Psychometric Properties of 2 Participation Measures in Veterans With Mild Traumatic Brain Injury.

Abstract

To compare the psychometric properties of 2 commonly used participation measures: the Community Reintegration of Service Members (CRIS) and the Participation Assessment with Recombined Tools-Objective (PART-O) in veterans with mild traumatic brain injury (mTBI). Data were collected from 2 cross-sectional observation studies conducted in 2 Veterans Affairs medical centers. Questionnaires were completed in-person or by mail. Veterans with mTBI (N=201) were recruited from the Michael E. DeBakey Veterans Affairs Medical Center in Houston (n=94) and the Malcom Randall North Florida/South Georgia Veterans Health System (n=107). Not applicable. CRIS and PART-O. We conducted Rasch analysis on the PART-O and on 3 subscales of the CRIS (extent of participation, perceived limitation, and satisfaction). For PART-O, results showed PART-O has questionable unidimensionality. For both instruments, some rating categories were underused, and rating scales did not advance accordingly. Compared with PART-O, the CRIS was able to distinguish more categories of person's ability (>5 vs 2 for PART-O) and had better internal consistency as indicated by higher Cronbach α (.96-.98 vs .65 for PART-O). To capture participation unique to veterans with mTBI, CRIS has greater potential to detect a change in participation and is therefore recommended over PART-O. Rating scales of both instruments, however, need further refinement. We suggest future studies examine collapsed rating categories and use qualitative methods to redefine categories.