Evaluating abusive head trauma in children < 5 years old: Risk factors and the importance of the social history

Background: Abusive head trauma (AHT) is the leading cause traumatic death in children ≤ 5 years of age. AHT remains seriously under-surveilled, increasing the risk of subsequent injury and death. This study assesses the clinical and social risks associated with fatal and non-fatal AHT. Methods: A single-institution, retrospective review of suspected AHT patients ≤ 5 years of age between 2010 and 2016 using a prospective hospital forensic registry data yielded demographic, clinical, family, psycho-social and other follow-up information. Descriptive statistics were used to look for differences between patients with AHT and accidental head trauma. Logistic regression estimated the adjusted odds ratios (AOR) for AHT. A receiver operating characteristic (ROC) curve was created to calculate model sensitivity and specificity. Results: Forensic evaluations of 783 children age ≤ 5 years with head trauma met the inclusion criteria; 25 were fatal with median[IQR] age 23[4.5–39.0] months. Of 758 non-fatal patients, age was 7[3.0–11.0] months; 59.5% male; 435 patients (57.4%) presented with a skull fracture, 403 (53.2%) with intracranial hemorrhage. Ultimately 242 (31.9%) were adjudicated AHT, 335(44.2%) were accidental, 181 (23.9%) were undetermined. Clinical factors increasing the risk of AHT included multiple fractures (Exp(β) = 9.9[p = 0.001]), bruising (Expβ = 5.7[p < 0.001]), subdural blood (Exp(β) = 5.3[p = 0.001]), seizures (Exp(β) = 4.9[p = 0.02]), lethargy/unresponsiveness (Exp(β) = 2.24[p = 0.02]), loss of consciousness (Exp(β) = 4.69[p = 0.001]), and unknown mechanism of injury (Exp(β) = 3.9[p = 0.001]); skull fracture reduced the risk of AHT by half (Exp(β) = 0.5[p = 0.011]). Social risks factors included prior police involvement (Exp(β) = 5.9[p = 0.001]), substance abuse (Exp(β) = 5.7[p = .001]), unknown number of adults in the home (Exp(β) = 4.1[p = 0.001]) and intimate partner violence (Exp(β) = 2.3[p = 0.02]). ROC area under the curve (AUC) = 0.90([95% CI = 0.86–0.93] p = .001) provides 73% sensitivity; 91% specificity. Conclusions: To improve surveillance of AHT, interviews should include and consider social factors including caregiver/household substance abuse, intimate partner violence, prior police involvement and household size. An unknown number of adults in home is associated with an increased risk of AHT. Study Type/Level of Evidence: Prognostic, Level III.