Latent Class Analysis of Prenatal Substance Exposure and Child Behavioral Outcomes.

Abstract

ObjectivesTo predict behavioral disruptions in middle childhood, we identified latent classes of prenatal substance use.Study designAs part of the Environmental influences on Child Health Outcomes (ECHO) Program, we harmonized prenatal substance use data and child behavior outcomes from 2,195 women and their 6- to 11-year-old children across 10 cohorts in the United States and used latent class-adjusted regression models to predict parent-rated child behavior.ResultsThree latent classes fit the data: low use (90.5%; n=1,986), primarily using no substances; licit use (6.6%; n=145), mainly using nicotine with a moderate likelihood of using alcohol and marijuana; and illicit use (2.9%; n=64), predominantly using illicit substances along with a moderate likelihood of using licit substances. Children exposed to primarily licit substances in utero had higher levels of externalizing behavior than children exposed to low or no substances (p=.001, d=.64). Children exposed to illicit substances in utero showed small but significant elevations in internalizing behavior than children exposed to low or no substances (p<.001, d=.16).ConclusionsThe differences in prenatal polysubstance use may increase risk for specific childhood problem behaviors; however, child outcomes appeared comparably adverse for both licit and illicit polysubstance exposure. We highlight the need for similar multi-cohort, large-scale studies to examine childhood outcomes based on prenatal substance use profiles.