Rasch Analysis of the Behavioral Assessment Screening Tool for Adolescents with Mild Traumatic Brain Injury

Research Objectives: To provide psychometric evidence for the Behavioral Assessment Screening Tool for Adolescents (BAST-A) using Rasch analysis according to the RULER guidelines.1,2

Design: Measurement Development

Setting: Community

Participants: Adolescents ages 12-17 with a documented concussion (n=107)

Interventions: Not Applicable

Main Outcome Measures: The Behavioral Assessment Screening Tool for Adolescents (BAST-A),3,4 adapted from the BAST for adults,5,6 is a 59-item measure of neurobehavioral functioning. The measure captures severity of symptoms in three domains: Negative Affect and Fatigue, Executive and Social Function, and Risk Behaviors.

Results: Each of the BAST-A subscales met criteria for unidimensionality. There were 5 misfitting items for Negative Affect and Fatigue and 1 misfitting item for both Executive and Social Function and Risk Behaviors. However, upon removal of these items, the average domain measures and standard errors did not improve so the items were retained. Item difficulty (i.e., a measure of symptom severity) also displayed a range across subscales (-1.14 to 2.11 logits). In this sample of adolescents with concussion, less severe symptoms were more commonly endorsed for each of the subscales. However, risk behaviors were endorsed the least. The Executive and Social Function subscale displayed appropriate targeting, but the Negative Affect and Fatigue and Risk Behaviors subscales had higher domain values (-1.06 and -3.38) indicating that more items need to be added to capture different severity levels.

Conclusions: The BAST-A displayed adequate psychometric properties for this sample according to RULER guidelines for the Negative Affect and Fatigue and Executive and Social Function subscales but not the Risk Behaviors subscale. Future work should include adolescents with traumatic brain injury (TBI) of various severity levels and causes of injury as this sample consisted of mostly sports concussion.

Author(s) Disclosures: The authors report no conflicts of interest.