(791) Alternative payment models for pregnancy care: a difference-in-difference analysis among state Medicaid programs

Friday, January 31, 2025

10:30 AM – 12:30 PM

Submitting Author and Presenting Author(s)

Jordan S. Stone, MD

Fellow
Hospital of the University of Pennsylvania
Philadelphia, Pennsylvania, United States

Coauthor(s)

AC

Arina Chesnokova, MD, MPH

Hospital of the University of Pennsylvania
Philadelphia, Pennsylvania, United States
AS

Allison Schachter, BA

University of Pennsylvania
Philadelphia, Pennsylvania, United States
Sindhu K. Srinivas, MD, MSCE

Professor
University of Pennsylvania Perelman School of Medicine
Philadelphia, Pennsylvania, United States
SM

Sunni Mumford, PhD

University of Pennsylvania Perelman School of Medicine
Philadelphia, Pennsylvania, United States

Objective: There is growing interest in using payment model design to improve pregnancy care delivery and reduce morbidity and mortality, particularly within Medicaid. Despite this, data on alternative payment models (i.e., those that establish a budget for a care episode, coupled with quality incentives) in pregnancy are sparse. The objective of this study is to evaluate whether the introduction of a mandatory Medicaid bundled payment program for pregnancy care was associated with changes in antenatal care delivery and maternal morbidity.

Study Design: This is a quasi-experimental difference-in-difference (DiD) analysis comparing two states that implemented mandatory Medicaid bundled payment programs for pregnancy care (TN in 2011 and OH in 2014) to control states (GA, KS, KY, OK). Demographic variables and pregnancy outcomes were collected from birth certificate records between 2009 and 2019. Inclusion criteria were births reimbursed by Medicaid in intervention or control states with concordance between state of residence and state of birth. A DiD analysis was conducted for co-primary outcomes—gestational age at entry into prenatal care and maternal transfusion at time of delivery. These were chosen as proxies for modifiable aspects of pregnancy care within the domains of antenatal care delivery and maternal morbidity. The extended two-way fixed effects estimator was used with the heterogenous DiD regression to account for staggered rollout in the intervention group.

Results: Between 2009 and 2019, we identified 2,227,136 births that met inclusion criteria. 55% were in the control group and 45% were in the intervention group. The pre-intervention parallel trends assumption was satisfied between groups. In the heterogenous DiD regression model, no significant difference-in-differences were identified for the specified outcomes between the two groups (see table).

Conclusion: Mandatory Medicaid bundled payment programs did not have a significant effect on gestational age at entry into prenatal care or maternal transfusion rates. Additional studies are needed to further evaluate the impact of these models.