Inflicted versus unintentional injury: towards improved detection and surveillance of supervisory neglect.

BACKGROUND:Many children evaluated for child abuse have non-inflicted injuries due to supervisory neglect. Despite its prevalence, supervisory neglect has received minimal attention. METHODS:We performed a retrospective review of patients maintained in a hospital forensic registry. Text analytics software was used to classify types of supervisory neglect from reported mechanism of injury (MOI). Logistic regression was used to assess risks associated with supervisory neglect. RESULTS:For 1185 eligible patients, 553 were classified as having unintentional injuries. Text analysis identified four types of supervisory neglect for falls (N = 376): interrupted supervision (53.4%); failure to adapt the home (26.7%); safety equipment non-compliance (12.8%); inadequate substitute care (7.0%); a fifth category was identified for non-falls (N = 54): rough handling (32.0%). Supervisory neglect was associated with MOI consistent with the injury (AOR = 15.5[p < 0.001]), no loss of consciousness (AOR = 6.8(p < 0.001]), no bruising away from the injury site (AOR = 3.7[p < 0.001]), and direct hospital presentation (AOR = 1.8[p < 0.05]). Of the 553 with unintentional injury, 62% had isolated head injuries; 20%, however, had evidence of prior head trauma. CONCLUSIONS:Interrupted supervision was the most common form of supervisory neglect. Twenty percent of forensics patients with head injury found to have non-inflicted injuries had evidence of prior head injury. Supervisory neglect suggests a high-risk population. LEVEL OF EVIDENCE:Level II, Prognosis.