Decision Making in Fluency Measures of Aphasic Spontaneous Speech

Kim, Hyanghee; Kwon, Miseon; LNa, Duk; Choi, Sang-sook; Lee, Kwang ho; Chung, Chin Sang

ORIGINAL ARTICLE

Commun Sci Disord. 1998;3(1): 5-19.

Hyanghee Kim^` , Miseon Kwon^` , Duk LNa^` , Sang-sook Choi^` , Kwang ho Lee^` , and Chin Sang Chung^`

ABSTRACT

Fluency judgements of an aphasic patient's spontaneous speech can be extremely subjective. A number of variables that might be related to the fluency produced by aphasic patients were investigated. Tape-recorded language samples of the picture description task of K-WAB were used for the analysis. A consensus judgement on the degree of fluency (i.e., hypo- fluent, normal-fluent, and hyper-fluent) was made by three speech-language pathologists. The picture description measures of the three groups (hypo-fluent = 16 patients; normal- fluent = 11; hyper-fluent=4) were analyzed using eight variables (i.e., the number of syllables per second or SPS the mean number of syllables per utterance or SPU; the mean number of words per utterance or WPU; the total number of utterances or TU; the total number of complex sentences/TU ratio; the total number of sentences /TU radio or TS/TU; the analysis of nine content-function words; and the per-centages of errors). We used the aforementioned eight variables as dependent variables and the three fluency groups as independent variables. The variables that categorize at least two patient groups were SPU, WPU, TU, and TS/TU. Based on a discriminant function analysis， the percentage of group cases correctly classified was 79.37%. This study reveals important variables that might contribute to improving the fluency judgement of language of patients with aphasia.