ABSTRACT
Objective
To estimate the effect of Medicaid continuous coverage provision (CCP) during the COVID‐19 public health emergency on postpartum insurance for low‐income noncitizen immigrants compared to citizens, and to assess whether state use of the from‐conception‐to‐the‐end‐of‐pregnancy (FCEP) policy, which excluded many immigrants from postpartum coverage, modified this effect.
Study Setting and Design
This study used a quasi‐experimental design (difference‐in‐differences, or DID) to compare the insurance coverage gains between low‐income postpartum noncitizens and citizens, using 2019 as the reference year. We further applied a triple difference (DDD) approach to assess the effect of the FCEP‐only coverage policy, hypothesizing that noncitizens in states using FCEP as the sole pathway would experience smaller gains than in other states using various pathways to cover pregnant noncitizens otherwise ineligible.
Data Sources and Analytic Sample
Using the 2016–2022 American Community Survey individual‐level data, we identified 61,572 low‐income citizens (weighted N = 7,553,823) and 10,444 low‐income noncitizens (weighted N = 1,487,373), with 8 FCEP‐only states and 23 other states with various pathways.
Principal Findings
During the study period, the average postpartum uninsurance rate was 44.6% for noncitizens and 14.0% for citizens, while Medicaid coverage was 34.5% for noncitizens and 57.5% for citizens. Medicaid gains among noncitizens were statistically similar to gains among citizens in 2021 compared to 2019 in both FCEP‐only states (DID = 2.3 percentage points [pp], 95% CI: −13.3, 8.7) and states with various pathways (DID = 4.6 pp, 95% CI: −1.2, 10.4), but these trends did not persist in 2022. We found no differences in insurance coverage gains among noncitizens between FCEP‐only and states with various pathways.
Conclusions
Findings suggest that the CCP did not substantially reduce postpartum coverage inequities for noncitizen immigrants in either FCEP‐only states or states with various pathways. Postpartum coverage inequities persisted under the CCPs, leaving many noncitizens uninsured.
CALLOUT BOX
- What Is Known on This Topic
-
○Medicaid coverage extensions during the COVID‐19 emergency reduced uninsurance after birth, but little is known about their effect on noncitizens.
-
○Some restrictive states limited coverage for immigrants to the pregnancy period only, excluding them from postpartum benefits; many noncitizens in these states were left out of the coverage extensions.
-
○
- What This Study Found
-
○Postpartum insurance disparities between citizens and noncitizens persisted.
-
○Coverage extensions did not significantly narrow this disparity in either restrictive or generous states, leaving many noncitizens uninsured.
-
○
1. Introduction
Many low‐income immigrants face significant barriers to accessing insurance coverage for perinatal care in the United States (US). Under federal law, immigrants who have not been lawfully present in the US for at least 5 years are typically ineligible for full‐scope Medicaid using federal funds [1]. Evidence has shown that legal exclusions in Medicaid benefits are negatively associated with perinatal care. For example, immigrants in states that do not offer Medicaid coverage for recent lawfully‐present immigrants are less likely to receive postpartum care than those in states with full coverage [2]. Similarly, postpartum immigrants enrolled in Emergency Medicaid are less likely to have postpartum visits compared to those with full‐benefit Medicaid [3], whereas its extension significantly increased the likelihood of postpartum visits among immigrants [4]. Given this landscape of unequal access to postpartum care, it is important to understand how recent efforts to extend Medicaid in the postpartum period have impacted immigrants.
In 2020, the Families First Coronavirus Response Act (FFCRA) established a de facto Medicaid postpartum coverage extension for pregnant and birthing people nationwide. Under this CCP, anyone enrolled in a qualifying Medicaid program on or after March 18, 2020, automatically had continuous coverage until the public health emergency ended in March 2023 [5]; prior to the CCP, postpartum women typically lost Medicaid benefits 60–90 days after birth. A growing body of evidence suggests that continuous coverage benefits postpartum women. For instance, Weber et al. found that postpartum women who were at the highest risk of losing Medicaid coverage before the CCP experienced the greatest Medicaid gains after FFCRA [6], and similarly, Daw et al. found an increase in overall postpartum Medicaid enrollment and a decrease in the uninsurance rate during the same period [7]. Gordon et al. also documented that the CCP led to fewer people losing Medicaid coverage and an increase in continuous 12‐month Medicaid enrollment postpartum [8]. However, not every Medicaid program was subject to the CCP, and thus, not every postpartum Medicaid beneficiary likely benefited from the policy change equally. Specifically, many noncitizens ineligible for full‐benefit Pregnancy Medicaid programs were likely excluded, yet little attention has been paid to this nuance. While one study found that postpartum uninsurance decreased among immigrants in New York City during the CCP [9], this result cannot be generalized; New York state uniquely covers nonqualified immigrant residents under pregnancy Medicaid, making pregnant immigrants in the state eligible for the CCP, unlike in many other states. Therefore, the extent to which postpartum immigrants benefited from the coverage extension nationwide remains unclear. This study aims to fill this gap by comparing coverage gains after the implementation of the CCP between low‐income postpartum citizens and noncitizens at the national level. We hypothesized that noncitizens would experience smaller coverage gains from the CCP than citizens.
Further, the from‐conception‐to‐the‐end‐of‐pregnancy (FCEP) (previously called the “unborn child option”) is a policy approach that, even prior to the pandemic, induced variation in how states covered pregnant citizens versus noncitizens. After the 1996 Personal Responsibility and Work Opportunity Reconciliation Act imposed a five‐year waiting period (also known as the 5‐year bar) that restricted Medicaid eligibility for pregnant noncitizen immigrants, some states adopted the FCEP under the Children's Health Insurance Program (CHIP) to provide prenatal Medicaid coverage regardless of immigration status. However, FCEP provides limited coverage because it focuses on supporting future US citizens rather than the noncitizen mothers themselves [10, 11, 12, 13]. Consequently, FCEP covers only services directly related to conception through birth, with postpartum care either excluded or limited [12, 13] some provide only one or two postpartum visits, while others extend coverage up to 60 days with varying benefit scopes [13]. In contrast, other states use pregnancy Medicaid or state‐funded coverage that provides more comprehensive coverage for immigrants than FCEP options [13].
