Introduction

Latinx children—more than half of whom are children of immigrants—are one of the fastest-growing U.S. child demographic groups, and have become increasingly dispersed geographically over time (Clarke et al., 2017; Johnson & Lichter, 2010). Over the last 30 years, Latinx families have increasingly settled in new urban, suburban, and rural areas throughout the U.S. (Johnson & Lichter, 2010; Lichter & Johnson, 2009; Massey & Capoferro, 2008). As a result, over one-quarter of the U.S. Latinx population now lives outside of an “established” Latinx settlement area (Lichter & Johnson, 2009). This geographic dispersion could be consequential for Latinx children’s health outcomes, including health care access (which we hereto also refer to broadly as “health access”).

Latinx children face some of the largest barriers to receiving regular health care in the U.S.—they are less likely to be insured and have a usual source of care, and are more likely to have delayed care, than children from most other racial/ethnic groups, even amidst the implementation of the Affordable Care Act (ACA; Branch & Conway, 2022; Flores & Tomany-Korman, 2008; Flores et al., 2005; Guzman et al., 2020; Langellier et al., 2016; Ortega et al., 2018; Perreira et al., 2021; Whitener & Corcoran, 2021). These patterns, however, may differ across U.S. Latinx destinations—geographic areas that vary in the historical presence and recent growth of the Latinx population. In particular, Latinx children living in newer Latinx destinations—areas with limited historical Latinx presence and high recent growth—may face more health access barriers than their peers in established destinations. Many new Latinx settlement destinations are located in places—the Southeast and rural areas—with limited health care resources, weak immigrant support systems, and strong anti-immigrant hostility (Massey, 2008). In contrast, established Latinx destinations—areas with an extensive historical Latinx presence—tend to have more health care resources and policies that promote health access for all children (e.g., Medicaid expansion, FQHCs), not just for Latinx children, as well as the co-ethnic presence and resources to help Latinx families and children navigate the local health care context (Ackert et al., 2021; Gresenz et al., 2012a; Waters & Jiménez, 2005). Consequently, research finds that Latinx adults in new versus established destinations are less likely to be insured and more likely to delay care (Gresenz et al., 2012; Monnat, 2017).

Research, however, has yet to examine if these same Latinx adult health access disparities in new versus established destinations extend to Latinx children. Because Latinx children—90% of whom are U.S.-born citizens—are more likely to qualify for public health insurance, they may face fewer health access barriers than adults in new destinations (Murphey et al., 2014), thus potentially mitigating Latinx children’s health access disparities across destinations. Then again, as with Latinx adults, new destinations may still present more barriers to health access for Latinx children—even for those children who are U.S. born—than established destinations, due to destination differences in factors that affect all children, including institutional and policy contexts for health access, and factors that specifically affect children of immigrants, including immigrant hostilities, targeted health resources, and the size of co-ethnic support networks (Ackert et al., 2021; De Jong et al., 2017; Ebert & Ovink, 2014; Marrow et al., 2022; Perreira et al., 2020; Waters & Jiménez, 2005).

To assess this issue, we use an integrated, multilevel dataset, with health access measures from the restricted-use National Health Interview Survey (NHIS), to compare how three health access outcomes—health insurance coverage, delayed care, and usual place of care—among Latinx children and their peers differ across Latinx settlement destinations at the county level. Following prior research (Ackert et al., 2019), we use a 4-category destination typology—established, fast-growing hubs, new, and minor—to capture county-level differences in both the historical presence and recent growth of the Latinx population. Our primary focus is on comparing new and established destinations. However, because research finds that some established destinations continue to experience high Latinx growth, with growth-strain challenges similar to those of new destinations, we distinguish between established destinations (large historical Latinx presence, low-to-median Latinx growth) and fast-growing hubs (large historical Latinx presence, high Latinx growth). We examine whether children in general, and Latinx children of immigrants in particular, have more negative health access outcomes in new destination counties versus historically established (i.e., established and fast-growing hub) counties.

