The Clinical Evaluation of a Modified Fmp in Infant Auditory Brainstem Response Tests.

Abstract

The performance of a modified calculation of F statistic using multiple points (Fmp) in auditory brainstem response (ABR) tests in infants was evaluated. Using UK national guidelines, the ABR threshold was established on 50 infant ears at 4 kHz and 41 infant ears at 1 kHz. A specificity-based Fmp criterion for response presence was established from the distribution of no-response values. This criterion was then applied to determine the sensitivity of Fmp in detecting responses. A 97.5% true negative rate in no-response waveforms corresponded to an Fmp of 2.2. This criterion detected 85% of 4 kHz and 68% of 1 kHz responses at 10 dB above the ABR threshold but only 51% of 4 kHz and 32% of 1 kHz responses at the ABR threshold. Fmp has reasonable clinical utility at stimulus levels above the ABR threshold but is not an adequate replacement for strictly applied conventional waveform interpretation at the ABR threshold. A proposal is offered that should improve Fmp sensitivity.