On-scene factors that predict severe injury of patients involved in frontal crashes of passenger cars

Purpose: We sought to determine on-scene factors that predict severe injury to the occupants of passenger cars involved in frontal crashes. Methods: From January 2011 to December 2014, we collected data from patients who were taken to two emergency centres following a frontal motor vehicle crash. Binomial logistic regression was used to model the effects of occupant characteristics (sex, age, body mass index), vehicle damage (according to the collision deformation classification code), and safety devices on severe injuries (injury severity score >15). Results: Of 344 subjects, 75 (21.8 %) had severe injuries. Sex, seat belt status, extent of vertical crash, intrusion, and deformation extent (DE) were significantly different between severe and non-severe injuries. After adjusting for confounders, non-use of seat belt tripled the odds of severe injury [odds ratio (OR) 2.7, 95 % confidence interval (CI) 1.461–5.105]. DE ≥4 and intrusion increased the risk of severe injury (OR 2.4, 95 % CI 1.120–5.204 and OR 5.2, 95 % CI 2.525–10.780, respectively). A combination model to predict severe injury using intrusion, seat belt use, and DE ≥4 demonstrated 56.0 % sensitivity, 88.9 % specificity, and 58.4 % positive predictive value (AUC = 0.781, 95 % CI 0.734–0.824). Conclusions: For passenger cars involved in a frontal crash, intrusion, unbelted status, and DE ≥4 are good predictors of severe injury. Sequential criteria using vehicle DE, seat belt use, and intrusion can be used by first responders to triage patients involved in a frontal collision.