PT  - JOURNAL ARTICLE
AU  - Singichetti, Bhavna
AU  - Marshall, Stephen
AU  - Cantrell, John
AU  - Breedlove, Katherine
AU  - Cameron, Kenneth
AU  - Pasquina, Paul
AU  - McCrea, Michael
AU  - Broglio, Steven
AU  - McAllister, Thomas
TI  - 122 School-level determinants of variability in observed concussion incidence: a care consortium study
AID  - 10.1136/injuryprev-2020-savir.44
DP  - 2020 May 01
TA  - Injury Prevention
PG  - A17--A18
VI  - 26
IP  - Suppl 1
4099  - http://injuryprevention.bmj.com/content/26/Suppl_1/A17.3.short
4100  - http://injuryprevention.bmj.com/content/26/Suppl_1/A17.3.full
SO  - Inj Prev2020 May 01; 26
AB  - Statement of Purpose To investigate variability in observed concussion incidence between NCAA colleges participating in a multisite prospective study of concussion, and to quantify the effect of selected school-level factors on concussion risk.Methods/Approach Data on sport-related concussion (SRC) were provided by the CARE Consortium, a multisite study of 30 collegiate institutions. School-level factors included NCAA Division (DI, DII, DIII) and school type (military or civilian) and a 3-level Sport Risk Index (Low, Medium, High). Random intercepts G-side log-binomial regression was used to model between- and within-school variability in concussion risk in NCAA athletes. Three concussion outcomes were modeled: all SRCs, competition SRCs only, and practice SRCs only.Results School-Level Risk Factors: In fully adjusted models with all SRCs as the outcome, Sport Risk Index was the strongest predictor (risk ratio (RR) of 6.0; 95%CI: 4.4, 8.1, for the High vs. Low categories of the Index). Concussion risk was higher in Division I schools than in Division 2 (RR=1.6, 95%CI: 0.6, 4.2) and Division 3 schools (RR=1.8, 95%CI: 0.9, 3.6) schools. Military academies had an elevated risk of SRC (RR=1.5; 95%CI: 0.7, 3.3; analysis limited to NCAA athletes). School-Level Variability: Most of the variability in SRC risk was at the level of the athlete, not at the school. For all three outcomes, the within-school (athlete-level) variance was over five times the between-school variance. Adjusting for our three school-level risk factors (Division, Mil/Civ, and Sport Risk Index) removed 40% of the variation between schools for competition SRC, and 25% for overall SRCs and practice SRCs.Conclusions Sport-level factors predict concussion risk, and a substantial portion of variability in concussion risk between schools is readily explainable.Significance and Contributions to Injury and Violence Prevention Science Understanding school-level determinants and variability in concussion risk may identify opportunities for interventions to reduce SRC incidence.