Abusive head trauma in young children: a population-based study

Pediatr Emerg Care. 2013 Mar;29(3):283-91. doi: 10.1097/PEC.0b013e31828503ea.


Objective: The objectives of this study were to provide population-based incidence estimate of abusive head trauma (AHT) in children aged 0 to 5 years from inpatient and emergency department (ED) and identify risk characteristics for recognizing high-risk children to improve public health surveillance.

Methods: This was a retrospective cohort study based on children's first encounter in ED or hospital admission with a diagnosis of head trauma (HT), 2000-2010. The relationship between clinical markers and AHT was examined controlling for covariables in the model using Cox hazards regression. Kaplan-Meier incidence probability was plotted, and the number of weeks elapsing from date of birth to the first encounter with HT established the survival time (T).

Results: Twenty-six thousand six hundred eighty-one children had HT, 502 (1.8%) resulted from abuse; 42.4% was captured from ED. Incidence varied from 28.9 (95% confidence interval [CI], 27.9-37.4) in infants to 4.1 (95% CI, 2.4-5.7) in 5-year-olds per 100,000 per year. Adjusted hazard ratio was 20.3 (95% CI, 10.9-38.0) for intracranial bleeding and 11.4 (95% CI, 8.57-15.21) for retinal hemorrhage.

Conclusions: Incidence estimates of AHT are incomplete without including ED. Intracranial bleeding is a cardinal feature of AHT to be considered in case ascertainment to improve public health surveillance.

MeSH terms

  • Chi-Square Distribution
  • Child Abuse / diagnosis*
  • Child, Preschool
  • Craniocerebral Trauma / diagnosis*
  • Diagnosis, Differential
  • Female
  • Humans
  • Incidence
  • Infant
  • Infant, Newborn
  • Intracranial Hemorrhages / diagnosis
  • Intracranial Hemorrhages / epidemiology
  • Male
  • Poisson Distribution
  • Population Surveillance
  • Proportional Hazards Models
  • Retinal Hemorrhage / diagnosis
  • Retinal Hemorrhage / epidemiology
  • Retrospective Studies
  • Risk Factors
  • South Carolina / epidemiology