Objective Vestibulo-oculomotor impairments are common after concussion. The Buffalo Concussion Physical Exam (BCPE) is a brief, pertinent physical examination for helping to diagnose concussion. We derived a scoring algorithm to identify adolescents who may take longer to recover (>30 days) and may benefit from early specific therapy.

Design Retrospective cohort.

Setting University clinics.

Participants 263 adolescents diagnosed with concussion using international guidelines (14.92±1.9 years, 62% male, 90% sport-related, 6.0±4 days from injury).

Interventions (or Assessment of Risk Factors) BCPE was performed at the initial visit and weekly until recovery. We used a binomial generalized linear model with complementary log-log error function. Significant predictor variables were stepwise selected using Akaike Information Criterion, and k-fold cross-validation was used for internal validity.

Outcome Measures Days since injury (1-point/day), high-velocity/multiple impacts (2-points), >2 previous concussion (4-points), dizziness on postural change (5-points), abnormal vestibulo-ocular reflex (VOR, 5-points), abnormal tandem gait (1-point), high-velocity/multiple impact plus abnormal VOR (5-points), and dizziness on postural change plus abnormal VOR (minus 4-points) were used to categorize low (0–10), medium (11–15) and high (>15 points) risk.

Main Results The algorithm correctly identified 73% of normal recoverers as low-risk, 23% as medium-risk, and 4% as high-risk. It correctly identified 28% of delayed recoverers as low-risk, 28% as medium-risk, and 44% as high-risk. Medium- to high-risk score within one week of injury was 72% sensitive and 74% specific for delayed recovery.

Conclusions An easily calculable score using 3 questions and 3 physical exam findings within one week of injury identified concussed adolescents who may benefit from early therapeutic intervention.