States' approaches to coverage of noncitizens prior to the CCP may have had consequences as to the differential impact of CCP on postpartum uninsurance based on citizenship status. The CCP under FFCRA did not apply to individuals only covered by FCEP, because FCEP was not subject to the maintenance of eligibility requirements [14, 15]. By contrast, many states that used pathways such as the Children's Health Insurance Program Reauthorization Act of 2009 (CHIPRA) 214—an option to waive the 5‐year bar for lawfully present pregnant immigrants—or state‐only funds in addition to FCEP or instead of FCEP to cover pregnant noncitizens may have been more likely to provide continuous enrollment for noncitizens. As a result, we hypothesize that low‐income postpartum noncitizens living in states solely using FCEP experienced fewer benefits compared to those living in states offering other Medicaid coverage pathways.
To test these hypotheses, we used data from the 2016–2022 American Community Survey (ACS). Using a quasi‐experimental design, we first assessed how the CCP differentially affected changes in insurance coverage among low‐income postpartum citizens and noncitizens nationwide. We then compared these effects in FCEP‐only states versus a comparison group of states.
2. Methods
2.1. Data Source and Study Population
We used 2016–2022 ACS individual‐level microdata obtained from IPUMS USA. Postpartum women were identified as females aged 15–50 who answered “yes” to the ACS question on recent fertility: “Has this person given birth to any children in the past 12 months?”. We excluded individuals under 19 years old, as they were continuously covered by CHIP and therefore not impacted by the policy. We restricted the sample to postpartum women from low‐income households, who were most likely to be affected by the CCP. We operationalized low‐income households as those with household incomes below 200% of the federal poverty level (FPL). To identify citizenship status, we used the ACS question, “Is this person a citizen of the United States?”.
We excluded respondents from Idaho, Nebraska, Oklahoma, Rhode Island, and Utah to eliminate potential confounding effects because they expanded Medicaid income eligibility during the study period, as well as Washington, DC, which provides coverage for certain low‐income adults regardless of citizenship status. Lastly, we obtained state‐level unemployment and poverty rates from the Current Population Survey Historical Poverty Tables [16] and the Bureau of Labor Statistics [17].
2.2. Policy Classification of States
Based on the previous policy analysis [13], we categorized states according to their perinatal care mechanisms for noncitizens not meeting immigration status requirements (i.e., under the 5‐year bar) (see Table A1) prior to FFCRA, specifically whether they used FCEP as their sole coverage option. Our hypothesis was that states using only FCEP may have de‐prioritized mothers' health, relative to other states. We identified eight states (Table A1) in the “FCEP‐only” group, which consists of states that exclusively used the FCEP to cover perinatal noncitizen immigrants under the 5‐year bar. As noted above, FCEP was exempt from the FFCRA CCP requirements; therefore, noncitizens in these states did not receive continuous postpartum coverage. In contrast, “States with Various Pathways” are 23 states that covered noncitizen pregnant immigrants under the 5‐year bar using other or multiple pathways, such as CHIPRA 214 or state‐only funded coverage, even if these states also used FCEP. For example, California used FCEP but was not assigned to the FCEP‐only group because noncitizens under the five‐year bar were also covered by CHIPRA 214 and state‐only‐funded programs. Thus, grouping states like California with FCEP‐only states could confound the effect of FCEP. Lastly, we excluded the 14 remaining states (“States with no coverage”) that do not provide any perinatal coverage for noncitizen immigrants under the 5‐year bar, as our analyses aimed to compare the FCEP with more comprehensive coverage.
2.3. Outcome Measures
We used the ACS question, “Is this person currently covered by any of the following types of health insurance or health coverage plans?”. Based on this, we created three binary variables for self‐reported health insurance status at the time of the survey: uninsurance (vs. any insurance), Medicaid (vs. uninsured or non‐Medicaid), and private insurance (vs. uninsured or public insurance). As many respondents reported multiple health coverages, we adjusted them into mutually exclusive insurance types based on a hierarchy used by the Kaiser Family Foundation: Medicaid, Medicare, Employer, Military, and Non‐Group [18].
2.4. Covariates
Individual‐level covariates included age, marital status, educational attainment, employment status, and race and ethnicity of the respondents. Given a small number of observations, in adjusted models, we combined Asian and Pacific Islander, American Indian or Alaska Native, and Other into a single group. We included state‐level time‐varying covariates plausibly associated with health insurance and Medicaid coverage of postpartum people: state unemployment rate (annual averages) and state poverty rate (% of the population below FPL).
2.5. Statistical Analysis
We first graphically examined yearly estimates of low‐income postpartum women's uninsurance, Medicaid, and private health insurance coverage by citizenship status for all states combined during the pre (2016–2019) and post (2021–2022) periods. Then, using a difference‐in‐differences (DID) event study approach, we estimated how many low‐income postpartum noncitizens gained insurance coverage after implementing the CCP compared to their citizen counterparts. Our analytic approach assesses whether a uniform federal policy functioned differently for noncitizens, a population structurally excluded from many full‐benefit coverage pathways. The DID framework is useful for measuring within‐group pre–post changes, with the interaction term capturing differential changes by citizenship status, providing a single causal contrast with effect modification by group. This comparison quantifies disparities in how the policy operated across populations.
We used an event study approach for two key purposes. First, it allowed us to examine pre‐policy trends in insurance coverage among citizens and noncitizens. Given the potential bias from differences in state‐level policy environments before the policy change, establishing this comparability is essential. To test the parallel trends assumption, we examined interactions between noncitizen status and each pre‐policy year, using 2019 as the baseline. Visual inspection and statistical analysis of these pre‐treatment trends allowed us to evaluate whether the groups followed similar trajectories prior to the policy implementation. Second, the event study approach estimates policy impacts at multiple time points relative to a reference period. In our case, it allows us to examine differences in outcomes for 2021 and 2022 separately, compared to 2019. The DID coefficients in the event study models represented effects of CCPs on citizen versus noncitizen disparity in 2021 (Post 1) and 2022 (Post 2) relative to 2019. We disaggregated the results into marginal effects for citizens and noncitizens to estimate the percentage point changes separately for each group.
