Overturning the ACA’s Medicaid Expansion Would Likely Decrease Low-Income,Reproductive-Age Women’s Healthcare Spending and Utilization

Abstract

In late 2020, the Supreme Court began hearing a case challenging the Affordable Care Act (ACA), which led to coverage gains for many low-income, reproductive-age women. To explore potential implications of a full ACA repeal for this population, we examined gains experienced after Medicaid expansion, assuming that such gains may be reversed. Using restricted 2013 to 2014 data from the Medical Expenditure Panel Survey for 1190 women ages 18 to 44 with household incomes below 138% of the federal poverty level, we compared the change in healthcare spending and utilization for women living in expansion states to the change in non-expansion states using a difference-in-differences design. We found that if Medicaid expansion were overturned, Medicaid coverage is likely to decrease, as well as Medicaid spending and prescription drug utilization.

What do we already know about this topic?

For low-income, reproductive-age women, the ACA was associated with increases in coverage, access, and use of preventive care and contraceptives.

How does your research contribute to the field?

We find that the ACA was also associated with increases in Medicaid spending and prescription drug utilization for low-income, reproductive-age women, suggesting that the ACA’s coverage gains translated into meaningful increases in use of necessary services.

What are your research’s implications toward theory, practice, or policy?

Our research suggests that the ACA led to important gains in healthcare access for low-income, reproductive-age women, as measured by spending and utilization increases; thus, if the Supreme Court invalidates the entire ACA, a Medicaid expansion repeal would likely have significant adverse consequences for low-income, reproductive-age women’s access to care.

Introduction

In late 2020, the Supreme Court began hearing a case on the Affordable Care Act’s (ACA’s) individual mandate, with possible implications for the entire law. ¹ With a full ACA repeal, researchers have predicted that uninsured rates would increase for those with incomes below 200% of the federal poverty level (FPL) and adults ages 19 to 34. ² Many low-income, reproductive-age women fall into these categories and thus may face unique risks with an ACA repeal.

Reproductive-age women are most likely to have healthcare expenditures for pregnancy and mental health care, which represent 20% and 10% of spending, respectively. ³ Before the ACA, women faced access challenges, due to pre-existing condition exclusions, reduced availability of employer-sponsored coverage, and insurance affordability concerns. ⁴

Existing research has found that the ACA led to significant coverage gains for reproductive-age women.^5-14 The ACA was associated with a 7.4 percentage point decrease in the probability of being uninsured for this group. ⁸ A systematic review found that the ACA has resulted in improved access to care, preventive care use, and contraceptive use. ⁵ However, less is known about whether coverage changes translated into overall spending changes.

In this policy brief, we examined whether the ACA’s 2014 Medicaid expansions were associated with changes in low-income, reproductive-age women’s healthcare spending and utilization, to inform potential implications of a full ACA repeal.

Methods

To gauge possible impacts of an ACA repeal, we estimated the effects of the Medicaid expansions implemented in January 2014. We used the natural experiment created by the state expansion option. Our estimates were based on difference-in-differences analyses, comparing the outcome change in expansion states to the change in non-expansion states. This project was approved by the authors’ institutional review board.

Data and Sample

We used restricted Medical Expenditure Panel Survey (MEPS) Household Component Panel 18 data for 2013 to 2014, which included non-public state Medicaid expansion status flags. We identified women ages 18 to 44 with 2013 incomes below 138% of the FPL, the eligibility cutoff for non-disabled adults without dependents if a state expanded Medicaid. Based on their 2013 state of residence, we classified women into living in expansion or non-expansion states, defined as of 01/01/2014 and excluding early expansion states. ¹⁵ We used 2013 income and state of residence, rather than 2014, because a 2014 sample may be biased by changes from the ACA itself. Pregnant women were included in coverage analyses, since the ACA may have had a “woodwork” effect to increase enrollment among previously eligible women, but excluded in spending analyses since we expected spending patterns to differ between pregnant and non-pregnant women. ¹⁶

Outcomes

Our outcomes included annual insurance coverage, annual per-person total healthcare spending, Medicaid spending, and out-of-pocket spending. We also examined the number of prescription fills as a utilization measure.

Statistical Analysis

We conducted difference-in-differences analyses using linear models, comparing the spending and utilization changes in expansion states to respective changes in non-expansion states. To assess our parallel trends assumption, we relied on prior work examining coverage changes associated with the ACA for reproductive-age women, and we examined Medicaid spending per age 19 to 64 enrollee for 2010 to 2013 from the Medicaid and CHIP Payment and Access Commission (Supplemental Appendix Exhibit 1). ¹⁷

We also conducted descriptive analyses using multinomial logistic and multivariate logistic models to assess coverage changes by payer and race/ethnicity.

Models included age, race/ethnicity, marital status, education, health status, and pregnancy status. Analyses were conducted in Stata using longitudinal MEPS survey weights and robust standard errors at the CFACT Data Center in Rockville, Maryland, due to the data’s restricted status.

