Development of a Multidimensional, Multigroup Measure of Cognitive-Communication for Inpatient Rehabilitation.

Abstract

To develop a comprehensive, repeatable measure of cognitive-linguistic function for use in inpatient rehabilitation facility (IRF) settings using contemporary methods and to compare the sensitivity of this measure to the functional independence measure (FIM) cognitive items. Retrospective analysis of clinician-reported assessment data collected as part of routine clinical operations. Free-standing IRF in the Midwestern United States. A convenience sample of inpatients (N=12,245) with motor speech, voice, aphasia and/or cognitive-communicative diagnoses, ≥18 years who were admitted to an IRF. Standard of care rehabilitation services. Cognitive-Communication Ability Quotient (AQ) composed of 10 standard measures used by speech-language pathologists. Confirmatory factor analysis resulted in good-fitting models (root-mean-square errors of approximation≤0.08, comparative fit indices, and nonnormed fit indices≥0.95) for 5 groups defined by primary communication impairment (Aphasia, Cognitive-Communication Disorder, Brain Injury, Right Hemisphere Dysfunction, Motor Speech/Voice). Re-estimation as a multigroup, MIRT model yielded scores more sensitive to change compared to the FIM cognitive score. True score equating analysis demonstrated a higher ceiling and lower floor for the Cognitive-Communication AQ compared to the FIM. We constructed 5 forms of a novel cognitive-communication measure, the Cognitive-Communication AQ, which demonstrates superior measurement characteristics compared with the FIM cognition score with a lower floor and higher ceiling. The AQ can detect changes in cognitive-communicative function that typically occur during an IRF stay. Repeated measurement during the IRF stay allows clinicians to monitor patients' progress and modify rehabilitation plans accordingly.