Next, we split the sample into two groups: low‐income postpartum women in FCEP‐only states (treated group) and their counterparts in states with various pathways (comparison group). Using the same event study approach, we estimated the effect of the continuous coverage extension as the intervention and citizenship status as a stratification variable. The coefficients of interest represent the effects of the CCP on the disparity between citizens and noncitizens in 2021 and 2022, relative to 2019. We compared the event‐study coefficients and disaggregated marginal effects for citizens and noncitizens across FCEP‐only and multiple‐pathways states.
Finally, using a triple DID approach, we formally tested the heterogeneous effects of CCP. Here, the triple interaction term between “noncitizen status,” “post,” and “FCEP‐only states” is the variable of interest. This term captures whether the impact of the CCP on the insurance disparity between citizens and noncitizens varied by a state's reliance on FCEP as its sole coverage pathway for noncitizens under the 5‐year bar.
As supplemental analyses, we re‐ran all regression models on only Hispanic low‐income postpartum women because Central America, South America, and the Caribbean are the largest regions from which noncitizens who are undocumented or Deferred Action for Childhood Arrivals (DACA) recipients are [19, 20, 21, 22]. We clustered standard errors at the state level and applied the ACS person‐level sampling weight to all analyses.
3. Results
3.1. Overall Comparison of Citizens Versus Noncitizens
Throughout the study period (2016–2019 and 2021–2022), we identified 72,016 low‐income postpartum women (weighted N = 9,041,196). Of these, 61,572 were citizens (weighted N = 7,553,823), and 10,444 were noncitizens (weighted N = 1,487,373) (Table 1).
TABLE 1.
Sample characteristics: pre‐ and post‐continuous coverage provision comparisons among low‐income postpartum women by citizenship status.
| Citizen | Noncitizen | |||||
|---|---|---|---|---|---|---|
| Pre (2016–2019) | Post 1 (2021) | Post 2 (2022) | Pre (2016–2019) | Post 1 (2021) | Post 2 (2022) | |
|
N (weighted N) |
42,473 (5,145,337) |
9388 (1,200,892) |
9711 (1,207,594) |
7337 (1,047,698) |
1550 (217,876) | 1557 (221,799) |
| Age (mean) | 28.3 | 29.0 | 29.5 | 31.2 | 31.9 | 31.9 |
| Currently not married | 61.6% | 61.1% | 58.6% | 34.8% | 36.1% | 38.4% |
| High school or less | 60.2% | 59.7% | 59.2% | 73.6% | 70.7% | 71.0% |
| Employment | ||||||
| Employed | 46.7% | 44.7% | 46.9% | 30.0% | 26.8% | 28.6% |
| Unemployed | 8.4% | 8.9% | 7.4% | 4.2% | 5.5% | 3.1% |
| Not in labor force | 44.9% | 46.4% | 45.6% | 65.8% | 67.8% | 68.3% |
| Race/ethnicity | ||||||
| Non‐Hispanic White | 46.8% | 44.7% | 44.0% | 8.4% | 7.2% | 5.8% |
| Non‐Hispanic Black | 23.4% | 22.3% | 22.8% | 9.6% | 9.8% | 10.3% |
| Non‐Hispanic Asian and Pacific Islander | 2.5% | 2.6% | 2.8% | 11.9% | 10.8% | 11.5% |
| Non‐Hispanic American Indian or Alaska Native | 1.3% | 0.9% | 0.7% | 0.0% | 0.0% | 0.0% |
| Non‐Hispanic Other a | 3.0% | 5.4% | 5.4% | 1.7% | 2.3% | 3.6% |
| Hispanic | 23.0% | 24.0% | 24.2% | 68.4% | 69.8% | 68.8% |
Note: Authors' analysis of American Community Survey (ACS) IPUMS data 2016–2019, 2021–2022. ACS person weight was applied.
“Other” category includes “some other race” and “two or more races.”
Table 2 shows persistent and substantial disparities in self‐reported uninsurance between citizens and noncitizens throughout the study period. Uninsurance rates for postpartum noncitizens were more than 30 percentage points (pp) higher than postpartum citizens, while the proportion insured by Medicaid was more than 20 pp lower among noncitizens compared to citizens. Unadjusted pre–post comparisons indicate that uninsurance rates decreased for citizens in both 2021 and 2022; however, similar declines were not observed among noncitizens, especially in 2022. Similarly, noncitizens experienced a 4.2 pp (95% CI: 0.7,7.5) increase in Medicaid coverage in 2021, but this gain did not persist in 2022. Both groups saw decreases in private insurance in 2021, though these changes were also not sustained in 2022 (see Figure A1 for detailed trends). Adjusted DID estimates comparing 2016–2019 to 2021–2022 (Table A2) suggest that the differences between the two groups were not statistically significant.
TABLE 2.
Unadjusted pre‐ versus postcontinuous coverage provision comparisons of insurance rates among low‐income postpartum women by citizenship status.
| Citizen | Noncitizen | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Pre (2016–2019) | Post 1 (2021) | Post 2 (2022) | DiffPost1‐Pre [95% CI] | DiffPost2‐Pre [95% CI] | Pre (2016–2019) | Post 1 (2021) | Post 2 (2022) | DiffPost1‐Pre [95% CI] | DiffPost2‐Pre [95% CI] | |
| Uninsured | 14.0% | 11.1% | 10.7% |
−2.9 pp [−3.8, −2.0] |
−3.3 pp [−4.2, −2.3] |
44.6% | 42.1% | 44.2% |
−2.4 pp [−6.0, 1.1] |
−0.3 pp [−3.8, 3.1] |
| Medicaid | 57.5% | 62.7% | 61.8% |
5.2 pp [3.7, 6.5] |
4.3 pp [2.9, 5.7] |
34.5% | 38.7% | 34.4% |
4.2 pp [0.7, 7.5] |
−0.2 pp [−3.4, 3.1] |
| Private Insurance | 27.6% | 25.4% | 26.7% |
−2.2 pp [−3.4, −0.9] |
−0.9 pp [−2.2, 0.3] |
20.3% | 18.8% | 21.1% |
−1.5 pp [−4.2, 1.1] |
0.8 pp [−2.0, 3.6] |
Note: Authors' analysis of American Community Survey (ACS) IPUMS data 2016–2019, 2021–2022. ACS person weight was applied.
