OBJECTIVE: To investigate the neurobehavioral pattern of recovery of consciousness as reflected by performance on the subscales of the Coma Recovery Scale - Revised (CRS-R).
DESIGN: Retrospective item response theory and factor analysis. SETTING: Inpatient rehabilitation facilities in the United States and Europe. PARTICIPANTS: Rehabilitation inpatients (N=180) with post-traumatic disturbance in consciousness who participated in a double-blinded randomized controlled drug trial INTERVENTIONS: None MAIN OUTCOME MEASURES: Coma Recovery Scale- Revised (CRS-R) subscales scores.
RESULTS: The CRS-R was found to fit factor analytic models adhering to the assumptions of unidimensionality and monotonicity. In addition, subscales were mutually independent based on residual correlations. Non-parametric IRT reaffirmed the finding of monotonicity. A highly constrained confirmatory factor analysis model, which imposed equal factor loadings on all items was found to fit the data well and used to estimate one parameter logistic item response theory parameters.
CONCLUSION: This study provides evidence of the unidimensionality of the CRS-R and supports the hierarchical structure of the CRS-R subscales, suggesting that it is an effective tool for establishing diagnosis and monitoring recovery of consciousness following severe traumatic brain injury (TBI).