Using administrative data in child protection research – A prospective cohort study of racial disparities in substance exposure and neonatal child protective service involvement

Background and Objective: Prenatal substance exposure is thought to be a factor in maltreatment reports made at or shortly after birth. Yet, in the United States there is broad discretion as to whether documented substance... [ view full abstract ]

Background and Objective: Prenatal substance exposure is thought to be a factor in maltreatment reports made at or shortly after birth. Yet, in the United States there is broad discretion as to whether documented substance prenatal exposure is reported to child protective services (CPS).

Methods: We examined racial differences in medically documented prenatal substance exposure and subsequent maltreatment reports using a unique administrative dataset consisting of linked birth, hospital discharge, and CPS records from California, USA. Diagnostic codes were used to document substance exposure type. CPS records provided information on a maltreatment report received during the neonatal period. The prevalence of infant substance exposure was calculated by race, substance type, and socio-demographic covariates. Multivariate generalized linear models were used to estimate racial differences in the neonatal likelihood of being reported for maltreatment among substance exposed infants.

Results: Based on a population-based California cohort of 474,071 black, white, and Hispanic infants born in 2006, medically documented substance exposure was identified for 1.6% (n=7,428). The prevalence of substance exposure varied significantly across groups (p<.001), with the highest rate observed among black infants(4.1%) and the lowest for Hispanic infants(1.0%). Among infants with white and Hispanic mothers, the most prevalent substance types were amphetamine (41.6% and 41.9%, respectively) followed by cannabis (35.2% and 30.5%, respectively); for infants with black mothers, cannabis was the most common(46.6%) followed by cocaine(30.5%). After adjusting for socio-demographic birth characteristics, no racial variations in the reporting of maltreatment were observed among infants in which substance exposure was documented.

Conclusions: Hospital discharge records linked to vital birth records provide an imperfect yet still valuable means to monitoring population-level patterns of documented substance exposure. Cross-sector linkages to administrative child protective services records provide valuable information concerning reporting decisions made by the medical system that may vary by race and substance type.