Background on Latinx Children’s Health Access and Settlement Patterns

U.S. Latinx children and their families face several major barriers to health and well-being because of their lower access to health care than other groups. Even though they are more likely to be insured than their parents, Latinx children are twice as likely as non-Latinx children to be uninsured—a health insurance disparity that has persisted over time and increased in recent years, despite the implementation of the ACA, especially among Latinx children with immigrant parents (Branch & Conway, 2022; Guzman et al., 2020; Langellier et al., 2016; Ortega et al., 2018; Perreira et al., 2021; Whitener & Corcoran, 2021). Latinx children are also more likely to delay care, less likely to have a usual source of care, and less likely to go to doctors’ visits—including preventive care visits—than both non-Latinx White and Black children (Langellier et al., 2016), and gaps with non-Latinx Whites in health care utilization have remained even after the implementation of the ACA (Ortega et al., 2018).

U.S. geographic area of residence plays a key role in shaping health care access. For example, state of residence is strongly associated with Latinx child uninsurance levels, with the highest percentages of uninsured Latinx children in many newer areas of Latinx settlement in the South (e.g., Mississippi) and Midwest (e.g., Nebraska) (Perreira et al., 2021; Whitener & Corcoran, 2021). This state-level uninsurance pattern points to one important dimension of place of residence that could differentiate Latinx health access—the history of geographic areas as Latinx-receiving communities. Prior to 1990, the Latinx population was concentrated in areas along the U.S.-Mexico border, Florida, New York, and Illinois, but in the 1990s, increasing numbers were “pulled” into new urban and rural destinations across the country, particularly the U.S. Southeast (Massey, 2008; Zúñiga & Hernández-León, 2005). This diversity of settlement is most pronounced among Latinx children, who are increasingly born outside of established destination areas (Johnson & Lichter, 2008, 2010).

How Latinx Settlement Destinations Shape All Children’s Health Access

Living in a new versus an established Latinx destination area may hinder health access for all children, not just for Latinx children, for several reasons. First, evidence suggests that health care resources are often more limited in some new Latinx destinations compared to established destinations. For instance, research finds that the availability of health care personnel (e.g., primary care physicians) and facilities (e.g., federally qualified health centers [FQHCs]) varies across U.S. communities (Gaskin et al., 2012; Ko & Ponce, 2013; Ko et al., 2014) and are more limited in rural areas and the U.S. Southeast—regions where many new Latinx destinations are located. Rural areas often have strained and understaffed health departments (Harris et al., 2016), and states in the Southeast rank among the lowest in the nation in health care access and quality (Kaiser Family Foundation, 2021; Ziegler, n.d. 2019). Even though physician shortages are less prevalent in new versus established Latinx destinations (Ackert et al., 2021; Gresenz et al., 2012b), affordable health care options, such as FQHCs, are more limited in these areas (Ackert et al., 2021). Consequently, families and children in new versus established destinations may not be able to find affordable care options, even if physician supply in these areas is sufficient overall.

State health care policies are another challenge that may hinder health access in new Latinx destinations for all children, not just for Latinx children. Health care policies, especially those related to health insurance coverage, vary substantially across U.S. states, leading to state variation in health insurance rates (Conway, 2019; Kaiser Family Foundation, 2020). For instance, despite federal 2008 provisions in the ACA to expand Medicaid coverage (public health insurance) in all states, most states in the U.S. Southeast—the region most impacted by Latinx dispersion—have refused to do so. Such policy restrictions could be harmful for all children, but may be particularly harmful for Latinx children. Evidence indicates that Latinx children living in Medicaid non-expansion versus expansion states are 2.5 times more likely to be uninsured, and within Medicaid non-expansion states Latinx children are 2.5 times more likely to be uninsured than non-Latinos/as (Whitener & Corcoran, 2021). Besides Texas and Florida, all of the states that have not expanded Medicaid (most of them in the U.S. South and Southeast) include many new Latinx destination communities (Kaiser Family Foundation, 2020).

