Reliability and Validity of the Sport Concussion Assessment Tool–3 (SCAT3) in High School and Collegiate Athletes

Abstract

Background: The Sport Concussion Assessment Tool–3 (SCAT3) facilitates sideline clinical assessments of concussed athletes. Yet, there is little published research on clinically relevant metrics for the SCAT3 as a whole. Purpose: We documented the psychometric properties of the major SCAT3 components (symptoms, cognition, balance) and derived clinical decision criteria (ie, reliable change score cutoffs and normative conversation tables) for clinicians to apply to cases with and without available preinjury baseline data. Study Design: Cohort study (diagnosis); Level of evidence, 2. Methods: High school and collegiate athletes (N = 2018) completed preseason baseline evaluations including the SCAT3. Re-evaluations of 166 injured athletes and 164 noninjured controls were performed within 24 hours of injury and at 8, 15, and 45 days after injury. Analyses focused on predictors of baseline performance, test-retest reliability, and sensitivity and specificity of the SCAT3 using either single postinjury cutoffs or reliable change index (RCI) criteria derived from this sample. Results: Athlete sex, level of competition, attention-deficit/hyperactivity disorder (ADHD), learning disability (LD), and estimated verbal intellectual ability (but not concussion history) were associated with baseline scores on ≥1 SCAT3 components (small to moderate effect sizes). Female sex, high school level of competition (vs college), and ADHD were associated with higher baseline symptom ratings ( d = 0.25-0.32). Male sex, ADHD, and LD were associated with lower baseline Standardized Assessment of Concussion (SAC) scores ( d = 0.28-0.68). Male sex, high school level of competition, ADHD, and LD were associated with poorer baseline Balance Error Scoring System (BESS) performance ( d = 0.14-0.26). After injury, the symptom checklist manifested the largest effect size at the 24-hour assessment ( d = 1.52), with group differences diminished but statistically significant at day 8 ( d = 0.39) and nonsignificant at day 15. Effect sizes for the SAC and BESS were small to moderate at 24 hours (SAC: d = −0.36; modified BESS: d = 0.46; full BESS: d = 0.51) and became nonsignificant at day 8 (SAC) and day 15 (BESS). Receiver operating characteristic curve analyses demonstrated a stronger discrimination for symptoms (area under the curve [AUC] = 0.86) than cognitive and balance measures (AUCs = 0.58 and 0.62, respectively), with comparable discrimination of each SCAT3 component using postinjury scores alone versus baseline-adjusted scores ( P = .71-.90). Normative conversion tables and RCI criteria were created to facilitate the use of the SCAT3 both with and without baseline test results. Conclusion: Individual predictors should be taken into account when interpreting the SCAT3. The normative conversion tables and RCIs presented can be used to help interpret concussed athletes’ performance both with and without baseline data, given the comparability of the 2 interpretative approaches.