Racial Inequities in Postpartum Coverage During Medicaid Continuous Coverage: Evidence from ACA Expansion vs Non-Expansion States

Abstract

Introduction:

Extending Pregnancy Medicaid coverage to 12 months postpartum is a key strategy to reduce maternal morbidity and mortality and may be most impactful among states that did not expand Medicaid as part of the Affordable Care Act (ACA). The continuous Medicaid coverage provision enacted at the onset of the COVID-19 pandemic provides a natural experiment to test the impact of postpartum Medicaid extension policies.

Methods:

A cross-sectional pre-post design was used to estimate changes in postpartum insurance after the introduction of continuous Medicaid coverage. Using 2016–2023 American Community Survey (ACS) data, non-Hispanic Black and non-Hispanic White respondents aged 19–50 who gave birth in the previous 12 months were included (n=157,016). Postpartum insurance was categorized as Medicaid, private, or uninsured. Linear regression models compared expansion and non-expansion states within Black and White respondents for each year pre- and post-continuous coverage, with 2019 as the reference year. Covariate-adjusted regressions included respondents’ age, employment status, and household income.

Results:

Pre-pandemic uninsurance rates among Black women were 16.5% in non-expansion states and 6.4% in expansion states, and among White women they were 11.1% and 5.4%, respectively. Among Black women, uninsurance decreased after continuous coverage in non-expansion states only. For example, in 2023, Black postpartum uninsurance decreased in non-expansion states 5.2 percentage points (pp) more (95%CI= −8.9, −1.5) than in expansion states; this was driven by a 3.9pp larger gain in Medicaid in non-expansion states than expansion states (95%CI= −1.8, 9.6). Among White women, uninsurance decreases in 2021–2023 were also larger in non-expansion states but were driven by gains in private insurance. As a result, Black-White disparities in postpartum uninsurance remained unchanged.

Conclusions:

Postpartum uninsurance decreased among Black women in non-expansion states after continuous coverage, but racial inequities persisted.

Introduction

Black women in the United States (US) experience persistently high rates of maternal morbidity and mortality. In 2022, Black women had two times the rate of pregnancy-related death within 42 days of delivery than non-Hispanic White women¹, and were more than four times more likely to die of pregnancy-associated deaths up to 12 months postpartum.² Maternal death, however, is only the tip of the iceberg, with increasing evidence that the postpartum period exacerbates racial health inequities more broadly. For example, in the postpartum period, Black women have elevated rates of severe maternal morbidity,³ postpartum depression,⁴ hospital readmissions and Emergency Department visits,⁵ all of which are associated with poorer access to health care.

The postpartum period is one of intense physiologic and psychosocial adaption, requiring monitoring of health to meet the needs of women in this critical period.⁶ For example, postpartum health care provides depression screening and treatment,⁷ family planning,⁸ behavioral health,⁹ and breastfeeding support.¹⁰ Access to postpartum health care is also crucial for connection to primary or specialty care to follow up on any conditions arising during pregnancy, such as hypertension or diabetes, to prevent later disease.¹¹ Evidence suggests that insurance coverage increases access to postpartum care^12,13 and contraception,^14,15 and women who lose insurance postpartum are more likely to have unmet need for health care.¹⁶ Thus, access to postpartum care is a key lever to improve Black maternal health.

One policy response to the maternal health crisis is extending Pregnancy Medicaid from 60 days to 12 months postpartum, with the goal of increasing health care access for women who qualified for Medicaid only due to pregnancy, but who previously would lose care at 60 days postpartum – a time crucial to preventing postpartum morbidity and mortality.¹⁷ 49 states have passed Medicaid postpartum extensions into law under the American Rescue Plan Act (ARPA), making it one of the single biggest policy responses to the crisis.¹⁸ However, the Families First Coronavirus Response Act (FFCRA) enacted by Congress at the onset of the COVID-19 emergency period jump-started these efforts through its continuous Medicaid coverage provision, which provided de facto postpartum Medicaid extensions to those with pregnancy Medicaid.

