Puretone Threshold Estimation from Acoustic Reflex Thresholds—A Myth?

Abstract

There have been repeated claims in the literature that hearing thresholds can be estimated by means of acoustic reflex threshold (ART) measurements. We have evaluated this thesis in depth by examining the statistical relationships between conventional puretone thresholds and ARTs for tonal and broadband stimuli in a large sample (1 207) of adult patients with sensorineural hearing loss; they were compensation claimants for noise-induced loss. All relevant data were stored in a computer database, and analysed using the Statistical Package for the Social Sciences (SPSS). The relationships were examined from two distinct standpoints; dependency of ARTs upon puretone thresholds, and prediction of audiogram features using ARTs. The average tonal ART is constant up to 70 dB average loss, and thereafter increases rapidly. The broadband noise ART increases steadily for average losses of 10 to 85 dB. The convergence principle is therefore valid up to 70 dB in this population. The predictor suggested by Niemeyer & Sesterhenn does not fit the data. The single ART which best predicts average loss is that for white noise, but a multiple regression predicition using the white noise and 500 Hz ARTs is preferable. Predictive accuracy is dependent upon audiogram slope, but the latter cannot be predicted usefully from the ART measures used. None of the predictors studied is adequate for medico-legal assessment and the method is probably not sufficiently accurate for clinical use in adults.