Limitations

While MEPS is nationally representative, the number of respondents who were low-income, reproductive-age women was limited. Though MEPS is less reliable than claims for spending, it allowed us to identify uninsured women and their spending. Our study also included only 2 years, so we examined the short-term impact of the ACA’s 2014 provisions, not other provisions, and we were not able to directly assess for parallel pre-trends in our data. As an observational study, our estimates cannot be interpreted as causal, due to potential confounding from historical trends; however, our study design of difference-in-differences helped reduce such concerns about unobserved confounding. In addition, Michigan and New Hampshire, which expanded later in 2014, were included as non-expansion states, which may have biased our results to the null. Finally, our results may not generalize to later Medicaid expansions.

Results

Supplemental Appendix Exhibit 2 presents sample characteristics of the 1190 low-income, reproductive-age women.

Coverage Increased, Especially in Medicaid Expansion States

In our multinomial logistic model, 38.7% of uninsured, low-income, reproductive-age women in 2013 gained Medicaid in 2014 in expansion states, compared to 19.5% in non-expansion states (P < .01, Supplemental Appendix Exhibit 3).

In our multivariate logistic model, Hispanic and Black women were less likely to gain Medicaid coverage, compared to White women (12.0% for Hispanic and 13.4% for Black, compared to 42.8% for White; P < .01 and P < .05, respectively, Supplemental Appendix Exhibit 4), which may be due to a higher proportion of Hispanic and Black women on Medicaid pre-expansion, or structural barriers to enrollment and racism.

Spending Increased for Medicaid

Medicaid expansion was associated with a differential increase in annual per-person Medicaid spending of $771 from 2013 to 2014, compared to non-expansion in our difference-in-differences analysis (P < .05, Table 1). While our estimates were also consistent with reduced out-of-pocket spending and increased overall spending, the estimated impacts were not significant.

Table 1.

Effect of ACA Medicaid Expansion on Medical Spending and Utilization among Low-Income, Reproductive-Age Women.

	Total	Out-of-pocket	Medicaid	Private
	Any spending (% points)
Dependent variable	(1)	(2)	(3)	(4)
Expansion × post	−0.57	0.88	1.28	−1.2
Mean for expansion state in pre-period	77.1	59.6	40.8	18.2
(SD)	(3.89)	(4.81)	(3.69)	(4.11)
P-value	P = .88	P = .85	P = .73	P = .76
Confidence interval	(−8.2, 7.1)	(−8.6, 10.4)	(−6.0, 8.6)	(−9.3, 6.9)
	Amount of spending ($)
Dependent variable	(5)	(6)	(7)	(8)
Expansion × post	568	−79	771	−127
Mean for expansion state in pre-period	2287	271	1177	443
(SD)	(549)	(92)	(383)	(244)
P-value	P = .30	P = .39	P = .046*	P = .61
Confidence interval	(−516, 1562)	(−260, 103)	(13, 1530)	(−608, 355)
		Utilization
		Prescription fills (#)
Dependent variable	(9)
Expansion × post	2.76
Mean for expansion state in pre-period	6.8
(SD)	(1.09)
P-value	P = .012**
Confidence interval	(0.61, 4.90)

Source. Authors’ analysis of 1004 non-pregnant women ages 18 to 44 with incomes below 138% of the federal poverty level using data from the Medical Expenditure Panel Survey Household Component Panel 18.

Note. Amount of spending analyses were calculated only including those with any spending during the 2 years. Prescription fills included refills. Difference-in-differences analysis compared the change in outcome for women living in expansion states to the change in outcome for women living in non-expansion states. Covariates included were age, race/ethnicity, education, marital status, and health status. Pregnant women were excluded because they have different overall health care use patterns than non-pregnant women.

Utilization of Prescription Drugs Increased

Medicaid expansion was associated with a differential increase of 2.8 refills in expansion states from 2013 to 2014, compared to non-expansion states in our difference-in-differences analysis (P < .05, Table 1).

Conclusion

We found that Medicaid expansion was associated with increased Medicaid coverage that did translate into increased Medicaid spending. We also found increased prescription drug utilization, consistent with prior work. ¹⁸ Under the assumption that an ACA repeal would yield opposite effects, our results suggest that overturning the ACA would likely increase the uninsured rate among low-income, reproductive-age women and decrease their healthcare utilization and spending.

Supplemental Material

sj-pdf-1-inq-10.1177_0046958020981462 – Supplemental material for Overturning the ACA’s Medicaid Expansion Would Likely Decrease Low-Income, Reproductive-Age Women’s Healthcare Spending and Utilization

Supplemental material, sj-pdf-1-inq-10.1177_0046958020981462 for Overturning the ACA’s Medicaid Expansion Would Likely Decrease Low-Income, Reproductive-Age Women’s Healthcare Spending and Utilization by Lucy Chen, Richard G. Frank and Haiden A. Huskamp in INQUIRY: The Journal of Health Care Organization, Provision, and Financing