Individual variability on discourse measures over repeated sampling times in persons with aphasia

Background: Although persons with aphasia typically have difficulty with the production of language at the level of discourse, there is a paucity of reliable measurement systems to quantify the characteristics of spoken language. Nicholas and Brookshire (1993) developed one of the few standardised, rule-based systems to quantify the informativeness of spoken language samples. While the authors reported temporal stability for all measures, they also noted variability at the individual level. Because individual data were not reported, it is difficult to determine the nature and extent of that variability. Aims: The aim of the current investigation was to further explore Nicholas and Brookshire's (1993) quantitative linguistic analyses, and to examine individual variability over time in persons with aphasia. Methods Procedures: Five fluent and six nonfluent persons with aphasia produced language samples over repeated sampling times in response to the 10 stimulus items used by Nicholas and Brookshire (1993). Measurements of mean number of words, mean correct information units (CIUs), percent CIUs, words per minute, and CIUs per minute were calculated, and results from the sessions were compared. To examine factors related to individual variability, correlations between linguistic measures, months post onset of aphasia, and scores on standardised assessment tools were explored. Outcomes Results: Visual inspection of the data and descriptive statistics suggested that participants were more variable in their repeated productions than previously described by Nicholas and Brookshire (1993). Repeated measures ANOVAs revealed non-significant effects at the group level. There was no pattern of variability uniquely associated with aphasia fluency type. Range of mean number of CIUs was positively and significantly correlated with all test measures; however, the other range correlations were non-significant. Conclusions: The current results suggest greater variability over repeated sampling times in the spoken language of persons with aphasia than previously reported by Nicholas and Brookshire (1993). Clinicians and researchers should consider this variability, and establish stable baselines prior to the initiation of treatment to document meaningful change over time.