Air-conduction estimated from tympanometry (ACET): 2. The use of hearing level-ACET discrepancy (HAD) to determine appropriate use of bone-conduction tests in identifying permanent and mixed impairments

Summary Objective The caseload at secondary care in paediatric otology is largely otitis media with effusion (OME) and highly recurrent acute otitis media (RAOM). Few of these cases merit suspicion for hearing loss beyond the middle ear. The companion paper showed that the air conduction estimated from tympanometry (ACET) formula, derived on a very large clinical sample referred for ear or hearing problems and pre-assessed for a clinical trial, gives usable although only approximate estimates for hearing level (HL) on such a caseload. Tympanometry corresponds to a conductive loss (i.e. air–bone gap) so the HL–ACET discrepancy (HAD) should approximate the bone-conduction (BC) threshold. Clinical criteria might enable HAD to substitute for BC tests where those are infeasible, or to identify those most needing BC tests. Method ACET had been derived for the 4-frequency binaural average on 3085 cases with tympanometry and air-conduction HL available. On the 2701 of those with BC data at 1 kHz, we re-calculated ACET for 1 kHz only, and then explored the sensitivity/specificity trade of the discrepancy (HAD) in detecting clinically significant BC levels and the correlation between these measures. We further illustrated the generalization of the formula and cut-off on a small separate retrospective clinical sample. Results Correlations were moderate in the clinically relevant region. There were five cases of BC ≥ 30 dB in the database. At a HAD cut-off of +5 dB, the sift would identify all (nominal 100% sensitivity). For marginal cases, two definitions were adopted (BC ≥ 25 dB and ≥20 dB; 9 and 23 cases, respectively). Sift sensitivity remained high (89% and 83%, respectively), and specificity was acceptable (75% for both definitions). Conclusions Given tympanometry and air-conduction HL, comparison of HAD with a recommended cut-off gives acceptable sensitivity and specificity for non-OME hearing problems. BC testing can be reserved for probable positive cases, provisionally only 25% of caseload. HAD could temporarily substitute for BC measurement in children too young to accept bone-conduction transducers in awake testing. Where a high proportion of the caseload is expected to have middle ear fluid, ACET and HAD together offer efficient possibilities for best use of available information.