Dual statistical models link baseline visual attention measure to risk for significant symptomatic concussion in sports

Abstract

Aim: Athletic pre-season testing can establish functional baseline for comparison following concussion. Whether impacts of future concussions may be foretold by such testing is little known. Materials & methods: Two sets of models for a significant burden of concussion were generated: a traditional approach using a series of logistic regressions, and a penalized regression approach using elastic net. Results: 3091 youth and adult athletes were baseline-assessed. 90 subsequently experienced concussion and 35 were still experiencing a significant burden of concussion when tested within two weeks. Both models associated prior history of head injury and visual attention-related metrics with a significant burden of concussion. Conclusion: Pre-season testing of visual attention may identify athletes who are at risk for significant sports-related concussion. Plain language summary: Athletic pre-season testing can establish functional baseline for comparison following concussion and may predict impacts of future concussions. In this study, 3,091 youth and adult athletes were baseline-assessed. 90 subsequently experienced concussion and 35 were still experiencing a significant burden of concussion when tested within two weeks. A statistical model and a machinelearning model both associated prior history of head injury and visual attention-related metrics with a significant burden of concussion. Pre-season testing of visual attention may identify athletes who are at risk for significant sports-related concussion.