A Four-Factor Symptom Model for Student Athletes with and without Prior Concussions
Abstract Purpose This study was designed to determine whether the factor structure of the Post-Concussion Symptom Scale (PCSS) is consistent (i.e., invariant) in subgroups of youth stratified by their history of prior concussions. A four-factor model of the PCSS has empirical support (i.e., cognitive-sensory, sleep-arousal, vestibular-somatic, and affective symptoms), and the current investigation examined (i) whether this model is invariant across athletes with 0, 1, or ≥2 prior concussions and (ii) group differences across symptom clusters. Methods Participants included a large sample of adolescent athletes (54.4% boys; 13–18 years-old, M=15.50±1.27 years) with 0 (n=32,668), 1 (n=4,110), or ≥2 prior concussions (n=1,681) administered the 22-item PCSS at pre-season baseline. Measurement invariance was tested (good fit=CFI≥0.95; significant change-in-fit=∆CFI≥0.01) and effect sizes for group differences in symptom clusters were calculated. Results The four-factor model showed configural (CFI=0.968) and weak invariance (∆CFI=0.006), but not strong invariance. Comparisons based on concussion history across symptom clusters indicated very small differences between athletes with 0 and 1 prior concussion(s) (d=0.05-0.15), slightly larger differences between athletes with 1 and ≥2 prior concussion(s) (d=0.12-0.24), and modestly larger differences between athletes with 0 and ≥2 prior concussion(s) (d=0.17-0.35). Conclusion These findings support partial invariance of the four-factor PCSS model across athletes with and without a history of concussion. The association with prior concussion was greatest for cognitive-sensory symptoms and smallest for vestibular symptoms. Future research should explore whether alternative factor models are invariant across athletes with different numbers of prior concussions.