How Latinx Settlement Destinations Shape Latinx Children’s Health Access

New Latinx destinations may pose additional health access barriers for Latinx children, specifically, because of differences with established destinations in Latinx group size, immigrant/Latinx targeted institutional resources, and intergroup relations (Waters & Jiménez, 2005). Given their rich Latinx settlement histories, established destinations are characterized by large Latinx populations and rooted Latinx social networks (Portes & Rumbaut, 2006). Established destinations are therefore more likely than new destinations to have group-specific institutional resources (e.g., bilingual medical professionals) as well as shared co-ethnic knowledge of how to access available resources (i.e., shared information about insurance and care options). In contrast, many health and social services in new destinations lack the support systems to meet the health care needs of immigrant-origin families (Derose et al., 2007), and they have fewer of the “safety net” care resources that immigrants are most likely to utilize, such as community health centers (Ackert et al., 2021; Parker, 2021). A lack of group-specific resources may also affect U.S.-born Latinos/as. Indeed, prior research finds that U.S.-born Mexican American adults in new destinations are more dissatisfied with the care providers that they have accessed than those in established destinations (Gresenz et al., 2012b). Because new destinations have smaller but rapidly growing Latinx populations and networks (Lichter & Johnson, 2009), the Latinx population may have less shared co-ethnic knowledge to access the health policies and services that do exist in these newer areas of settlement.

Latinx families in new destinations may also face more negative and hostile intergroup relations than those in established areas, which could lead to lower uptake of available insurance policies and the under-utilization of existing health care resources among Latinx families. Discrimination towards immigrants and immigrant-origin groups, no matter where it occurs, can negatively affect health and well-being (Krieger, 2014; Samari, 2016). Discrimination and negative intergroup relations may specifically affect Latinx child health access in two ways. First, states may discriminate against Latinx families by limiting access to health insurance for Latinx subgroups such as low-income undocumented children or undocumented pregnant women (Ayón, 2015; Samari et al., 2021). Second, at both the state and local levels, discriminatory policies and perceived discrimination may reduce health care utilization among targeted groups (Alcalá & Cook, 2018; Perreira & Pedroza, 2019).

Even though many established destinations have been described as discriminatory places (Telles & Ortiz, 2008), several studies suggest that Latinos/a in new destinations face even higher levels of discrimination than those in established destinations. For example, new destinations have higher Latinx-White residential segregation levels, a more negative “receptivity climate,” and more anti-immigrant policies than established destinations (De Jong et al., 2017; Hall, 2013; Lichter et al., 2010; Walker & Leitner, 2011; Wong, 2012). Moreover, Latinx families in new destinations may be harmed more than those in established destinations by similar levels of immigrant hostilities, because there are fewer co-ethnics to mitigate discrimination effects on access to health insurance and health care (Ebert & Ovink, 2014).

Hypotheses

Based on the prior literature, we test several hypotheses regarding the associations between Latinx destinations and three health access outcomes (uninsurance, delayed care, and having a regular place of care). Overall, we expect more negative health access outcomes (i.e., lower health insurance coverage, more delayed care, and less likely to have a usual place of care) in new versus established Latinx destinations.

Hypothesis 1:

Children of all races/ethnicities in new Latinx destination counties will have more negative health access outcomes than children in established and fast-growing hub counties due to the more limited health care resources in these destinations.

Hypothesis 2:

The health access disparities between children in new versus established and fast-growing hub Latinx destinations will be even wider for Latinx children than for non-Latinx White children of U.S.-born parents, particularly Latinx children of immigrants, because of heightened health access barriers for Latinx and immigrant populations in new destination areas.

Hypothesis 3:

Differences in local immigrant hostilities and health care resources will partially account for differences in health access outcomes across destinations. Adjusting models used to test Hypotheses 1 and 2 for differences in these factors will attenuate between-destination differences in health access outcomes.

Data and Methods

Data

We use individual-level data from the restricted-access National Health Interview Survey (NHIS) and a comprehensive county-level dataset. We measure Latinx destinations at the county level to capture within-state variation in Latinx settlement, localized immigrant hostilities, and local health care resources.

