The Dysphonia Severity Index ( DSI )

Basics The Dysphonia Severity Index (DSI) is designed as an objective and quantitative correlate of voice quality. It is based on a multivariate analysis of an extended database of more than 1000 normal and pathologic voices for the multi-center study trial initiated by the Belgian Study Group on Voice Disorders (Van de Heyning, et al., 1996). Per patient, more than 45 voice characteristics and measurements were collected. Additionally, the voice quality as perceived by the clinician and expressed by means of the grade of hoarseness was obtained for all subjects (G from the GRBAS scale: normal, light, moderate or severe: Hirano, 1981). The Dysphonia Severity Index was then constructed by a multivariate statistical approach that calculated a combination of weights variables such that the different types of voices according to the perceived voice quality (G) were optimally identified. This resulted in only four voice characteristics: maximum phonation time (MPT, sec.) highest frequency (Fo-high, Hz), lowest intensity (I-low, dB(A)) and jitter (%). The calculated weights are given in the following expression: