Immigrant Inequities in Uninsurance and Postpartum Medicaid Extension: A Quasi-Experimental Study in New York City, 2016–2021

Abstract

Objectives. To determine if de facto postpartum Medicaid extension during the Families First Coronavirus Response Act (FFCRA) reduced immigrant versus US-born inequities in uninsurance.

Methods. We assessed self-reported uninsurance at 2 to 6 months postpartum among people with Medicaid-paid births using the New York City Pregnancy Risk Assessment Monitoring System (PRAMS), comparing immigrant and US-born people. We created a pre-FFCRA cohort of 2611 births from 2016 to 2019 and a post-FFCRA implementation cohort of 1197 births from 2020 to 2021. We calculated risk differences using log binomial regression.

Results. Self-reported postpartum uninsurance among immigrants decreased from 13.6% to 9.3% after FFCRA (adjusted risk difference = −4.9%; 95% confidence interval = −7.8%, −2.0%). Immigrant versus US-born inequities in postpartum uninsurance decreased except among Hispanic birthing people, among whom 1 in 6 reported they were uninsured during FFCRA, despite continued eligibility.

Conclusions. De facto postpartum Medicaid extension decreased immigrant inequities in insurance coverage, but Hispanic immigrants may have been unaware of continued coverage.

Public Health Implications. Postpartum Medicaid extension policies that are inclusive of all immigrants may decrease inequities, but community-integrated implementation is needed to raise awareness of coverage and advance postpartum maternal health equity. (Am J Public Health. 2025;115(5):732–735. https://doi.org/10.2105/AJPH.2024.307968)

In New York City (NYC), Hispanic and Black birthing people have 2 and 6 times the risk of pregnancy-related death as non-Hispanic White birthing people.¹ Inequities in maternal death extend into the postpartum period, and loss of Medicaid insurance after birth may contribute.² Lack of a federal requirement for Medicaid to insure undocumented and otherwise nonqualified immigrants means that immigrants are disproportionately uninsured in the postpartum period.³ In New York State, before 2020, undocumented immigrants were eligible for pregnancy Medicaid, but coverage ended at 60 days after birth.⁴ The Medicaid Maintenance of Eligibility in Families First Coronavirus Response Act (FFCRA) resulted in a de facto extension of Medicaid after birth.⁵ In New York State, no Medicaid recipient was disenrolled after birth during FFCRA, regardless of immigration status. Our objective was to estimate if immigrant versus US-born inequities in postpartum uninsurance declined during FFCRA, overall and by race/ethnicity.

METHODS

We used 2016–2021 data from the NYC Pregnancy Risk Assessment Monitoring System (PRAMS), a population-based survey administered via mail and phone to individuals with a recent live birth, identified through a stratified, random sample of birth certificate records.⁶ NYC PRAMS data are weighted to be representative of all live births to NYC residents. Over the study period (2016–2021), NYC PRAMS had an unweighted response rate of 61.7% and 8018 survey respondents. NYC PRAMS is unique to national PRAMS data because it links survey data to birth certificate details on birth country and years in the United States. We selected respondents with Medicaid-paid births, as ascertained by hospital staff (59.3%). We created a pre-FFCRA period of births (January 1, 2016–March 17, 2019; n = 2611) and a post-FFCRA implementation period of births (March 18, 2020–December 31, 2021; n = 1197). We considered March 18, 2019 to March 17, 2020 (n = 694) a “washout” period, so that our pre-FFCRA period included only those whose pregnancy-related Medicaid coverage would have expired prior to FFCRA implementation, whereas our post-FFCRA implementation period would include only those who gave birth during FFCRA and thus were “exposed” to continuous enrollment. We obtained self-reported race, ethnicity, country of birth, time in the United States, age, education, and parity from the birth certificate. We categorized self-reported postpartum insurance (none vs any) at survey response (2–6 months after birth) to ascertain “self-reported postpartum uninsurance” (henceforth “postpartum uninsurance”); we excluded from the analysis those with missing data on country of birth, race/ethnicity, or postpartum insurance status (3.2%). We created a 3-month time variable. Missingness was low for all variables, ranging from 0.0% (age) to 3.2% (postpartum insurance type).

We examined postpartum uninsurance graphically and sample characteristics by pre- and post-FFCRA implementation periods. We estimated marginal probabilities from logistic regression to calculate risk differences comparing the pre- and post-FFCRA implementation periods, adjusting for education, age, parity, and race/ethnicity. We used difference-in-difference (DID) parameterization to test immigrant versus US-born differences in postpartum uninsurance in the pre- and post-FFCRA implementation periods. We tested an interrupted time series approach; the time coefficient and post-FFCRA implementation period slopes were null, so we used DID models for simplicity. We used sampling weights and robust standard. We tested the parallel trends assumption visually, ran a placebo regression with a pre-FFCRA period of 2016 to 2017 and post-FFCRA implementation period of 2018 to 2019, and used alternate washout periods (October, 1 2019–March 17, 2020, March 18, 2019–January 16, 2020, and October 1, 2019–January 16, 2020).

RESULTS

Sociodemographic characteristics in the pre- and post-FFCRA implementation periods differed by years living in the United States (among immigrants), race/ethnicity, and education (Appendix Table A, available as a supplement to the online version of this article at http://www.ajph.org; P < .05 on χ² test). Appendix Figure A shows parallel trends between immigrants and US-born people in the pre-FFCRA period, followed by a decrease in postpartum uninsurance after FFCRA implementation.

Among immigrants, postpartum uninsurance decreased from 13.6% to 9.3% (adjusted risk difference [ARD] = −4.9%; 95% confidence interval [CI] = −7.8%, −2.0%), whereas among the US-born it changed from 1.2% to 0.7% (ARD = −0.5%, 95% CI = −5.6%, 4.6%; Table 1). Postpartum uninsurance decreased among all immigrant groups, but least among Hispanic immigrants, from 23.5% to 17.7% (ARD = −4.0%; 95% CI = −9.9%, 1.9; Table 1). Overall, the immigrant versus US-born disparity in postpartum uninsurance decreased 3.7 percentage points (adjusted DID = −4.0%; 95% CI = −7.1%, −0.9%). Disparity reductions of similar magnitude were present among Black and Hispanic subgroups. Results did not differ by years in the United States (Appendix Figure B).

TABLE 1—

Changes in Immigrant Versus US-Born Inequity in Self-Reported Postpartum Uninsurance Among Persons Insured by Medicaid at Birth Before and After Families First Coronavirus Response Act (FFCRA) Implementation: New York City, 2016–2021

	US-Born				Immigrant				Difference-in-Differences
	Pre-FFCRA, % (95% CI)	Post-FFCRA, % (95% CI)	Risk Difference, % (95% CI)	Adjusted Risk Difference,a % (95% CI)	Pre-FFCRA, % (95% CI)	Post-FFCRA, % (95% CI)	Risk Difference % (95% CI)	Adjusted Risk Difference,a % (95% CI)	Difference-in-Difference, % (95% CI)	Adjusted Difference-in-Difference,a % (95% CI)
All race/ethnicities	1.2 (0.6, 2.3)	0.7 (0.2, 2.6)	−0.5 (−1.7, 0.007)	−0.5 (−5.6, 4.6)	13.6 (11.8, 15.6)	9.3 (7.1, 12.2)	−4.2 (−7.4, −1.0)	−4.9 (−7.8, −2.0)	−3.7 (−7.1, −0.3)	4.0 (7.1, 0.9)
Hispanic	1.9 (0.8, 4.4)	2.2 (0.6, 8.3)	0.3 (−3.1, 3.7)	−0.1 (−6.0, 5.8)	23.5 (20.0, 27.3)	17.7 (13.4, 22.9)	−5.8 (−11.8, 0.2)	−4.0 (−9.9, 1.9)	−6.1 (−13.0, 0.8)	−4.0 (−11.0, 3.0)
Non-Hispanic Whiteb	0.4 (0.1, 2.8)	0.0	…	…	10.2 (6.3, 16.0)	0.4 (0.1, 2.6)	−9.8 (−14.6, −5.0)	−10.1 (−18.4, −1.8)	…	…
Non-Hispanic Black	1.5 (0.5, 4.2)	0.1 (0.0, 0.8)	−1.4 (−3.0, 0.2)	−1.3 (−21.6, 19.0)	5.7 (3.4, 9.5)	0.2 (0.0, 1.1)	−5.5 (−8.5, −2.5)	−5.4 (−8.5, −2.3)	−4.2 (7.5, −0.8)	−4.1 (−10.0, 1.8)
Non-Hispanic Asian or Pacific Islanderc	…	…	…	…	5.8 (3.7, 9.0)	1.2 (0.3, 4.5)	−4.6 (−7.6, −1.6)	−4.7 (−9.0, −0.4)	…	…

