Objective The purpose of this study is to rigorously evaluate the SCAT symptom list and improve the sensitivity and sensitivity by reducing the list.

Design Cohort prospective study.

Setting Collegiate athletics and military service academies.

Participants 46,836 athletes and cadets enrolled in the NCAA/DOD CARE consortium; 3,362 were diagnosed with a concussion.

Outcome Measures The independent variables are the individual SCAT Symptom Evaluations and the sum of specific individual symptoms.

Main Results Individual symptoms demonstrate a variety of Cohen’s-d effect sizes with the smallest effect sizes being Nervous or Anxious (d=0.19) and Sadness (d=0.36). The largest effect sizes are Pressure in Head (d=2.29), Don’t Feel Right (d=2.30), and Headache (d=2.54). A Machine Learner (ML) utilizing all 22 symptoms has an AUC=0.91, however, this may not be viable because clinicians would need to use a phone/web app to interpret the data. Clinicians often use Symptoms Severity Score because this can be easily performed. A ML utilizing Symptom Severity Score has an AUC = 0.81. A ML utilizing a sum of Pressure in Head, Don’t Feel Right, and Headache has an AUC of 0.89.

Conclusions These results suggest a reduction in the future SCAT-6 symptom list should be considered, possibly to 3 questions for an AUC improvement from 0.81 to 0.89 with a 22% improvement in sensitivity and 6.2% improvement in specificity.