National Health Interview Survey

To measure individual-level health access outcomes and control variables, we use the 2010–2014 NHIS, conducted by the National Center for Health Statistics (NCHS), which is the principle source of U.S. population health data. We first pool multiple data years to ensure sufficient sample size, using the public-use integrated NHIS data from IPUMS (Blewett et al., 2021). Using a multi-stage stratified sample, the NHIS collects yearly information on a broad range of health and healthcare topics from approximately 40,000 housing units containing 100,000 individuals. The survey contains a set of core demographic and health questions for all household members, and supplementary questionnaires for a subsample of one randomly selected child and adult per household. The individual-level public-use NHIS data is then merged into the restricted-access NHIS data via NHIS household and person identifiers. We access the restricted-access NHIS data at the Missouri Federal Statistical Research Data Center (RDC). We requested the use of geographic variables (state and county) in the restricted-access NHIS through the RDC to merge county-level data on Latinx destinations with NHIS individual-level data on children’s health access.

The advantage of using the restricted-access NHIS is that it is the only nationally-representative dataset that links geographic place of residence to a broad range of health and health access indicators, which is essential for this analysis. The disadvantages are that these data are highly restricted. The data can only be used in a secure, onsite data facility, which creates additional time delays, particularly during COVID-19, and due to distance limitations. Users must undergo lengthy processes to access the data and to have all output, results, and manuscripts approved by NCHS for dissemination. (As a disclaimer, it is also important to note that the findings and conclusions in this research are those of the author[s] and do not necessarily represent the views of the Research Data Center, the National Center for Health Statistics, or the Centers for Disease Control and Prevention.) For these reasons, we are only able to include NHIS data through 2014. However, we complement this data by running robustness checks using more recent data from the American Community Survey (ACS 5-year estimates, 2015–2019), which has information on children’s health insurance status, but not other health access information. Overall, the results for destinations are consistent across these two data sources and time periods.

County-Level Dataset

Our county-level dataset integrates information from four data sources: The American Community Survey (ACS; 5-year estimates), U.S. Decennial Censuses (1990, 2000, 2010), Health Resources and Services Administration (HRSA) data, and a self-compilation of internal immigration enforcement policies. The U.S. Decennial Census data provides needed information to classify counties into a Latinx destination typology (e.g., Latinx population growth rates). We use the ACS data to capture county-level socio-demographic differences (e.g., poverty). The HRSA data, managed by the US Department of Health and Human Services, offers comprehensive details on healthcare services, workforce, funding, and access. Our specific focus revolves around county-level health resources, particularly data concerning the healthcare workforce and services. For this reason, we access data on the total number of active MDs in the county from the HRSA Area Health Resources File, which provides county-level tabulations of the number of people working in health service occupations (U.S. Health Resources & Services Administration, 2020). Lastly, we use self-compiled data on different internal immigration enforcement policies to proxy immigrant hostilities. These data are based on a variety of cross-checked sources, including FOIAs, ICE website data, and public datasets (Pedroza, 2019) and organizations that track immigration enforcement policies (Syracuse Transactional Records Access Clearinghouse, n.d. 2021). All county-level measures except those from the decennial censuses are measured annually from 2010 to 2014.

Sample

We focus on children ages 4 to 17 for which NHIS has consistent health access measures. Because results did not differ between younger (age 4–11) and older (age 12–17) children, we combine these age groups. Based on reports of child race/ethnicity and parents’ country of birth, we identify Latinx children—divided into children of immigrants (one or both parents foreign-born; n = 22,689) and children of U.S.-born parents (both parents U.S.-born; n = 10,572)—and non-Latinx White (n = 42,292) and Black (n = 14,343) children of U.S.-born parents. The full sample size (n = 89,896) is lower for the delayed care outcome (n = 89,838) due to missing data, and for the no usual place of care outcome (n = 41,663) because this measure is only available in the NHIS child sub-sample. To reduce missing data, we create a dummy variable for missing on parent education (n = 3,518). No other data were missing. Results were robust to different missing data approaches. Notably, we found similar results using multiply-imputed data (see Appendix A).

Measures

Health Care Access

Following the prior literature (Artiga & Orgera, 2019; Livingston, 2009; Office of Disease Prevention & Health Promotion, 2020), we use three dichotomous outcomes to comprehensively capture children’s health care access: (1) child has no health insurance (versus has public or private insurance); (2) child experienced delayed medical care due to cost in the past 12 months; and (3) child has no usual place of medical care. The first two measures are available for all NHIS respondents and the latter for the child sub-sample only.