Note. CI = confidence interval.

^a Adjusted for education, age, parity, race/ethnicity, and nativity.

^b Non-Hispanic White subgroup excluded from risk difference and difference-in-difference analyses because zero events of uninsurance were observed in the post-FFCRA implementation US-born subgroup, resulting in uninterpretable results.

^c Non-Hispanic Asian or Pacific Islander US-born subgroup suppressed because of small sample size (< 30).

The placebo analysis produced null results, as desired (Appendix Table B), and in sensitivity analyses, alternative washout periods produced similar results (Appendix Table C).

DISCUSSION

We found that continuous Medicaid enrollment during FFCRA was associated with a 4 percentage point decrease in postpartum uninsurance in NYC among immigrants, resulting in a reduction, but not elimination, of immigrant versus US-born inequities. However, 18% of Hispanic immigrants still reported they were uninsured.

Our equity impact analysis contributes to evidence that continuous Medicaid coverage during FFCRA decreased uninsurance. Medicaid-to-uninsured postpartum churn decreased nationally, from 10.3% in 2019 to 3.7% in 2021,⁷ and Medicaid beneficiaries used more preventive services in the postpartum period.⁸ We build on these findings by assessing the impact on equity on immigrants, the subpopulation most likely to lose coverage postpartum.

A strength of our study is the availability of birth country, years in the United States, and hospital-reported delivery payer. One limitation is that postpartum insurance was self-reported; however, this does yield the patient’s perspective of coverage, which is vital to motivating patients to seek follow-up and preventive care. Another limitation is that we did not know the legal status of immigrants in our data, which would have allowed us to better understand the fraction of immigrants expected to gain coverage (Appendix Table D).

Public Health Implications

Many Hispanic immigrants reported no health insurance despite being eligible for continuous Medicaid coverage during FFCRA, suggesting a lack of awareness of their coverage. Other research has suggested that people enrolled in Medicaid during the pandemic thought they were uninsured.⁹ One explanation for our findings is the fear and misunderstanding about Medicaid eligibility related to immigration status and public charge criteria among immigrants, which was thought to influence the uptake of Medicaid and Supplemental Nutrition Assistance Program (SNAP) at the onset of the pandemic.¹⁰ Other explanations include poorer dissemination of information related to the continuous Medicaid provision in immigrant-serving health institutions or immigrant communities. The consequences of this Medicaid coverage knowledge gap could be profound: in 2020, postpartum maternal mortality increased 41% during the COVID-19 pandemic, with the largest increase for Hispanic people.¹¹ To combat this knowledge gap, agencies need targeted strategies to disseminate information on changes in postpartum Medicaid policy—for example, by working with community-based organizations and federally qualified health centers. However, these strategies will prove a challenge given the continued level of anti-immigrant sentiment in the United States and fear in some immigrant communities.¹²

Our findings have national implications. To date, of 47 states that implemented postpartum Medicaid extensions, New York and 11 other states include enrollees regardless of immigration status. As such, the gains in equity we identified in our natural experiment estimate what to expect in these states. However, states that exclude categories of immigrants may not see similar gains, thus reducing the effectiveness of the policy while leaving behind many immigrants.

Conclusion

Postpartum Medicaid extension policies have the potential to reduce inequities, but a targeted, community-integrated approach is needed to raise awareness of coverage and ensure optimal implementation to advance maternal health equity.