Identifying risk factors for child maltreatment in Alaska: a population-based approach

Abstract

Background: Child maltreatment has been linked to multiple negative health outcomes and many leading causes of death. Statewide population-based evaluations are needed to identify high-risk populations early in life for targeted interventions.

Purpose: To assess the utility of combining Pregnancy Risk Assessment Monitoring System (PRAMS) data with child protective services (CPS) records to identify risk factors associated with Protective Services Reports (PSR) suggestive of child maltreatment.

Methods: This was a retrospective population-based cohort study conducted in the spring of 2010 using weighted survey data from Alaska PRAMS for birth years 1997-1999. PRAMS responses were linked with CPS records for the sampled child. The outcome of interest was any PSR made to CPS after the survey was returned through 48 months after birth. Validation of the PRAMS data set occurred through direct comparison between the total population and PRAMS weighted sample for birth certificate factors. Multivariate logistic regression models were constructed to identify risk groups.

Results: In the final multivariate model among the main effect variables, three of the top five strongest associated factors were derived all or in part from PRAMS. Public aid as a source of income had a significant interaction with Alaska Native status, and among Alaska non-Natives had an AOR of 3.37 (95% CI=2.2, 5.1). Six significant modifiable factors were identified in the multivariate model. Three quarters (75%) of the maltreatment cases occurred among children with two or more of these factors, despite being found in about one third (32%) of the total population.

Conclusions: Although birth certificates remained a valuable source of risk factor information for child maltreatment, PRAMS identified additional risk factors not available from birth certificates.