A latent class analysis of injury patterns among rear-seated, seat-belted children

Background. The objective of the study was to determine the constellation of injury patterns in rear-seated, seat-belt-restrained children using data-driven latent class methodology novel to injury prevention research. Methods: A cross-sectional probability sample of rear-seated, belt-restrained children aged 5 to 15 years in crashes was obtained via insurance claims records and a telephone survey. Eight body regions of AIS 2 or greater injury (concussion, nonconcussive head injury, face, chest, abdomen, neck/spine/back, upper extremity, lower extremity) were determined, and a latent class model was fit to determine whether underlying "injury clusters" were present. Results. A three-class model appears to best fit to observed data: an "abdominal/spine" cluster that contained 11% of the population, a "concussion" cluster that contained 56% of the population, and a residual "mixture" cluster that contained the remaining 33% of the population. When compared with the mixture cluster, the abdomen/spine cluster was associated with 4 to 8 year old children, lap-only belt restraint, frontal impacts, and minivans. The concussion cluster was also more common among 4 to 8 and 9 to 12 year olds, with side or rear impact crashes, and with pickup trucks. Conclusions. Latent class analysis allows injury clusters to be estimated from the data, not predetermined by the investigator, and suggests that distinct mechanisms of abdominal injury and concussive head injury exist in a population-based sample of children in motor vehicle crashes.