Latinx Destinations

Following prior research on immigrant and Latinx destinations (Ackert, 2017; Hall, 2013), using U.S. Decennial Census data, we define destination type based on the “base” size of the Latinx population in 1990, and the percent growth in the Latinx population from 1990 to 2010. Additionally, because there is wide variation in Latinx growth rates among new and established destinations, we further divide each of these categories into two sub-groups based on differential growth rates from 1990 to 2010. In total, we identify four Latinx destination types: (1) Established, (2) fast-growing hub, (3) new, and (4) minor.

To create these groups, we first divide counties into two broad “established” and “new” Latinx destination categories, based on if a county falls above or below the average U.S. Latinx population size in 1990 (9% Latinx): (1) “established” counties fall at or above the average (> = 9% Latinx in 1990); and (2) “new” counties fall below the average (< 9% Latinx in 1990). Then, for each group separately, we calculate Latinx growth rates from 1990 to 2010 and set median cut-off levels that identify low versus medium/high growth rates within each group. For the final typology: (1) Established destinations are “established” counties that fall below the median growth rate for all “established” counties (< 70% growth); (2) Fast-growing hubs are “established” counties that fall at or above the median growth rate for all “established” counties (> = 70% growth); (3) New destinations are “new” counties that fall at or above the median growth for all “new” counties (> = 272% growth); and (4) Minor destinations are “new” counties that fall below the median growth rate for all “new” counties (< 272% growth). A map of this destination typology is available in Appendix B.

Our destination measurement approach, especially our categorization of new and established destinations, is largely consistent with prior research, though not all studies distinguish between “established” and “fast-growing” hub categories (Ackert et al., 2019; Hall, 2013; Lichter & Johnson, 2009; Stamps & Bohon, 2006). Note that our results were robust to alternative typology classification specifications that adjust for potential cut-off challenges (e.g., exaggerated growth rates due to small denominator challenges in counties with a very small 1990 Latinx total number). To account for potential cut-off challenges, we made the following revisions: (1) Re-classified new destinations as minor if their 2010 Latinx population was < 5%, and (2) re-classified minor destinations as new if their 2010 Latinx population was > 16% (the national average). Our results were similar using these revised cut-offs (available upon request).

County Health Care Resources

To capture variation in county health care resources, we use the HRSA data and include an indicator of the number of MD physicians per 1,000 residents. Because total number of active MDs are only available for years 2010 and 2014, we use linear interpolation to create estimates for in-between years. The denominator for the per capita measure is the estimate of the total county population from the census in 2010. We examined other measures of health care service resources (e.g., number of hospitals per 10,000 residents, county health shortage status), but ultimately excluded these measures due to high collinearity with the total number of active MDs per capita measure (e.g., r = 0.87 for total active MDs and total hospitals per capita).

County Immigrant/Latinx Hostility

To capture county-level variation in immigrant/Latinx hostilities, we use our self-compiled data and the local internal immigration enforcement program, 287(g), as our primary measure. Using publicly available data on 287(g) program locations (Pedroza, 2019), we classify counties based on if they ever had a 287(g) agreement (1 = yes; 0 = no) from 2003 to 2014. Adopted in 1996, the 287(g) program delegated federal immigration powers to local law enforcement for the first time. This program and subsequent local internal enforcement policies (e.g., Secure Communities) primarily targeted Latinos/as and created significant fear and feelings of hostility among Latinos/as, no matter their citizenship status (Capps et al., 2015; Perreira & Pedroza, 2019). Compared to other local internal enforcement policies, 287(g) best captures localized immigrant/Latinx hostilities because counties had to actively apply for the program, suggesting that immigration enforcement efforts in these counties were internally motivated (Capps et al., 2011, 2015; Rhodes et al., 2015). We also explored including other internal enforcement policy measures, including county-level detention rates from the Secure Communities program (Pedroza, 2013), whether the county had experienced a federal immigration raid (Santillano et al., 2020), and if the county had adopted a local-level anti-immigrant policy (e.g., landlord laws) (O’Neil, 2011). These measures, however, were largely non-significant and had a weaker association with our focal outcomes than the 287(g) measure. Consequently, adding them to our models did not change our results; thus, for simplicity, we include the 287(g) measure only.

