|| Checking for direct PDF access through Ovid
Predictors of caregiver-reported behavior problems for 3-year-olds with prenatal cocaine exposure (PCE) and matched controls were examined using structural equation modeling. We tested whether PCE had a direct effect on child behavior problems in a model that included other prenatal drug exposure, child sex, caregiver depression, and the quality of the child's home environment. The sample (N = 256) was drawn from a longitudinal, prospective study of children of (predominantly crack) cocaine-using women and controls matched on race, socioeconomic status, parity, and pregnancy risk. Child Behavior Problems was modeled as a latent variable composed of the 48-item Conners' Parent Report Scale Conduct Problem and Impulsive-Hyperactive scales and the Eyberg Child Behavior Inventory Intensity scale. Caregiver depression was the only significant predictor of Child Behavior Problems. Mean levels of caregiver self-reported depression and reported child behavior problems did not differ between groups. Mean depression scores were well above the recommended clinical cutoff while mean child behavior problems scores were within normal limits. The model explained 21% of the variance in caregiver-reported child behavior problems in our sample of rural African American, low SES youngsters. Non-maternal caregivers of cocaine-exposed children had significantly lower mean depression scores and mean child behavior problems ratings for 2 of 3 scales used in the study compared to biological mothers of children with PCE and controls. For all groups, much larger proportions of children were rated as having clinically significant behavior problems than would be expected based on the prevalence of behavior problems in the general population.