DESCRIPTION (provided by applicant): Prenatal opioid agonist therapy with methadone (MMT) or buprenorphine (BMT) prevents maternal illicit opioid use and withdrawal and improves pregnancy outcomes. However, exposing the fetus to these therapies increases the risk and severity of neonatal withdrawal (NAS, neonatal abstinence syndrome) vs. heroin exposure only. Forty to 80% of exposed neonates require prolonged hospitalization and treatment for NAS. The number of opioid dependent pregnant women in the US is ~900,000 annually and continues to increase. There is an urgent need to identify the safest treatments for opioid dependency in pregnancy. While the effectiveness of opioid agonist therapies on maternal relapse and withdrawal is established, it is less clear whether BMT vs. MMT is superior for fetal and neonatal well-being. Some studies suggest that prenatal BMT decreases NAS severity compared to MMT but the interpretation of these studies is limited by important differences in women treated with BMT vs. MMT (i.e., bias from study drop out and confounding by indication). The proposed study will overcome these limitations by using an existing large database paired with a prospective clinical cohort and employing modern methods to measure and adjust for bias. Thus, the results of this study will guide providers and opioid dependent pregnant women in safe treatment choices. We will use the Massachusetts Medicaid Analytic eXtract (MAX) database from 2006 (when BMT began to be used in pregnancy in Massachusetts) through 2011 (Medicaid data available at the proposed funding start date) to identify the largest cohort of
agonist therapy exposed mother-infant pairs (~9,000) studied to date (the MAX Cohort). In addition, we will enroll 115 women and their neonates (the RESPECT Cohort) attending Boston Medical Center's Project RESPECT Clinic, a leading clinic that specializes in caring for opioid dependent pregnant women. Data on a wide range of potential confounders will be collected from RESPECT Cohort participants; their neonatal outcomes will be abstracted from medical charts following birth. The RESPECT Cohort data will be used to create a propensity score that captures how maternal characteristics influence the probability of prenatal treatment with BMT vs. MMT. The propensity scores will be used in established bias models to determine the amount of bias from uncontrolled confounding by indication that is included in the estimated effect of BMT vs. MMT on each neonatal outcome. This measured bias will then be used to remove confounding in the MAX Cohort analysis of the comparative safety of BMT vs. MMT. The proposed translational study is innovative in two key respects. First, it will address the problem of confounding by indication of prior studies by using an existing large database-from a state with virtually unrestricted Medicaid coverage for BMT and MMT-with supplemental confounder information from a clinical patient population. Second, it will develop novel methods for using the Medicaid database for ongoing surveillance of prenatal opioid agonist therapy exposure and infant outcomes in a larger nationwide population.