Abbreviations: CI, confidence interval; pp, percentage points.
Next, we estimated DID event study models adjusting for differences in baseline covariates between citizens and noncitizens. While the CCP had a reduced uninsurance rate and increased Medicaid within both groups in 2021, there were no significant between‐group differences. Figure 1 presents disaggregated marginal changes for citizens and noncitizens for each year, using 2019 as a reference year. The DID coefficients for each year indicate whether the differences in marginal changes between the two groups are statistically significant. In 2021, uninsurance among noncitizens decreased by 7.3 pp (95% CI: −13.0 to −1.5), compared to a 5.2 pp (95% CI: −7.6 to −2.7) decrease among citizens. By 2022, only citizens maintained a 4.3 pp (95% CI: −6.6 to −1.9) decrease in uninsurance relative to 2019. However, the between‐group differences were not significant in either year. Medicaid coverage increased by 8.2 pp (95% CI: 2.2, 14.2) among noncitizens and 6.2 pp (95% CI: 2.8, 9.6) among citizens in 2021. In 2022, only citizens retained a 5.6 pp (95% CI: 2.6, 8.6) increase in Medicaid coverage compared to 2019, but again, the between‐group difference was insignificant. There were no notable changes in private insurance rates among either group following the implementation of the CCP (see Figure A2 and Table A3 for full regression results).
FIGURE 1.
Average marginal effects of continuous coverage provision among low‐income postpartum citizens versus noncitizens: difference‐in‐differences event study estimates. The blue circle denotes the citizen group and the red square denotes the noncitizen group. 95% Confidence Intervals. The X‐axis represents year, and the Y‐axis represents marginal changes in the probability of uninsurance, Medicaid, and private insurance coverage. Each point estimate represents the predicted difference in uninsurance, Medicaid, and private insurance coverage relative to 2019 (the reference year) for noncitizens and citizens. For example, in 2021, noncitizens experienced a 7.3 pp decrease in the uninsurance rate compared to 2019. The “Diff” label for each year reflects the difference in marginal effects between noncitizens and citizens (noncitizens minus citizens), corresponding to the difference‐in‐differences coefficients. All models applied American Community Survey person‐level weights, included state fixed effects, and clustered standard errors at the state level. Covariates included age, marital status, education level, race/ethnicity, employment status, state‐level poverty rate, and state‐level unemployment rate. Idaho, Utah, Nebraska, Oklahoma, the District of Columbia, and Rhode Island were excluded due to Medicaid expansion and changes in eligibility criteria during the study period. Full results are presented in Table A2.
3.2. Comparison of Citizens Versus Noncitizens by State Policy Group
Baseline characteristics by state group and citizenship status are provided in Tables A4 and A5. Standard pre–post DID results (Tables A6 and A7) indicate that in FCEP‐only states, noncitizens had smaller reductions in uninsurance and smaller increases in Medicaid coverage compared to citizens, though these differences were not significant.
Figure 2 presents event study results showing disaggregated marginal changes in outcomes for citizens and noncitizens stratified by state policy group, using 2019 as the reference year. In 2021, states with various pathways (Panel B) experienced significant declines in uninsurance for everyone (−5.7 pp for citizens, 95% CI: −9.5, −1.8; −10.1 pp for noncitizens, 95%, 95% CI: −17.0, −3.1), while FCEP‐only states (Panel A) showed insignificant increases. In 2022, uninsurance declined in FCEP‐only states for both citizens (−5.3 pp) and noncitizens (−8.8 pp). In contrast, in states with various pathways, the uninsurance decline continued for citizens (−3.6 pp, 95% CI: −5.6, −1.4) but not for noncitizens. Medicaid coverage increased substantially in states with various pathways in 2021 for both citizens (+6.1 pp, 95% CI: 1.4, 10.9) and noncitizens (+10.8 pp, 95% CI: 3.6, 18.0). In 2022, only citizens in these states maintained gains (+4.4 pp, 95% CI: 12.3, 7.6), while noncitizens' gains diminished. However, the differences in marginal changes between citizens and noncitizens were not statistically significant (see Tables A8 and A9 for full regression results).
FIGURE 2.
Average marginal effects of continuous coverage provision on low‐income postpartum citizens versus noncitizens: difference‐in‐differences event study estimates by from‐conception‐to‐the‐end‐of‐pregnancy (FCEP)‐only states and states with various pathways. PANEL A: FCEP‐only states. PANEL B: states with various pathways. The blue circle denotes the citizen group and the red square denotes the noncitizen group. 95% Confidence Intervals. The X‐axis represents year, and the Y‐axis represents marginal changes in the probability of uninsurance, Medicaid, and private insurance coverage. Each point estimate represents the predicted difference in uninsurance, Medicaid, and private insurance coverage relative to 2019 (the reference year) for noncitizens and citizens. The “Diff” label for each year reflects the difference in marginal effects between noncitizens and citizens (noncitizens minus citizens), corresponding to the difference‐in‐differences coefficients. All models applied American Community Survey person‐level weights, included state fixed effects, and clustered standard errors at the state level. Covariates included age, marital status, education level, race/ethnicity, employment status, state‐level poverty rate, and state‐level unemployment rate. Idaho, Utah, Nebraska, Oklahoma, the District of Columbia, and Rhode Island were excluded due to Medicaid expansion and changes in eligibility criteria during the study period. Full results are presented in Tables A5 and A6.
