The main concept analysis: Sensitivity, specificity and inter-rater reliability in Dutch-speaking people with aphasia

This study aimed to assess the sensitivity, specificity and inter-rater reliability of the Dutch Main Concept Analysis (MCA), a diagnostic tool that evaluates the effectiveness and efficiency of information transfer during verbal discourse production, in people with aphasia (PWA).

The MCA was administered to 30 persons with aphasia (17 fluent, 13 nonfluent). Sensitivity and specificity were assessed through (1) single-subject level comparison of MCA scores to normative data, (2) group-level comparison of MCA scores between PWA and 30 age-matched controls, and (3) ROC-analysis for total main concept score (MCS; effectiveness) and total number of accurate and complete main concepts per minute (AC/min; efficiency). Furthermore, the effect of aphasia type (fluent/nonfluent) and severity on these measures was assessed using a linear regression. Discourse samples were scored by two independent raters to calculate inter-rater reliability.

Over 85 % of PWA showed aberrant outcomes on at least four out of seven MCA outcomes based on single-subject comparison to normative data. ROC analysis revealed that MCS and AC/min are sensitive and specific measures to differentiate between discourse of PWA and unimpaired speakers. Increased aphasia severity was associated with a reduced MCS. Good to excellent inter-rater reliability was achieved for all MCA measures.

The Dutch MCA is a sensitive and specific tool, characterized by a good to excellent inter-rater reliability, for identifying impaired effectiveness and efficiency of verbal information transfer during discourse production in PWA. Aphasia severity, but not type, is a determining factor for the effectiveness of information transfer in PWA.