Objective To develop an algorithm that accurately discriminates recovery duration following sport-related concussion (SRC).

Design Retrospective case-control study.

Setting Regional concussion center.

Participants Athletes ages 12–18 presenting within 28 days of injury between 11/2017–10/2020 were included and randomized into training (70%) and test (30%) samples for algorithm validation. Those with positive head imaging, non-sport mechanism, or incomplete records were excluded.

Assessment of Risk Factors Independent variables included age, sex, sport contact level, loss of consciousness, post-injury amnesia, initial Post-Concussion Symptom Scale (PCSS) score, days to presentation, concussion history, personal or family history of ADHD, migraines, or psychiatric diagnosis.

Outcome Measures The primary outcome was days to recovery, grouped as early (≤14 days), typical (15–27 days), or delayed (≥28 days). Recovery was defined as symptom resolution or return to baseline PCSS or initiation of graduated return-to-play.

Main Results A total of 493 patients were included (mean age 15.7±1.5 years, 68.2% male, 70.0% white). Median time to presentation was 5 days (IQR: 2–10 days). Recovery duration was ≤14 days in 52.3% of patients, 15–27 days in 21.5%, and ≥28 days in 26.2%. The variables most predictive of recovery duration were initial PCSS (≤6, 7–28, or ≥29), time to presentation (≤7 vs >7 days), and prior concussions (0 vs ≥1). The model accurately discriminated early versus typical or delayed recovery (AUC = 0.80), and correctly identified >90% of patients recovering early.

Conclusions This predictive algorithm accurately discriminated recovery duration and may aid in allocating resources, counseling patients, and making timely referrals.