The reliability and validity of the brief symptom inventory-18 in persons with traumatic brain injury.

OBJECTIVE: To investigate the psychometric properties of the Brief Symptom Inventory-18 (BSI-18) among persons with traumatic brain injury (TBI).
DESIGN: Inception cohort design with cross-sectional follow-up of 6 months to 15 years.
SETTING: Rehabilitation hospital. PARTICIPANTS: Adults (N=257) with moderate to severe TBI (81 inpatients and 176 follow-up participants, analyzed separately).
INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: The BSI-18 is a brief screen of psychologic distress with a Global Severity Index (GSI), and 3 clinical subscales: somatization, anxiety, and depression.
RESULTS: Internal consistency of the GSI was high in both follow-up participants (alpha=.91) and inpatients (alpha=.84), whereas estimates for the somatization, anxiety, and depression subscales were more variable (alpha range, .61-.84). As would be expected for a measure of affective state, retest reliability estimates were only moderate. The BSI-18 GSI correlated with multiple measures of psychosocial adjustment. After accounting for demographics, injury severity, inpatient functional status, years since injury, and various psychosocial factors, the BSI-18 showed incremental validity in predicting concurrent functional, psychosocial, and psychologic status.
CONCLUSIONS: The BSI-18 GSI had excellent reliability and validity among inpatients and follow-up participants. Modest reliability estimates may place an upper bound on the validity of the BSI-18 clinical subscales in inpatient TBI populations.