Diagnostic classification through content analysis of patients' speech.

Abstract

Speech samples from 71 patients in four diagnostic groups were analyzed by two quantitative methods of speech content analysis, the results of which were entered into a discriminant analysis to test whether patients could be accurately classified back into their appropriate diagnostic groups. These classifications were compared with classifications made by two psychiatrists, blind to the patients' diagnoses, who read transcripts of the speech samples. The results suggest that data from the systematic quantification of lexical choice can be used to classify patients into their respective diagnostic groups and that this classification compares favorably with that done by psychiatric raters.