Medicaid Accountable Care Organizations and Childbirth Outcomes

Introduction: States are considering adoption of the Accountable Care Organization (ACO) model to transform their Medicaid programs. Few studies have addressed the impact of Medicaid ACOs on costs and outcomes. Many Medicaid enrollees are pregnant women, some of whom have mental health conditions and substance use disorders that require intensive services. ACOs have the flexibility to invest in novel and cost-effective prenatal care programs that address social determinants, such as group prenatal care programs that are oriented toward preventing premature births and improving maternal outcomes. This study analyzes the impact of Medicaid ACOs on maternal and neonatal outcomes. Methods: This study used inpatient hospital data (SID) from the Healthcare Cost and Utilization Project (HCUP) from 2008 to 2015. Maternal and neonatal outcomes from three states, Colorado, Oregon, and New Jersey, implementing Medicaid ACO programs were compared to three adjacent states without these models: New Mexico, Washington, and New York. Following use of a coarsened exact matching approach to filter the sample to comparable discharges from the ACO and non-ACO states, difference-in-differences regression models were estimated to study Medicaid ACO impacts on rates of infant mortality, neonatal intensive care unit utilization, low birthweight, uncomplicated C-section delivery, and severe maternal morbidity, as well as cost per birth. Results: On average, Medicaid ACOs were associated with a $535 reduction in cost per birth (p=0.01). Medicaid ACOs were associated with an estimated 0.31 percentage point reduction in C-section probability on average, although this result was not statistically significant (p=0.38). Medicaid ACOs were associated with a 0.41 percentage point reduction in the probability of severe maternal morbidity in Colorado (p<0.05) and a 1.0 percentage point increase in NICU usage (p=0.38), although the latter estimate was not statistically significant. In addition, Medicaid ACOs were associated with a 0.74 percentage point increase in the probability of low birthweight (p=0.04). Conclusion: Medicaid ACOs appear to be more successful at reducing childbirth costs than improving key birth outcomes. Affecting these outcomes may require a deeper intervention or a longer amount of time. The association uncovered between Medicaid ACOs and increased rates of low birthweight may suggest that Medicaid ACOs have improved prenatal care such that more fetuses survived to birth but were born preterm. Despite having leadership and funding to support new initiatives, Medicaid ACOs still face challenges, including lack of provider engagement and information systems that do not communicate.