The structure of post-concussion symptoms in adolescent student athletes: Confirmatory factor analysis and measurement invariance

Abstract

Objective: This study examined factor models for the Post-Concussion Symptom Scale (PCSS) at baseline and after suspected sport-related concussion, and measurement invariance from pre-injury to post-injury assessments and across age, gender, and health history groups (e.g., attention-deficit/hyperactivity disorder, psychiatric history). Methods: Adolescent student athletes (ages 13-18) completed a baseline PCSS (n = 39,015; 54.3% boys) and a subsample within 21 days of a suspected concussion (n = 1,554; 56.7% boys) completed a post-injury PCSS. Five models were evaluated for fit and invariance. Results: Confirmatory factor analyses showed good baseline and post-injury model fit for a previously supported four-factor model (i.e., cognitive-sensory, sleep-arousal, vestibular-somatic, and affective), an alternative four-factor model (i.e., cognitive, sleep-arousal, physical, and affective), and an incomplete bifactor model with vestibular-somatic and affective specific factors, along with partial invariance from pre-injury to post-injury assessments. Partial-to-full invariance was established for each model at baseline across demographic and health history groups. Conclusions: Results showed empirical and conceptual support for both PCSS subscales (i.e., cognitive, sleep-arousal, physical, and affective) and a total score for use in pre-injury to post-injury assessments and across demographic and health history groups at baseline. Future normative data, stratified by demographics and health history, could provide more precise symptom assessments for concussion management.