting · Cochlear implants and aural rehabilitation · Speech and language development and delay · Audiology ethics, RCI practice and disability legislation · Public health audiology and hearing conservation