Figure A4 illustrates the triple DID coefficients, which indicate differential changes in the citizen versus noncitizen insurance disparity pre‐ versus post‐period for FCEP‐only states versus states with various pathways (see Table A12 for full regression results). We did not find significant triple DID coefficients across the three outcomes, suggesting that coverage gains among noncitizens were statistically similar between the two state groups.
As supplemental analyses, we focused on the Hispanic sample and found similar but pronounced patterns (Tables A10 and A11). Hispanic noncitizens experienced smaller Medicaid gains than Hispanic citizens in FCEP‐only states in both 2021 and 2022. In states with various pathways, Hispanic noncitizens experienced larger Medicaid gains than citizens in 2021 but smaller gains in 2022. Lastly, triple DID results in the Hispanic sample remained largely consistent with the full sample (Table A13).
4. Discussion
Our study examined self‐reported insurance coverage among low‐income postpartum citizens and noncitizens before and during the continuous provision of Medicaid during the COVID‐19 pandemic. To our knowledge, it is the first multistate study that measures the impact of the CCP on postpartum immigrants. Using the national‐level data, this study aims to estimate population‐level effects and provide insights for the broader population impact of the CCP. Our hypothesis that noncitizens gained less coverage than citizens during the continuous coverage period was not supported. This suggests that postpartum noncitizens may have remained enrolled through different pathways, with some meeting Medicaid eligibility requirements and others accessing other programs. It likely reflects the heterogeneity within the noncitizen population: some noncitizens qualified for full‐scope Medicaid coverage (e.g., lawful permanent residents meeting the 5‐year residency requirement or those in exempt categories), while others may have accessed state‐funded programs or emergency Medicaid, or even remained covered while states sorted through the administrative complexity of understanding who was eligible for the CCP.
Although our results signal that a sizable number of low‐income postpartum noncitizens remained covered during the CCPs, substantial disparities in self‐reported insurance persisted. While previous studies found that CCP during the COVID‐19 pandemic improved Medicaid coverage among postpartum women, our findings demonstrate that disparities between citizens and noncitizens persist. Over 40% of postpartum noncitizens were uninsured from 2016 to 2022, compared to 12% of citizens, and about 35% of noncitizens had Medicaid, versus 60% of citizens. These persistent disparities reflect the legal exclusions in eligibility that prevent many noncitizens from accessing Medicaid for perinatal care.
We also hypothesized that postpartum noncitizens would be largely excluded from coverage gains in states using FCEP exclusively, and indeed, we found no Medicaid coverage gains among this group. However, contrary to the hypothesis, we found no significant difference in any gains between citizens and noncitizens, likely because estimates in these states have limited statistical power due to small sample sizes. Although the uninsurance rate dropped in 2022 in FCEP‐only states, likely because Medicaid and private insurance both increased, this finding applied to both citizens and noncitizens. Meanwhile, in states with various pathways, we found that the CCP had a significant impact on noncitizens in 2021, reducing uninsurance and increasing Medicaid coverage. Despite a few studies focusing on immigrant postpartum women, these findings align with those of Janevic et al., who reported a notable decline in uninsurance among immigrants in New York City in 2021, part of a state with comprehensive postpartum Medicaid [9]. However, such coverage gains among noncitizens in this group were not significantly different from their citizen counterparts. The overall findings, including the triple DID results, suggest that, while noncitizens in FCEP‐only states did not benefit from the CCP, they neither fell further behind their citizen counterparts in their own states nor in states with various pathways.
Another important finding is the complexity of measuring state differences in post‐FFCRA postpartum Medicaid policies vis‐à‐vis noncitizen immigrants, especially given that some states were concurrently using multiple pathways to cover pregnant and birthing people with different immigration statuses. While we initially hypothesized that using FCEP as the sole mechanism would serve as a good proxy for measuring the post‐FFCRA policy environment, our null findings could indicate that it captured only part of the complexity of what occurred. Across all states, the rapidly evolving pandemic policy environment in which this analysis was conducted and a corresponding lack of documentation made it further difficult to ascertain which groups actually received postpartum coverage extensions and for how long. Future work documenting details of states' actual postpartum extension policies on noncitizen immigrants during the public health emergency could be informative.
Our study has both limitations and strengths. A major strength is the availability of the citizenship data in the ACS, which is not available in most perinatal data sources. However, it would have been helpful if we could identify the noncitizens' exact legal status, since different states may treat undocumented immigrants differently from legal permanent residents. Because some postpartum noncitizens qualified for Medicaid (e.g., those meeting the 5‐year residency requirement) while others did not, our analysis could not separate out these eligible noncitizens from the others. Unfortunately, the ACS does not report the legal status of noncitizens [23]. In addition, national surveys undercount some noncitizen populations because of issues such as a lack of standard mail delivery, internet access, language, literacy, and fear that information might be used to find or deport undocumented immigrants [24, 25, 26]. A second limitation relates directly to this issue: the small sample size of postpartum noncitizens reduced the statistical power of our analysis. This underscores the need for larger datasets on immigrant populations to better assess the policy impacts on them. Additionally, our data cannot distinguish whether the persistent inequity for noncitizens stems from legal ineligibility versus barriers to enrollment among those who are eligible. While our analysis focuses on legal exclusions embedded in eligibility rules, factors such as language barriers, limited outreach, and fear of information disclosure may have prevented eligible noncitizens from enrolling or maintaining coverage. Third, our analyses relied on self‐reported insurance, which cannot be verified. Studies found that ACS self‐reported data undercounted Medicaid coverage compared to administrative data from the Centers for Medicare and Medicaid Services [27, 28]. Postpartum women in our sample might not have realized that they still had Medicaid coverage at the time of survey responses, which could have led to underestimating the impact of the CCP. However, self‐reported coverage is useful because the perceived insurance coverage can greatly affect their use of postpartum follow‐up care. In addition, underreporting might lead to measurement error in our key outcomes. However, measurement error in the dependent variable is typically less problematic than in the independent variable because it produces consistent coefficient estimates with increased standard errors, or it may bias estimates toward zero if the measurement error is systematic [29]. This suggests our estimates may understate or provide very conservative estimates of the true impact of continuous coverage on uninsurance and Medicaid coverage. Fourth, we excluded six states that expanded Medicaid during our study period. This strengthened our analysis by eliminating potential confounding effects of the expansion and improving internal validity. However, the results cannot be generalized to the excluded states, which represents a trade‐off of this analytic decision. Lastly, another complexity is that pandemic‐related employment swings may have impacted individuals' Medicaid eligibility, apart from the CCP. This suggests that a complex mix of policy and individual factors influences Medicaid gains. However, in our analysis, we controlled for individual‐level employment status and state‐level unemployment rates.