A key aspect to the expected impact of postpartum Medicaid extensions on Black maternal health is a state’s pre-policy level of postpartum uninsurance. The expansion of Medicaid as part of the Affordable Care Act (ACA) of 2014 broadened eligibility to nearly all low-income adults by increasing income thresholds; by 2023 40 states plus DC expanded Medicaid.¹⁹ As a result, many states increased the number of women entering pregnancy enrolled in traditional Medicaid, reducing the number of women with pregnancy Medicaid losing insurance at 60 days postpartum.²⁰ Another result was that more Black women were uninsured in the postpartum period in non-expansion states than in expansion states.²¹ Thus the ability of postpartum Medicaid extension to improve maternal health equity may vary by state Medicaid expansion status, with the policy most likely to have the largest impact in states that did not participate in the ACA Medicaid Expansion.

The goal of this study was to evaluate changes in postpartum uninsurance, Medicaid insurance, and private insurance before versus during the continuous coverage provision periods among Black and White women, by ACA Medicaid expansion status. Previous analyses have reported gains in postpartum Medicaid coverage during the continuous coverage provision but did not explicitly investigate the equity impact while investigating the intersection with ACA expansion.^22,23 The primary hypothesis was that the largest gains in postpartum uninsurance among Black postpartum women occurred in states that did not expand Medicaid under the ACA. The secondary hypothesis was that this would shrink overall racial inequities in postpartum uninsurance, particularly in non-expansion states.

Methods

Study Sample

A cross-sectional pre-post design was used to estimate changes in postpartum insurance after the introduction of continuous Medicaid coverage, while examining the moderating influence of race and ACA expansion status.

American Community Survey (ACS) data were obtained from years 2016–2019 and 2021–2023 for 50 states and District of Columbia, obtained from IPUMS USA.²⁴ 2016 was chosen as the start year because most ACA expansion states had completed ACA implementation, and 2023 was the end year because the continuous coverage provision ended in 2023. Four states that expanded Medicaid under the ACA during the intervention period (Idaho, Utah, Nebraska, Oklahoma) were excluded. The Program for the Protection of Human Subjects at Icahn School of Medicine at Mount Sinai and the Columbia University Human Research Protection Office approved this study.

People aged 19–50 who answered “yes” to “Has this person given birth to any children in the past 12 months?” were included (n=196,639), all of whom were women. People <19 years were not included because they would already be eligible for continuous Medicaid coverage due to child eligibility rules in the Children’s Health Insurance Program (CHIP) or Medicaid, although some fraction may have benefited by maintaining eligibility despite aging out of CHIP. The sample was further restricted to non-Hispanic Black and non-Hispanic White women, for a total unweighted analytic sample size of n=157,016.²⁵ The focus of this study is on Black women, due to the disproportionate maternal mortality burden. The sample size was insufficient in the “American Indian and Alaskan Native” group for analysis, although this group also has heightened maternal mortality burden. The rationale for choosing non-Hispanic White women as the reference group is that historically they have a position of privilege and the lowest rates of maternal mortality. Race was conceptualized as a social construct and a proxy for intergenerational experiences of oppression and privilege. ACS data are cleaned and missing variables imputed by allocation or assignment prior to release.

The continuous coverage provision under FFCRA was in effect March 18, 2020-March 31, 2023. The pre-period was 2016–2019 (n=91,622) and post-period was 2021–2023 (n=65,394). Data from 2020 were not included because data quality issues with the ACS in 2020 are well-documented, and it is advised not to use these data.²⁶ Also, because the ACS asks about births in the past 12 months, some proportion of respondents in 2020 would have given birth prior to March 18, 2020, making them ineligible for continuous coverage. When the public health emergency ended on March 31, 2023, FFCRA’s continuous coverage also ended. However, states with postpartum extension legislation were able to segue straight to the new policy. Therefore, for most states, the FFCRA and legislated postpartum extension policy periods are indistinguishable in 2023.

