Discourse-based assessment in clinical context for persons with Alzheimer's disease: A scoping review
Guillaume Duboisdindien, Nancy Durieux, Mireille Kieu, Julie CattiniBackground
Discourse assessment captures communicative abilities integrating linguistic, cognitive, and pragmatic processes and is widely used to characterize communication changes in Alzheimer's disease (AD).
Objective
To identify and map clinically applicable methods and linguistic measures for assessing discourse changes in persons with AD.
Methods
This scoping review followed JBI methodology and PRISMA-ScR guidelines. Observational studies assessing spoken discourse in persons with AD compared with normocognitive controls were identified through systematic searches of MEDLINE, PsycINFO, Linguistics and Language Behavior Abstracts, and Scopus. Studies based exclusively on automated language-processing approaches were excluded to maintain a focus on clinically applicable assessment. English or French-language studies were included. To assess inter-rater reliability, data from 17/76 studies (22.4%) in the broader review were independently extracted by two reviewers, with high agreement; remaining studies were extracted by one reviewer, with uncertainties resolved collectively. Extracted data included elicitation tasks, coding procedures, linguistic domains, and outcome measures and were summarized descriptively.
Results
Fifty-nine papers were included. Picture description predominated, followed by conversational and narrative formats. Across papers, 291 distinct linguistic measures were identified. Discourse/pragmatic measures constituted the largest domain and had the greatest number of reported statistically significant findings. Fluency measures were also frequently associated with statistically significant differences, whereas morphosyntactic measures showed fewer such findings.
Conclusions
Substantial variability in elicitation procedures, operational definitions, and linguistic measures limits cross-study comparability. Methodological harmonization and further evaluation of the psychometric and clinical properties of discourse measures are needed to support evidence-based clinical assessment.