Five Years of Experience with Electric Response Audiometry

Abstract

The results from electric response audiometry (ERA) tests of 83 persons, including experimental subjects (those whose hearing thresholds were known) and clinical subjects between the ages of one month and 70 years, and of 40 newborns, were reviewed and scored for thresholds by three independent and sophisticated readers. Test stimuli were delivered in 10-dB steps. In the first group, the readers agreed on only 36%, and two readers agreed 43% of the time on the same 10-dB interval. When two agreed and the third did not, the average disagreement was 30 dB. When all three readers diverged, the average spread was 43 dB. There appeared to be no difFerence in three-reader accuracy between known normals and clinical cases. ERA agreement with audiograms obtained later was 56% (within ± 15 dB). In the newborns, three-reader agreement was much worse, and only 20% of the subjects yielded thresholds at or below 70 dB. It would appear, therefore, that ERA is not as reliable a tool, at least in our experience, as had been supposed. ERA should be considered as only part of diagnostic audiometry, and clinical judgments made solely on the basis of ERA should be strictly avoided.