Our study offers several policy implications. First, we found a national trend suggesting that the CCP increased Medicaid coverage and reduced uninsurance rates among noncitizens. However, moderate effect sizes and weak statistical significance of our findings indicate that the CCP did not substantially reduce postpartum coverage inequities between noncitizens and citizens in either FCEP‐only or states with various pathways, leaving many noncitizens uninsured or underinsured. Therefore, states should establish inclusive pathways to provide comprehensive Medicaid or CHIP coverage during pregnancy, regardless of duration in the US or legal status, while also leveraging state‐based Marketplace plans under the Affordable Care Act to offer affordable options. Furthermore, as prior studies documented, a lack of federal guidance and communication on the postpartum continuous coverage requirement among state agencies and providers posed great challenges to the more effective implementation, especially for immigrant populations [30, 31]. Persistent disparities between postpartum citizens and noncitizens highlight the need to address both legal exclusions in eligibility criteria and enrollment barriers that prevent eligible immigrants from accessing coverage [31, 32]. This could involve providing clearer, more accessible information, expanding language services, and partnering with community‐based organizations. Ensuring continuous coverage during this vulnerable period is crucial for advancing health equity and maternal health in the US, and future public health emergencies should guarantee uninterrupted Medicaid coverage for all postpartum individuals. Lastly, the American Rescue Plan Act of 2021 allowed states to voluntarily extend postpartum Medicaid coverage to 12 months after the Public Health Emergency period ended. Future analyses on unwinding and the effect of the American Rescue Plan Act will be valuable for understanding immigrant exclusions in postpartum Medicaid coverage extensions.
Funding
This work was supported by National Institutes of Health, R01MD018180.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Table A1: State classification based on pre‐FFCRA policies.
Figure A1: Trends in uninsurance, Medicaid, and private insurance rates among low‐income citizen and noncitizen postpartum women, 2016–2019.
Table A2: Difference‐in‐difference results of the impact of continuous coverage provision on insurance rates: low‐income noncitizens (vs. citizens), pre (2016–2019) versus post (2021–2022).
Figure A2: Difference‐in‐difference event study coefficient plots of the impact of continuous coverage provision on insurance rates: low‐income noncitizens (vs. citizens) 2016–2022 (reference year: 2019).
Table A3: Full result table of event study estimation of the impact of continuous coverage provision on insurance rates: low‐income citizens versus noncitizens, 2016–2022 (reference year: 2019).
Table A4: Pre versus post comparisons of sample characteristics by state policy group.
Figure A3: Trends in uninsurance, Medicaid, and private insurance rates among low‐income citizen and noncitizen postpartum women in from‐conception‐to‐the‐end‐of‐pregnancy (FCEP)‐only states versus states with various pathways, 2016–2022.
Table A7: Difference‐in‐difference results of the impact of continuous coverage provision on insurance rates: low‐income noncitizens (vs. citizens), pre (2016–2019) versus post (2021–2022) in states with various pathways.
Table A8: Event study full results: low‐income noncitizens versus citizens in from‐conception‐to‐the‐end‐of‐pregnancy (FCEP)‐only states, 2016–2022.
Table A9: Event study full results: low‐income noncitizens versus citizens in states with various pathways, 2016–2022.
Table A10: Event study full results: Hispanic low‐income noncitizens versus citizens in from‐conception‐to‐the‐end‐of‐pregnancy (FCEP)‐only states, 2016–2022.
Table A11: Event study full results: Hispanic low‐income noncitizens versus citizens in states with various pathways, 2016–2022.
Figure A4: Coefficient plots from triple difference‐in‐differences estimates of the impact of continuous coverage provision: citizen versus noncitizen postpartum women in from‐conception‐to‐the‐end‐of‐pregnancy (FCEP)‐only states versus states with various pathways.
Table A12: Triple difference‐in‐differences event study full results: pre versus post, low‐income citizen versus noncitizen, from‐conception‐to‐the‐end‐of‐pregnancy (FCEP)‐only states versus states with various pathways.
Table A13: Triple difference‐in‐differences event study full results: pre versus post, Hispanic low‐income citizen versus noncitizen, from‐conception‐to‐the‐end‐of‐pregnancy (FCEP)‐only states versus states with various pathways.
Acknowledgments
Research reported in this publication was supported by the National Institute on Minority Health and Health Disparities of the National Institutes of Health under Award Number R01MD018180. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Data Availability Statement
The data that support the findings of this study are available in IPUMS USA at https://www.ipums.org/. These data were derived from the following resources available in the public domain: IPUMS USA, https://www.ipums.org/.
