Typologies of Maternal Substance Use in Pregnancy: Latent Classes and Sociodemographic Correlates in a U.S. Sample

OBJECTIVE: Substance use patterns vary considerably in the general population, yet little is known about patterns before and during pregnancy. The purpose of this study was to describe single substance and polysubstance use (PSU) before and during pregnancy among recent births in the United States and compare exposure patterns. METHOD: We used data from the Pregnancy and Risk Assessment Monitoring System (PRAMS) postpartum survey for 2016-2018 to estimate the prevalence and identify patterns of substance use by participants 1 to 3 months before and during pregnancy. Data on use of cigarettes, cannabis, opioids, amphetamines, heroin, and cocaine were available for seven states. We used latent class analysis to identify patterns of substance use and tested whether participants with those patterns differed by age, income, race/ethnicity, and pre-pregnancy alcohol use. Analyses were weighted using PRAMS' state-level weights. RESULTS: We studied 15,429 participants representing 384,918 live, singleton births. Approximately half (51.3%) were 20-29 years old, 42.3% were above 200% of the Federal Poverty Level, and 73.3% were non-Hispanic White. We identified seven latent classes of maternal substance use: minimal users (70.7%), pre-pregnancy cigarette users (10.5%), persistent cigarette users (6.8%), pre-pregnancy cannabis users (5.5%), broad PSU (3.6%), opioid-only users (1.9%), and persistent cigarette/opioid co-users (1.0%). Participants in these groups differed significantly by age, income, race/ethnicity, and pre-pregnancy alcohol use (p < .001). CONCLUSIONS: This novel empirical classification of single substance and PSU could help to further our understanding of the impact of PSU on perinatal health and to design interventions for maternal substance use.