Assessing Language in Unstructured Conversation in People With Aphasia: Methods, Psychometric Integrity, Normative Data, and Comparison to a Structured Narrative Task

• Published in: Journal of speech, language, and hearing research Vol. 64; no. 11; pp. 4344 - 4365
• Main Authors: Leaman, Marion C., Edmonds, Lisa A.
• Format: Journal Article
• Language: English
• Published: United States American Speech-Language-Hearing Association 01.11.2021
• Subjects: Allied Health Personnel; Aphasia; Coherence; Cohesion; Communication; Communication (Thought Transfer); Comparative Analysis; Correlation; Dialect Studies; Evaluation Methods; Grammar; Grammaticality; Humans; Integrity; Interpersonal Communication; Interrater Reliability; Intervention; Interviews; Language; Language assessment; Language tests; Language Variation; Linguistics; Measures (Individuals); Narration; Narratives; Normative data; Outcome Measures; Physiological aspects; Psychometrics; Quantitative psychology; Rating Scales; Reproducibility of Results; Social Networks; Speech Communication; Speech Evaluation; Speech Language Pathology; Speech production; Speech therapists; Speech Therapy; Speech-language pathologists; Statistical Analysis; Task Analysis; Test validity and reliability; Variance analysis; Verbal communication; Verbs; New York (New York); United States
• ISSN: 1092-4388; 1558-9102; 1558-9102
• DOI: 10.1044/2021_JSLHR-20-00641

Purpose: This study evaluated interrater reliability (IRR) and test-retest stability (TRTS) of seven linguistic measures (percent correct information units, relevance, subject-verb-[object], complete utterance, grammaticality, referential cohesion, global coherence), and communicative success in unstructured conversation and in a story narrative monologue (SNM) in persons with aphasia (PWAs) and matched participants without aphasia (M-PWoAs). Furthermore, the relationship of language in unstructured conversation and SNM was investigated for these measures. Methods: Twenty PWAs and 20 M-PWoAs participated in two unstructured conversations on different days with different speech-language pathologists trained as social conversation partners. An 8- to 12-min segment of each conversation was analyzed. Additionally, a wordless picture book was used to elicit an SNM sample at each visit. Correlational analyses were conducted to address the primary research questions. Normative range and minimal detectable change data were also calculated for the measures in both conditions. Results: IRR and TRTS were moderate to good for parametric measures and moderate to excellent for nonparametric measures for both groups, except for TRTS for referential cohesion for the PWAs in conversation. Furthermore, in PWAs, a strong correlation was demonstrated for three of eight measures across conditions. Moderate or weaker correlations were demonstrated for three of eight measures, and correlations for two of eight measures were not significant. An ancillary finding was no significant differences occurred for sample-to-sample variability between the two conditions for any measure. Conclusions: This study replicates previous research demonstrating the feasibility to reliably measure language in unstructured conversation in PWAs. Furthermore, this study provides preliminary evidence that language production varies for some measures between unstructured conversation and SNM, contributing to a literature base that demonstrates language variation between different types of monologue. Thus, these findings suggest that inclusion of the specific types of discourse of interest to the PWA may be important for comprehensive assessment of aphasia.