Objectives: To determine the incidence of and risk factors associated with infant (< 1 year of age) physical abuse in Alaska.
Methods: A population-based retrospective cohort study for the 1994-2000 resident birth cohort was conducted by linking data from birth certificates, Child Protective Services, a statewide hospital-based trauma registry, hospital discharge data, and the Alaska Infant Mortality Review (including death certificates). The main outcome measures were the incidences of overall physical abuse and abuse resulting in hospitalization or death. A case of child abuse was defined as an instance of substantiated physical abuse to an infant identified in the Child Protective Services database or an infant death with homicide identified on the death certificate as the manner of death.
Results: During the 7-year study period, there were 70,842 births and 325 cases of physical abuse including 72 that led to hospitalization (n = 58), death (n = 4), or both (n = 10); respective incidences for all abuse and abuse leading to hospitalization or death were 4.6 and 1.0 per 1000 live births. Following multivariate analyses, the risk factors with the highest population attributable risks were maternal or paternal education < or = 12 years, unmarried mother, and maternal prenatal substance use. To determine if the study methodology was likely to have missed cases of severe abuse, we examined information for all 216 infants hospitalized for trauma during the study period who did not have identification of abuse in one of the study databases; of these, at least 39 had injuries inconsistent with the reported mechanism (a long bone or skull fracture that reportedly resulted from a fall of less than 3 feet or from a caretaker's arms or for which the caretaker denied a history of trauma).
Conclusions: Alaska has one of the highest documented infant physical abuse incidences reported in the literature and abuse is associated with potentially modifiable-primarily social-risk factors. Despite this high incidence, substantial under-reporting of hospitalized cases likely occurs.