Association between punitive policies and neonatal abstinence syndrome among Medicaid-insured infants in complex policy environments.

aimsTo estimate the association between punitive policies for women with substance use during pregnancy and odds of neonatal abstinence syndrome (NAS) diagnosis among Medicaid-insured infants, and to estimate this association controlling for the presence of four other policies related to substance use in pregnancy. DESIGN, SETTING AND

participantsAnalysis of live births in Medicaid claims data from 39 US states in varying years between 2006 and 2014 using weighted generalized linear models with clustered standard errors and state and year fixed-effects. MEASUREMENTS: NAS rates in states without punitive policies were compared with rates in states with policies before and after policy enactment using logistic regression models adjusted for individual and county-level factors and including state and year fixed-effects. We estimated odds of NAS controlling for the presence of a potentially treatment-deterring policy requiring reporting of suspected prenatal substance use, and three treatment-supportive policies that create targeted programs for pregnant and postpartum women, prioritize pregnant women's access to substance use disorder treatment programs and prohibit discrimination towards pregnant women in treatment programs.

findingsAmong 9 714 798 weighted live births (1 896 082 unweighted), 49 670 (0.51%) had an NAS diagnosis. The findings were inconclusive as to whether or not punitive policies were associated with odds of NAS either in the short or long term. Odds of NAS among infants born in states with reporting policies were lower than those born in states without such policies [adjusted odds ratio (aOR) = 0.77, 95% confidence interval (CI) = 0.61-0.98].

conclusionsIn the United States, punitive policies for women with substance use during pregnancy do not appear to be associated with lower odds of neonatal abstinence syndrome (NAS). Reporting policies, which are heterogenous in their components and implementation, appear to be associated with lower odds of NAS when controlling for other relevant policies.