Modeling of the Opioid and Neonatal Abstinence Syndrome Co-Epidemic in Medicaid Population of Indiana

Background:With the onset of a second decade of opioid use creating devastating outcomes in the United States, there is value in studying the prevalence of Opioid Use Disorder (OUD) in pregnant women, a group critical to this outbreak. We used Medicaid claims 2014-2019, to analyze the medical, social and economic aspects of OUD in pregnant women and their babies. Objectives:The research aims to study the impact of Opioid Use Disorder (OUD) in Medicaidenrolled pregnant women and model the vertical perinatal effect of OUD, known as Neonatal Abstinence Syndrome (NAS), to the newborn baby. We also attempt to understand the effects of state legislation on doctor shopping and the role of prescribers in fraudulent solicitation of opioids. Methods:The research uses multivariate logistic regression to create a predictive model for highrisk pregnant women based on their claims history. Doctor shopping trends pre- and postlegislation are analyzed using regression discontinuity and graph analysis of the co-prescription network of physicians. Finally, OUD and NAS are modeled together as a probabilistic Bayesian belief network to simulate the cost of interventions, namely MAT enrollment, pharmacotherapy and dyad rooming-in. Results:Pregnant women who may have NAS offsprings are likely to have a history of nicotine addiction, alcohol use, dependence on pain medication and mental health diagnoses in the years leading up to pregnancy. State legislation is found to reduce prescription opioid shopping over the years, though the research highlights the need for policies to target complicit prescribers in addition to prescription monitoring. Finally, the compartment model calculates an incidence of 19-22% for downstream neurological delay in babies born to high-risk mothers. Lower delay rates are found to be associated with high MAT enrollment in the mother, high rooming-in of mother-infant dyad and low rate of pharmacotherapy in infants