Measures

The three insurance coverage outcomes were: uninsurance, Medicaid, and private. The rationale for examining these three coverage outcomes is to view substitution patterns between coverage sources. Insurance status was ascertained from the ACS question, “Is this person currently covered by any of the following types of health insurance or health coverage plans?”. Mutually exclusive insurance types were created based on a hierarchy used by Kaiser Family Foundation as follows: Medicaid, Medicare, employer, military (Veterans Affairs, TRICARE), and non-group (purchased directly from insurance companies).²⁷ Three binary variables for self-reported health insurance status were: “uninsurance” (1=zero coverage or Indian Health Services; 0=any insurance), “Medicaid” (1=Medicaid; 0=uninsured, Medicare, employer, military, and non-group), and “private” (1=employer, military, non-group; 0=Medicaid, Medicare).

Covariates were chosen based on prior knowledge of respondent characteristics that may have influenced insurance status besides the intervention and may be time-variant. Covariates included age, employment status (employed, unemployed, not in labor force), marital status, and household income as a % of the Federal Poverty Level (FPL).

Statistical analysis

The study team first descriptively examined all three insurance outcomes by race and expansion status over time. The survey year (2016–2023) was the time variable because no smaller time variable is available in the ACS. Next, an event study approach was used which compared each survey year in the pre- and post-periods to 2019 as a reference year (see Appendix Table 7 for equations).²⁸ The event study approach was chosen to identify per-year policy effects and allow for formal testing of pre-trends.²⁹ The intent was to identify per-year policy effects because awareness of the policy may have increased over time. The causal contrasts of interest were each post-policy year relative to the reference year of 2019, assuming no simultaneous confounding events and no pre-policy trends. However, because all states were exposed to the continuous Medicaid coverage provision, a treatment vs. control contrast was not possible. The ACS survey sampling person-level weights were used as recommended by IPUMS.

Fixed effects linear probability regression models were estimated. A robust standard error accounted for potential correlation of error terms within states due to unobserved state-level policies or events. Differences in frequencies of insurance outcomes pre- and post-continuous coverage among Black postpartum women were calculated. The interaction between the continuous coverage provision and ACA expansion was tested by comparing pre-post insurance changes among Black women in Medicaid expansion and non-expansion states. Models were repeated among White women. Next, pre-post differences in Black-White insurance coverage disparities were modeled by expansion status. Finally, a triple interaction term estimated the equity impact of the policy, comparing the joint-policy difference (continuous coverage and ACA expansion) coefficients among Black versus White postpartum respondents. After these event study models showed heterogeneity in the coefficients in the continuous coverage period, it was deemed inappropriate to present models averaging the pre- and post-periods and thus present yearly coefficients for all models. 95% confidence intervals (alpha=0.05) were calculated for all coefficients.

As a sensitivity analysis, 11 states were excluded (N=47,264) (See Appendix Table 1) that did not implement postpartum Medicaid extension legislation following the termination of continuous coverage. These states underwent an “unwinding” period in which postpartum Medicaid beneficiaries who were no longer eligible were disenrolled.

Results

Our total unweighted analytic sample size was n=157,016. Sample characteristics differed at baseline across study groups (Table 1, Appendix Table 2). For example, percent with < high school degree, percent unemployed, and percent with household income < 200% FPL are significantly higher among Black vs White respondents. Percent with household income <200% FPL was also higher in non-expansion vs expansion states. Characteristics were mostly stable across time, with small changes. For example, unemployment increased slightly across all groups from 2019–2021, then decreased in 2022–2023.

Table 1.