References
- 1. Swartz J. J., Hainmueller J., Lawrence D., and Rodriguez M. I., “Expanding Prenatal Care to Unauthorized Immigrant Women and the Effects on Infant Health,” Obstetrics and Gynecology 130, no. 5 (2017): 938–945. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Steenland M. W., Fabi R. E., Bellerose M., Desir A., White M. S., and Wherry L. R., “State Public Insurance Coverage Policies and Postpartum Care Among Immigrants,” Journal of the American Medical Association 330, no. 3 (2023): 238–246, 10.1001/JAMA.2023.10249. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Rodriguez M. I., McConnell K. J., Skye M., et al., “Disparities in Postpartum Contraceptive Use Among Immigrant Women With Restricted Medicaid Benefits,” AJOG Global Reports 2, no. 1 (2022): 100030, 10.1016/j.xagr.2021.100030. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Rodriguez M. I., Skye M., Lindner S., et al., “Analysis of Contraceptive Use Among Immigrant Women Following Expansion of Medicaid Coverage for Postpartum Care,” JAMA Network Open 4, no. 12 (2021): e2138983–e2138983, 10.1001/jamanetworkopen.2021.38983. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. NM R. L., “D N 17. Text – H.R.6201 – 116th Congress (2019‐2020): Families First Coronavirus Response Act,” March 18, 2020, accessed February 22, 2026, https://www.congress.gov/bill/116th‐congress/house‐bill/6201/text.
- 6. Weber E., Kim H., Ng A., Howell F. M., Fox A., and Janevic T., “Medicaid Eligibility Gaps and Pandemic‐Era Postpartum Insurance Rates,” JAMA Health Forum 6, no. 3 (2025): e250109, 10.1001/jamahealthforum.2025.0109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Daw J. R., MacCallum‐Bridges C. L., Kozhimannil K. B., and Admon L. K., “Continuous Medicaid Eligibility During the COVID‐19 Pandemic and Postpartum Coverage, Health Care, and Outcomes,” JAMA Health Forum 5, no. 3 (2024): e240004, 10.1001/jamahealthforum.2024.0004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Gordon S. H., Chen L., DeLew N., and Sommers B. D., “COVID‐19 Medicaid Continuous Enrollment Provision Yielded Gains in Postpartum Continuity Of Coverage,” Health Affairs 43, no. 3 (2024): 336–343, 10.1377/hlthaff.2023.00580. [DOI] [PubMed] [Google Scholar]
- 9. Janevic T., Birnie L., Belfon K., et al., “Immigrant Inequities in Uninsurance and Postpartum Medicaid Extension: A Quasi‐Experimental Study in New York City, 2016–2021,” American Journal of Public Health 115, no. 5 (2025): 732–735, 10.2105/AJPH.2024.307968. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Fabi R. E., Saloner B., and Taylor H., “State Policymaking and Stated Reasons: Prenatal Care for Undocumented Immigrants in an Era of Abortion Restriction,” Milbank Quarterly 99, no. 3 (2021): 693–720, 10.1111/1468-0009.12519. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Drewry J., Sen B., Wingate M., Bronstein J., Foster E. M., and Kotelchuck M., “The Impact of the State Children's Health Insurance Program's Unborn Child Ruling Expansions on Foreign‐Born Latina Prenatal Care and Birth Outcomes, 2000–2007,” Maternal and Child Health Journal 19, no. 7 (2014): 1464–1471, 10.1007/S10995-014-1650-5. [DOI] [PubMed] [Google Scholar]
- 12. Fox A., Howell F. M., Weber E., and Janevic T., “Left Behind: Medicaid Immigrant Exclusions and Access to Maternal Health Care Across the Reproductive‐Perinatal Continuum,” Medical Care Research and Review 80, no. 6 (2023): 582–595, 10.1177/10775587231170066. [DOI] [PubMed] [Google Scholar]
- 13. Ng A., Janevic T., Fox A. M., Howell F. M., Kim H., and Weber E., “Twelve‐Month Medicaid Postpartum Extensions Ring Hollow for Immigrant Communities,” Health Affairs Forefront (2024), 10.1377/forefront.20240801.693621. [DOI] [Google Scholar]
- 14. Centers for Medicare & Medicaid Services , “COVID‐19 Frequently Asked Questions (FAQs) for State Medicaid and Children's Health Insurance Program (CHIP) Agencies,” 2020, accessed March 3, 2026, https://www.medicaid.gov/sites/default/files/2020‐06/covid‐19‐faqs.pdf.
- 15. Centers for Medicare & Medicaid Services , “COVID‐19 Frequently Asked Questions (FAQs) for State Medicaid and Children's Health Insurance Program (CHIP) Agencies,” 2021, accessed March 3, 2026, https://www.medicaid.gov/state‐resource‐center/downloads/covid‐19‐faqs.pdf.
- 16. Bureau UC , “Historical Poverty Tables: People and Families ‐ 1959 to 2023. Census.gov,” accessed December 29, 2024, https://www.census.gov/data/tables/time‐series/demo/income‐poverty/historical‐poverty‐people.html.
- 17. Local Area Unemployment Statistics Map , “Bureau of Labor Statistics”, accessed December 29, 2024, https://data.bls.gov/lausmap/showMap.jsp;jsessionid=65501768C79EA386C0068D1268087407._t3_08v.
- 18. Health Insurance Coverage of the Total Population , “KFF State Health Facts. KFF,” accessed February 22, 2026, https://www.kff.org/state‐health‐policy‐data/state‐indicator/total‐population/.
- 19. US.migrationpolicy.org , “Profile of the Unauthorized Population,” accessed February 21, 2025, https://www.migrationpolicy.org/data/unauthorized‐immigrant‐population/state/US.
