Development of a Probability Model to Predict Head Injury Risk in Pediatric Falls

• NCJ Number: 306560
• Author(s): Gina Bertocci; Mary Clyde Pierce; Raymond Dsouza; Angela Thompson
• Date Published: 2023
• Length: 9 pages
• Grant Number(s): 2019-DU-BX-0029
• Agencies: NIJ-Sponsored

Annotation

This document reports on a project to develop evidence-based statistical models capable of predicting the probability of head injury in young children involved in falls, to improve forensic biomechanical investigations.

Abstract

The authors of this report had three specific aims: to describe the fall, child, and injury characteristics associated with pediatric falls resulting in head injury; to develop and validate a biofidelic 3D human child computer model used to determine head accelerations in simulations of reconstructed common pediatric fall scenarios; and to develop evidence-based statistical models to predict probability of head injury associated with falls involving young children. The overall project goal was to improve biomechanical investigations by developing evidence-based statistical models capable of predicting the probability of head injury in young children involved in falls. The research resulted in the creation of a dataset of 99 falls involving young children with a head injury; each fall in the dataset contains descriptions of patient demographics, anthropometrics, and relevant medical history, fall characteristics and environment details, injury outcomes and severity, a skin assessment body diagram, head injury details, imaging data, biomechanical outcomes of the fall, and accident versus abuse determination. The authors suggest that the evidence-based pediatric-specific head injury probability models developed through this research will help to predict the likelihood of head injury in fall reconstruction, which is a key element for improving confidence when performing biomechanical forensic investigations.