Using cluster analysis to classify audiogram shapes

AbstractThe purpose of this study was to design a statistical classification system of audiogram shapes in order to improve and integrate shape recognition across clinical settings. The study included 1633 adult subjects with normal hearing or symmetric sensorineural hearing impairment who underwent pure-tone audiometry between July 2007 and December 2008. K-means cluster analysis was employed to categorize audiometric shapes. Eleven audiogram shapes were identified: rising, flat, peaked 8-kHz dip, 4-kHz dip, 8-kHz dip, mild sloping, severe 8-kHz dip, sloping, abrupt loss, severe sloping, and profound abrupt loss. By using the classification system and nomenclature identified for audiogram shapes as outlined in this study, errors based on personal experiences can be reduced and a consistency can be developed across clinics.SumarioEl proposito de este estudio fue disenar un sistema de clasificacion estadistica de formas de audiogramas para mejorar e integrar el reconocimiento de formas en los diferentes co...