Sample Characteristics by Race, Year^a and Medicaid Expansion Status under the Affordable Care Act, 2019, 2021–2023

	Non-Expansion States				Expansion States
	2019	2021	2022	2023	2019	2021	2022	2023
Non-Hispanic Black
N (Unweighted)	1,378	1,318	1,353	1,325	1,656	1,594	1,557	1,321
N (Weightedb)	227,946	229,224	236,480	243,660	286,907	273,398	289,806	258,143
Age (mean)	29.9	30.8	31.2	31.4	329.5	30.0	30.2	30.5
	sd=6.4	sd=6.5	sd=6.6	sd=6.6	sd=63	sd=6.8	sd=6.5	sd=6.4
Currently not married	65.9%	67.0%	60.2%	62.5%	63.4%	59.9%	59.5%	60.3%
High school or less	47.3%	49.3%	45.8%	47.1%	48.2%	49.3%	45.3%	49.9%
Employment Status
Employed	66.8%	64.5%	65.2%	67.5%	65.2%	62.2%	63.8%	63.0%
Unemployed	8.7%	9.5%	9.6%	7.9%	8.7%	10.4%	9.3%	7.9%
Not in labor force	24.5%	26.1%	25.2%	24.6%	26.1%	27.3%	26.9%	29.1%
	189.0	195.7	200.6	206.7	200.3	216.1	201.7	208.1
Household income as % of FPL	sd=152.0	sd=156.1	sd=152.7	sd=152.5	sd=156.7	sd=166.5	sd=164.4	sd=161.9
Non-Hispanic White
N (Unweighted)	5,791	5,341	5,910	5,870	13,573	12,918	13,473	13,414
N (Weightedb)	571,356	569,572	598,748	580,759	1,328,217	1,331,708	1,324,206	1,335,716
Age (mean)	30.8	30.9	30.8	31.2	31.3	31.6	31.8	31.7
	sd=6.1	sd=6.0	sd=6.0	sd=6.0	sd=6.0	sd=6.0	sd=6.0	sd=6.0
Currently not married	25.6%	22.9%	21.6%	20.8%	23.8%	22.9%	22.1%	22.3%
High school or less	33.5%	30.9%	31.0%	47.1%	30.7%	29.6%	28.9%	30.0%
Employment Status
Employed	63.41	65.23	65.08	66.61	67.01	66.78	68.53	70.37
Unemployed	2.77	2.74	2.57	2.32	2.56	3.74	2.39	2.33
Not in labor force	33.83	32.03	32.35	31.08	30.42	29.48	29.08	27.31
	299.9	306.9	308.2	312.9	319.6	323.3	324.7	326.7
Household income as % of federal poverty level	sd=166.1	sd=168.2	sd=164.7	sd=162.9	sd=170.6	sd=170.3	sd=168.0	sd=167.6

^a2020 data excluded in study design

^bACS person weight was applied.

Figure 1 and Appendix Table 3 show postpartum uninsurance, Medicaid, and private insurance trends for Black and White women living in expansion and non-expansion states. Before continuous enrollment, postpartum uninsurance rates were higher in Black women in non-expansion states than in any other group. In 2019, the rate of postpartum uninsurance among Black women in non-expansion states was 16.5%, compared to 11.0% among White women. In expansion states, postpartum uninsurance rates were 6.4% and 5.4% for Black and White women, respectively. While uninsurance declined in non-expansion states after 2019 among both racial groups (from 16.5% in 2019 to 11.8% in 2023 for Black women and 11.0% to 7.0% among White women), it remained relatively constant in expansion states during those years. Notably, what drove declines in uninsurance in non-expansion states appears to differ for Black versus White women. For Black individuals in non-expansion states, decreasing uninsurance was driven by large increases in Medicaid (from 37.8% in 2019 to 42.5% in 2023, with a high of 45.3% in 2021). In contrast, for White individuals in non-expansion states, uninsurance decreases were driven by small gains in both Medicaid (from 19.4% to 20.4% in 2023) and private insurance (69.3% in 2019 to 72.1% in 2023).

Figure 1. Trends in postpartum insurance coverage in Affordable Care Act expansion and non-expansion states, by race, 2016–2023.

Figure 2, Appendix Table 4, and Appendix Table 5 report coefficients comparing non-expansion states to expansion states for each year 2016–2023 relative to 2019, stratified by race. Among both Black and White women, differences between expansion groups in postpartum uninsurance were similar in the pre-continuous coverage period.

Figure 2. Differences^a in insurance pre-post continuous Medicaid policy in non-Affordable Care Act expansion states compared to Affordable Care Act expansion states, by race, 2016–2023 (Reference Year: 2019).

