OBJECTIVES: To identify predictors of prolonged recovery (ie, >28 days) using patient demographic factors and healthy, norm-based cutoffs on a multimodal test battery in adolescents after sport-related concussions (SRCs). DESIGN: Retrospective cohort. Patients were deemed recovered after successful completion of return-to-play/school protocols and received medical clearance. SETTING: Community concussion clinic. PATIENTS: Male and female adolescent student athletes diagnosed with a SRC and evaluated within 1 week of injury. INDEPENDENT VARIABLES: Patient demographics, medical history, injury description, computerized neurocognitive testing, vestibular/ocular testing, and symptoms at initial clinical visit. MAIN OUTCOME MEASURES: Performance on clinical testing to predict recovery duration, classified as normal (<28 days) or prolonged (>28 days). RESULTS: A total of 201 adolescent student athletes (age = 15.3 +/- 1.4 years) were included (female 35%). Average recovery duration for the entire cohort was 22.3 +/- 13.3 days, with 22% (n = 45) of adolescent student athletes taking >28 days to recover. The final model was 88.3% accurate in classifying normal and prolonged recovery. Predictor variables included sex, loss of consciousness, history of ocular disorder, history of concussion, performance on visual motor speed composite, visual motion sensitivity symptom provocation and near point of convergence distance, number-naming total time, and symptom count. CONCLUSIONS: These findings suggest that using norm-based cutoffs from cognitive, oculomotor, and vestibulo-ocular testing and symptom reporting, clinicians can accurately predict a prolonged recovery (sensitivity = 81%) and normal recovery (specificity = 83%) in an adolescent, SRC cohort.