Objective: To study interrater reliability and factorial and criterion validity of the Neurobehavioral Rating Scale-Revised (NRS-R).
Design: Validity study on persons with traumatic brain injury (TBI) and test-retest reliability study on a randomly selected subset of patients. Factor analyses, kappa statistics, intraclass correlation coefficients, and Cronbach's alphas were used.
Setting: Inpatients from 15 French hospitals, mainly rehabilitation units. Other recruitment sites included a neurology hospital unit and a psychiatry hospital specifically devoted to TBI rehabilitation.
Patients: Two hundred eighty-six TBI patients ages 16 to 70 years (convenience sample).
Results: For the reliability study, the average of percentages of agreement among the items was 74.3% and the average of kappa statistics was .40. Factor analyses disclosed a maximum likelihood extraction of 5 correlated factors (F), explaining 42.2% of total variance: (F1) deficits in intentional behavior and in memory, (F2) lowering of emotional state, (F3) emotional and behavioral hyperactivation, (F4) lowering of arousal state and of attention, and (F5) language and speech problems. Results support the criterion validity of the factors. Reliability of the factor scores and internal consistencies of factors were very good.
Conclusions: Results describe some important properties of the NRS-R and, through an understanding of its underlying structure and relationships with the patients' clinical characteristics, contribute to the conceptual framework of neuropsychologic impairments after TBI.