^aCoefficients adjusted for age, marital status, employment status, and household income as % of federal poverty level.

PANEL A: Non-Hispanic Black

PANEL B: Non-Hispanic White

Table 2 compares Black-White differences in postpartum uninsurance for each year relative to 2019, by states’ expansion status. The null coefficients for 2021–2023 show that Black-White disparities in postpartum uninsurance did not change during continuous coverage. For example, in non-expansion states, the Black-White disparity difference in 2021 was −0.3 (95%CI=−7.1, 6.6) pp, −0.5 (95%CI= −4.0, 3.0) pp in 2022, and −0.6 (95%CI=−3.5, 2.2) pp in 2023.

Table 2.

Pre-post continuous Medicaid policy differences in racial inequity in postpartum uninsurance, by states’ Affordable Care Act expansion status, 2016–2023

	Non-expansion States	Expansion States	Non-expansion States	Expansion States
	Black vs. White pre-post differences [95% CI]	Black vs. White pre-post differences [95% CI]	Adjusteda Black vs. White pre-post differences [95% CI]	Adjusteda Black vs. White pre-post differences [95% CI]
N (weighted)	50591	106425	50591	106425
2016	0.003 [−0.056,0.063]	0.004 [−0.024,0.033]	0.005 [−0.052,0.063]	0.004 [−0.024,0.033]
2017	0.01 [−0.045,0.065]	0.014 [−0.013,0.042]	0.009 [−0.046,0.063]	0.014 [−0.013,0.040]
2018	0.012 [−0.058,0.082]	0.019 [−0.006,0.043]	0.011 [−0.058,0.081]	0.018 [−0.006,0.043]
2019	Referent	Referent	Referent	Referent
2021	−0.003 [−0.071,0.066]	0.011 [−0.010,0.032]	−0.004 [−0.070,0.062]	0.011 [−0.009,0.032]
2022	−0.005 [−0.040,0.030]	0 [−0.027,0.027]	−0.005 [−0.039,0.029]	−0.001 [−0.028,0.025]
2023	−0.006 [−0.035,0.022]	0.013 [−0.014,0.041]	−0.006 [−0.036,0.024]	0.01 [−0.018,0.038]

^aAdjusted for age, marital status employment status, household income as a % of federal poverty level

Appendix Table 6 shows the results of the DDD analysis, which tested if the differential in pre-post changes in racial inequity differed by expansion status. These 3-way interactions were of small magnitude and not statistically significant.

The sensitivity analysis removing states in which Medicaid disenrollment began in 2023 showed similar results (Appendix Figure 1). For example, the coefficient comparing 2023 vs. 2019 changes in uninsurance between expansion and non-expansion states among Black women changed from −0.052 to −0.059 when excluding states who underwent unwinding in 2023.

Discussion

This study examined the dual impact of existing and new US Medicaid policies on inequities in postpartum uninsurance. At baseline, states that did not expand Medicaid under the ACA had higher rates of postpartum uninsurance than those that did expand, and Black vs. White inequities were present primarily in non-expansion states. Trends from 2021–2023 showed decreased uninsurance rates in non-expansion states among Black women, but because decreases were similar among White individuals, the uninsurance disparity persisted. As hypothesized, Medicaid insurance increased primarily in non-expansion states, especially in 2021. However, these Medicaid gains happened concurrently with losses of private insurance among Black individuals and were not sustained through the end of the study period. While overall, findings supported the hypothesis that decreases in postpartum uninsurance among Black women following extension of Medicaid would be primarily in non-ACA expansion states, effect sizes were moderate and offset by changes in private insurance coverage, diminishing the equity impact.