- 20. Venkataramani A. S., Shah S. J., O'Brien R., Kawachi I., and Tsai A. C., “Health Consequences of the US Deferred Action for Childhood Arrivals (DACA) Immigration Programme: A Quasi‐Experimental Study,” Lancet Public Health 2, no. 4 (2017): e175–e181, 10.1016/S2468-2667(17)30047-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. García‐Pérez M. D., “DACA Recipients and Their Health Insurance Dream: Employment, Schooling, and Health Coverage,” Journal of Economics, Race, and Policy 2, no. 1 (2019): 77–108, 10.1007/s41996-019-00029-x. [DOI] [Google Scholar]
- 22. Krogstad J. M. and Passel J. S., “What We Know About Unauthorized Immigrants Living in the U.S,” Pew Research Center (2024). accessed December 11, 2025, https://www.pewresearch.org/short‐reads/2024/07/22/what‐we‐know‐about‐unauthorized‐immigrants‐living‐in‐the‐us/. [Google Scholar]
- 23. Capps R., Bachmeier J. D., and Van Hook J., “Estimating the Characteristics of Unauthorized Immigrants Using U.S. Census Data: Combined Sample Multiple Imputation,” Annals of the American Academy of Political and Social Science 677, no. 1 (2018): 165–179, 10.1177/0002716218767383. [DOI] [Google Scholar]
- 24. Bates N. and Trejo Y. G., “2020 Census Barriers, Attitudes, and Motivators Study (CBAMS) Focus Group Final Report,” U.S. Census Bureau (2019): 1–176. [Google Scholar]
- 25. Kissam E., Mines R., Quezada C., et al., “San Joaquin Valley Latino Immigrants: Implications of Survey Findings for Census 2020”.
- 26. Karpman M., Zuckerman S., and Gonzalez D., “On Eve of 2020 Census, Many People in Hard‐to‐Count Groups Remain Concerned About Participating. Urban Institute,” 2020. accessed December 29, 2024, https://www.urban.org/research/publication/eve‐2020‐census‐many‐people‐hard‐count‐groups‐remain‐concerned‐about‐participating.
- 27. McIntyre A., Smith R. B., and Sommers B. D., “Survey‐Reported Coverage in 2019‐2022 and Implications for Unwinding Medicaid Continuous Eligibility,” JAMA Health Forum 5, no. 4 (2024): e240430, 10.1001/jamahealthforum.2024.0430. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Hest R., “Tracking the Medicaid Undercount in the 2021 ACS Coverage Data. State Health Access Data Assistance Center,” 2023. accessed March 5, 2024, https://www.shadac.org/publications/medicaid‐undercount‐2021‐ACS.
- 29. Bound J., Brown C., and Mathiowetz N., “Chapter 59 ‐ Measurement Error in Survey Data,” in Handbook of Econometrics, vol. 5, ed. Heckman J. J. and Leamer E. (Elsevier, 2001), 3705–3843, 10.1016/S1573-4412(01)05012-7. [DOI] [Google Scholar]
- 30. Johnston E. M., Haley J. M., and Thomas T., “Promoting Continuous Postpartum Medicaid Coverage: Lessons From the Public Health Emergency,” 2021, accessed May 31, 2025, https://policycommons.net/artifacts/1817192/promoting‐continuous‐coverage‐during‐the‐postpartum‐period/2554130/.
- 31. Bernstein H., Gonzalez D., Echave P., and Guelespe D., “Immigrant Families Faced Multiple Barriers to Safety Net Programs in 2021. Urban Institute,” 2022. accessed May 31, 2025, https://www.urban.org/research/publication/immigrant‐families‐faced‐multiple‐barriers‐safety‐net‐programs‐2021.
- 32. Gonzalez D. and Bernstein H., “One in Four Adults in Mixed‐Status Families Did Not Participate in Safety Net Programs in 2022 Because of Green Card Concerns. Urban Institute,” 2023. accessed May 31, 2025, https://www.urban.org/research/publication/one‐four‐adults‐mixed‐status‐families‐did‐not‐participate‐safety‐net‐programs.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table A1: State classification based on pre‐FFCRA policies.
Figure A1: Trends in uninsurance, Medicaid, and private insurance rates among low‐income citizen and noncitizen postpartum women, 2016–2019.
Table A2: Difference‐in‐difference results of the impact of continuous coverage provision on insurance rates: low‐income noncitizens (vs. citizens), pre (2016–2019) versus post (2021–2022).
Figure A2: Difference‐in‐difference event study coefficient plots of the impact of continuous coverage provision on insurance rates: low‐income noncitizens (vs. citizens) 2016–2022 (reference year: 2019).
Table A3: Full result table of event study estimation of the impact of continuous coverage provision on insurance rates: low‐income citizens versus noncitizens, 2016–2022 (reference year: 2019).
Table A4: Pre versus post comparisons of sample characteristics by state policy group.
Figure A3: Trends in uninsurance, Medicaid, and private insurance rates among low‐income citizen and noncitizen postpartum women in from‐conception‐to‐the‐end‐of‐pregnancy (FCEP)‐only states versus states with various pathways, 2016–2022.
Table A7: Difference‐in‐difference results of the impact of continuous coverage provision on insurance rates: low‐income noncitizens (vs. citizens), pre (2016–2019) versus post (2021–2022) in states with various pathways.
Table A8: Event study full results: low‐income noncitizens versus citizens in from‐conception‐to‐the‐end‐of‐pregnancy (FCEP)‐only states, 2016–2022.
Table A9: Event study full results: low‐income noncitizens versus citizens in states with various pathways, 2016–2022.
Table A10: Event study full results: Hispanic low‐income noncitizens versus citizens in from‐conception‐to‐the‐end‐of‐pregnancy (FCEP)‐only states, 2016–2022.
Table A11: Event study full results: Hispanic low‐income noncitizens versus citizens in states with various pathways, 2016–2022.
Figure A4: Coefficient plots from triple difference‐in‐differences estimates of the impact of continuous coverage provision: citizen versus noncitizen postpartum women in from‐conception‐to‐the‐end‐of‐pregnancy (FCEP)‐only states versus states with various pathways.
Table A12: Triple difference‐in‐differences event study full results: pre versus post, low‐income citizen versus noncitizen, from‐conception‐to‐the‐end‐of‐pregnancy (FCEP)‐only states versus states with various pathways.
Table A13: Triple difference‐in‐differences event study full results: pre versus post, Hispanic low‐income citizen versus noncitizen, from‐conception‐to‐the‐end‐of‐pregnancy (FCEP)‐only states versus states with various pathways.
Data Availability Statement
The data that support the findings of this study are available in IPUMS USA at https://www.ipums.org/. These data were derived from the following resources available in the public domain: IPUMS USA, https://www.ipums.org/.