323 - The Impact of State Medicaid Expansion for Pregnant Women on Medicaid Utilization and Premature Birth Risk

Poster Number: 323
Publication Number: 323.1380

Background: Medicaid programs have variably expanded eligibility as percent of the federal poverty level (FPL) during pregnancy for decades. While Medicaid expansion for pregnant women has been found to increase coverage and access, the impacts of expansion on insurance, prenatal care, and risk of premature birth are unclear.

Objective: To estimate the impact of Medicaid eligibility expansion on Medicaid coverage for prenatal care and birth and on the risk of premature birth.

Design/Methods: We utilized data from the Pregnancy Risk Assessment Monitoring System (PRAMS), a stratified sample of eligible birth records from participating states which is combined with self-reported maternal and infant data. We utilized 2 state dyads: Ohio (150 to 200% of the FPL in 2008) vs Pennsylvania (185% of the FPL 2007-2010) and Wisconsin (185% to 250% of the FPL in 2009) vs Michigan (185% of the FPL 2008-2009). We used a differences-in-differences paradigm to estimate the Medicaid expansion effect with logistic regression for Medicaid utilization and ordinal logistic regression for gestational age (GA), adjusting for relevant factors. PRAMS analysis weights were utilized to make estimates State representative.

Results: Prior to Medicaid expansion, Ohio had more non-Hispanic Black mothers and Pennsylvania had higher family income; Wisconsin had higher maternal education and marriage and Michigan had more WIC participation and non-Hispanic Black mothers (Table 1). At Baseline, the distribution of GA was similar in Ohio vs Pennsylvania but lower in Michigan vs Wisconsin (Table 2). Medicaid expansion was associated with an increase in Medicaid coverage for birth in the Ohio vs Pennsylvania dyad on univariate analysis (OR 1.42, 95% CI 1.03-1.94) and multivariate analysis (OR 1.65, 95%CI 0.99-2.75); in Wisconsin compared to Michigan, no effect was found on univariate analysis but multivariate analysis found increases in Medicaid coverage for both prenatal care (OR 1.73, 95% CI 1.19-2.51) and birth (OR 1.79, 95% CI 1.23-2.59) (Table 3). Medicaid expansion was not associated with a significant change in GA distribution or preterm birth in either dyad (Table 2).

Conclusion(s): Unadjusted, Medicaid expansion was associated with increased utilization in only the low-income dyad, but, after accounting for income, increased use was observed with a greater effect in the high-income dyad. Medicaid expansion appears most effective when targeted at lower income patients. However, after accounting for income and likely-associated private insurance coverage, higher income patients are equally or more likely to uptake Medicaid if offered and needed.