This study adds to a growing literature on postpartum insurance coverage as a lever to reduce maternal morbidity and mortality. Previous analyses of ACS data showed increases in postpartum Medicaid insurance following FFCRA implementation among women in the eligibility gap.³⁰ Other researchers have also found gains in postpartum Medicaid coverage during FFCRA.^23,31–34 One of these studies also found greater gains in Medicaid insurance in non-expansion states during continuous enrollment using 2018–2021 Pregnancy Risk Assessment Monitoring System data, but did not examine racial inequities within these categories.²³ This analysis extends the study frame to 2023 and 46 states plus DC, demonstrating that gains were largely present among Black women in states that did not expand Medicaid. This new finding suggests that postpartum Medicaid extension policies may be insufficient to address the Black maternal health crisis in expansion states, while in non-expansion states, increased coverage is critical to increase access to care and improve outcomes.

Evidence from this study sheds light on the potential effectiveness of strategies to reduce maternal morbidity and mortality. Large differences in postpartum insurance at baseline in non-expansion states vs. expansion states were present, with 16.5% of Black women and 11.1% of White women uninsured in non-expansion states in 2019, compared to 6.4% and 5.4% in expansion states. These differences were reduced, although not eliminated, by continuous eligibility. This pattern underscores several points. First, the ACA Medicaid expansion is a critical part of health care access for postpartum women. Any plan to cut the Medicaid program³⁵ could have negative implications on access to care in pregnancy and postpartum, disproportionately affecting Black women, who are more likely to be insured by Medicaid, as shown in the results. Second, postpartum Medicaid extension policies are unlikely to have a substantial impact on racial disparities in postpartum maternal morbidity and mortality in expansion states, due to the relatively low percentage of uninsured Black women in these states. Therefore, states need to pursue additional policies to improve maternal health postpartum.

These findings also contribute insights into shifts from private insurance to Medicaid during the COVID-19 pandemic. Unemployment during the pandemic led to increases in new Medicaid enrollment that were larger in expansion states.³⁶ Black people were more likely than White people to become unemployed during the pandemic,³⁷ which is reflected in the results with the largest increases in unemployed status and largest losses of private insurance in 2021 in this group. Private insurance rebounded in 2022–2023 only among White women in non-expansion states. Because some of the women in the study sample may have been newly eligible and enrolled in Medicaid during the COVID-19 pandemic because of job loss, and this proportion was most likely greater in expansion states, estimates comparing Medicaid gains in non-expansion states vs. expansion states during continuous enrollment may be an underestimate of the true effect. Although adjustment for individual-level unemployment status may partially have accounted for this noise, evidence of shifts in unemployment and private insurance coverage in postpartum women should be considered when interpreting reports on postpartum insurance.

Limitations

Limitations exist in the study design. Because the continuous coverage provision applied to all states, only a pre-post study was feasible, with no “untreated” group for comparison. Also, 2023 was a transition year with the ending of continuous eligibility and the segue into new legislated postpartum extensions, a longer observation time may be needed to learn the eventual equity impact. The ACS question is asked about the past 12 months, so the data could not pinpoint which months postpartum coverage existed. Assuming that the distribution of months postpartum was similar in the pre- and post-periods, bias is unlikely. Another limitation is that insurance was self-reported. Previous research suggests that many people, including postpartum women, were unaware of their Medicaid coverage during continuous enrollment.^34,38 As a result, the reported increases in postpartum Medicaid coverage in the post-period may be an underestimate. It is unclear if this bias would be differential by race or expansion status. In Janevic et al, it was primarily Hispanic immigrant women who were unaware of their continued coverage, which is coherent with the findings of Ding et al., but one cannot rule out that differential reporting could further bias the reported findings toward the null. However, because coverage knowledge is crucial to seeking postpartum care, this self-reported measure is also a strength. Other strengths include examining three types of insurance coverage simultaneously for substitution effects, as well as a national focus.

Conclusions

In conclusion, the results of this study suggest that continuous Medicaid enrollment policies increased coverage among Black postpartum Medicaid beneficiaries, especially in states that did not previously expand Medicaid under the ACA. However, postpartum uninsurance was not eliminated, and racial inequities in uninsurance persisted. Overall, this study suggests that postpartum Medicaid extension plays an important role in improving access to postpartum maternal health care but it is not a sufficient policy intervention to address the US Black maternal health crisis.