Data Can Help Predict Children Most Vulnerable to Injury and Death

​A population-based study was able to identify children at birth who, by age three, were most likely to be hospitalized or die by using a predictive model built from administrative data. The study, "Injury and Mortality among Children Identified as at High Risk of Maltreatment" (published online Jan. 29), will be published in the February 2018 issue of Pediatrics. The authors followed 60,006 infants born in 2011 in New Zealand, assigning them "risk scores" based on the likelihood they would be substantiated for maltreatment by the age of two. Those scores were generated from government records, and included information such as maternal age and if the parents had a criminal justice history. Findings document that although a relatively small number of children experience injury hospitalizations and death during the first three years of life, these children are concentrated among the highest risk deciles of a model built using birth records and predicting child protection involvement. Among children in the study population who died from inflicted injury deaths or unintentional injuries, half would have been identified at the time of birth as falling in the top 10 percent for risk of substantiated maltreatment. Likewise, half of all children hospitalized for maltreatment-related injury had scores at birth placing them in the top 10 percent for substantiated maltreatment. The authors suggest a predictive model could help target children and families who would benefit from intensive, voluntary supports and preventive services to reduce rates of injury and mortality.

###

The American Academy of Pediatrics is an organization of 66,000 primary care pediatricians, pediatric medical subspecialists and pediatric surgical specialists dedicated to the health, safety and well-being of infants, children, adolescents and young adults. For more information, visit www.aap.org and follow us on Twitter @AmerAcadPeds