Hospital Variation in Child Protection Reports of Substance Exposed Infants

Background: It is estimated that 298,000 women in the United States used illicit drugs or alcohol during pregnancy in 2015, of whom 109,000 used illicit drugs and 105,000 engaged in binge alcohol use (SAMHSA, 2016). Federal law includes a requirement for health personnel to inform the child welfare system of infants prenatally exposed to substances (CAPTA, PL 93-247, 2010). However, recent research indicates that just over half of infants diagnosed with prenatal substance exposure were reported to child protective services (Putnam-Hornstein, Prindle, & Leventhal, 2016). It is unknown whether hospital-level factors contribute to these discrepancies in reporting.

We conducted a statewide retrospective cohort study to examine hospital-level and birth-level differences in reporting substance-exposed infants to CPS. Enhancing our understanding of these potential differences will provide insights on CPS reporting discrepancies and whether improved policies are necessary at the hospital level.

Methods: We utilized a population-based linked dataset of birth, hospital, and CPS administrative records using diagnostic (ICD-9) codes to identify infants prenatally exposed to substances. Performing multilevel and conventional logistic regression models, we examined hospital-level and individual birth-level factors in relation to a report to CPS among those infants prenatally exposed to substances.

Results: Of the 760,863 infants born in Washington State between 2006 and 2013, 12,308 (1.6%) were diagnosed with prenatal substance exposure. The multilevel model was identified as a better fit for the data by comparing Bayesian Information Criterions (BIC). Infants born at hospitals that served larger populations of patients with Medicaid (OR: 1.25) and hospitals with higher occupancy rates (OR: 1.43) were more likely to be reported to CPS. Infants exposed to amphetamines (OR: 2.58) and cocaine (OR: 2.33) were more likely to be reported while infants exposed to cannabis (OR: 0.62) were less likely to be reported to CPS.

Conclusions: We found that hospital-level and individual birth-level factors impact the likelihood of infants prenatally exposed to substances being reported to CPS, providing additional knowledge about which infants are reported to CPS. Hospitals that treat larger proportions of low-income patients, as indicated by payments via Medicaid, and hospitals that are busier, as indicated by higher occupancy rates, have increased likelihoods to report infants diagnosed with substance exposure to CPS. The variation in reporting practices based on hospital-level factors indicates that reports to CPS are not made in a standardized way. Results of these differences present opportunities for targeted training and prevention efforts.