Research

Job Market Paper

The Indirect Effects of Medicaid Payment Changes: Evidence of Spillovers to the Commercially Insured

(Download)

+ Show/Hide Abstract

This paper examines the spillover effects of a Medicaid payment change to the commercially insured population. I focus on a Medicaid policy change implemented in four states between January 1, 2014 and January 1, 2015 that discontinued payment for Early Elective Deliveries (EEDs), a low-value mode of childbirth defined as a scheduled, non-medically necessary induction or cesarean section (c-section) before 39 weeks’ gestation. The aim of the policy was to reduce EEDs, which are convenient for the physician and patient, but are not clinically beneficial. While previous evidence has shown significant reductions in EEDs in the Medicaid population, no studies have examined indirect effects of the policy in the non-Medicaid population. I use a difference-in-differences approach to assess whether the change in Medicaid payment policy impacted physician behavior across all payers. I find that all-payer EEDs declined 3.35% more in states with the Medicaid payment policy compared to control states with no measures aimed at reducing EEDs. Effects on EEDs were also larger in treatment states relative to states with other initiatives to reduce EEDs, including a voluntary “hard stop” policy and a Medicaid bonus payment. I find greater effects in areas with a higher share of for-profit hospitals, consistent with a stronger response to financial incentives where physician and hospital objectives align. I test for the presence of physician-induced demand in response to the Medicaid payment change, but find no evidence of this behavior.

Dissertation Papers in Progress

Place-Based Disparities in Maternal Care: The Role of Episode-Based Payments

+ Show/Hide Abstract

Black mothers experience 3.7 times more deaths from pregnancy-related complications and inadequate prenatal and postpartum care compared to White mothers. In Arkansas, low-risk women in metropolitan statistical areas (MSAs) with a high proportion of Black patients receive 9.6% more unnecessary cesarean sections than women in MSAs with a high share of White patients (27.4% versus 17.8%, respectively), suggesting that physician behavior is a potential contributor. Despite the increasingly salient role of physician decision making on racial disparities, little is known about effects of potential policy solutions. Using the 2010 to 2016 Truven MarketScan Commercial Claims database, I assess the impact of mandatory bundled payments for perinatal episodes on place-based racial disparities, in three types of perinatal services: (1) low-risk c-sections, (2) appropriate prenatal screenings, and (3) timely postpartum visits. I use Truven MarketScan Commercial Claims in a difference-in-differences analysis to compare changes in perinatal quality in the treatment state (Arkansas) versus a group of control states created via multiple propensity score weights. The effects of payment reform on racial disparities has only been assessed in clinical settings that rely on a binary indicator for treatment receipt. Maternal care allows me to test whether Black patients receive inferior treatment across a range of treatment options. This study aims to refine our understanding of the role of physician incentives on disparities. This is a critical question given inequalities in maternal care and the potential for payment policies to close or widen this gap. Results show significantly greater improvements in quality among areas with a high proportion of White patients in Arkansas, after policy implementation. Areas with a high White population experienced a 48% reduction in low-risk c-sections relative to MSAs with a high Black population (OR: 0.62). These MSAs also experienced relative improvements in prenatal screenings and postpartum visits. Results suggest that additional policy solutions with an equity focus may be needed to reduce racial disparities.

Do Effects of Mandatory Bundled Payments Vary by Baseline Quality? Evidence From Perinatal Care in Arkansas.

+ Show/Hide Abstract

In a difference-in-differences analysis using the Truven MarketScan Commercial Claims Database from 2010-2016, I test whether baseline quality is associated with heterogeneity in quality and spending effects of a mandatory bundled payment program for maternity care. Providers with lower quality face a higher marginal cost to change behavior, but also have greater room for improvement under value-based incentives than providers with higher quality. The analysis also explores whether the mechanism for savings (i.e., changes in price and/or quantity) varies between low vs. high quality providers. This paper seeks to inform whether the effects of mandatory payment reform differ from voluntary policies, and whether the approach to achieving savings is different, to inform the design of value-based reimbursement policies.