Individual and County Control Measures

To control for potential child and familial demographic differences across destinations that may influence health access, we include child’s age and gender, highest parental level of education (less than high school, high school, some college, or college or higher), family structure (two-parent, single-parent, or other), child’s self-rated health status based on parental reports (fair/poor health versus good/very good/excellent health indicator), and, in the delayed and usual place of care analysis, child health insurance coverage (any public, private only, none). To capture potential immigrant-related demographic differences, we control for children’s and household members’ citizenship status using the following dummy indicators: (1) non-citizen child (no matter other household members’ citizenship status), (2) citizen child in a household with mixed-citizenship statuses, and (3) citizen child in a household with all citizens (reference category). In sensitivity analyses, we included additional immigrant-related indicators—children’s Latinx heritage country background, household poverty status, head of households’ years in the U.S., and whether the interview was completed in English (the best proxy for English language ability)—and found similar results (i.e., their inclusion did not explain cross-destination differences). Lastly, we use ACS data to control for potential county-level compositional differences that may differ across destinations and also influence health access, including racial/ethnic composition (proportion of non-Latinx Whites), educational composition (proportion of adults 25 + with a BA), economic composition (proportion of the population in poverty), and urbanicity (1 = metropolitan; 0 = non-metropolitan).

Analysis

We use descriptive statistics to examine overall differences in children’s health care access across racial/ethnic groups and different Latinx settlement destinations. We then use logistic regression analyses (odds ratios reported) to examine how overall differences in children’s health care access across settlement destinations are shaped by differences in child, family, and county-level demographics, physician supply, and 287(g) agreements. We estimate the following general model:

$${Y}_{ijt}={\alpha }_{0}+ {\beta }_{1}{Dest}_{jt}+{\beta }_{2}{Race}_{ijt}+{\beta }_{3}{Ind}_{ijt}+{\beta }_{4}{Cnty}_{jt}+ {\beta }_{5}{T}_{t}+ {\varepsilon }_{ijt}$$

where i indexes individuals, j indexes counties, and t indexes year. Yijt is the outcome variable of interest; Destjt is a vector of three dummies indicating our classification of Latinx destination type (Established is the reference category); Raceijt is vector of three dummies that capture children’s race/ethnicity and parental nativity status (Non-Latinx White is the reference category); Indijt is a vector of individual-level controls; Cntyjt is a vector of county-level focal predictors and controls; \({T}_{t}\) captures year fixed effects; and εijt is an error term. In this model, the coefficients on the Destjt variables captured the mean differences in health care access between Latinx destination types (with established as the reference).

Similar to other studies on new destinations (Ackert et al., 2019; Monnat, 2017), we use different models to assess how the addition of different theoretical blocks of variables shape health care access disparities across destinations. Model 1 provides a baseline assessment and includes our Latinx settlement destination categories and race/ethnicity and parent nativity controls. Model 2 captures differences in individual and family compositional background across destinations, and Model 3 captures county-level compositional differences. Model 4 adds our two focal county-level predictors (total MDs per 1,000 residents and 287[g] adoption) to assess the influence of settlement differences in health care resources and immigrant hostilities. In sensitivity analyses, we capture differences in health care resources and immigrant hostilities at the state-level by including state fixed effects. In the delayed and usual place of care analyses, only, we add an additional model (Model 5) that includes children’s health insurance status, because access to health insurance is likely to influence children’s use and place of care.

Lastly, using our full model with all controls, we add interactions between racial/ethnic and parental nativity groups and destinations to see if destination associations with the health access measures vary across child sub-groups, particularly Latinx children of immigrants. We do so by adjusting our equation as follows:

$${Y}_{ijt}={\alpha }_{0}+ {{\beta }_{1}{Dest}_{jt}*{Race}_{ijt}+\beta }_{2}{Dest}_{jt}+{\beta }_{3}{Race}_{ijt}+{\beta }_{4}{Ind}_{ijt}+{\beta }_{5}{Cnty}_{jt}+ {\beta }_{6}{T}_{t}+ {\varepsilon }_{ijt}$$

In this equation, the coefficients on the Destjt * Raceijt interaction terms capture racial/ethnic and parent nativity group-specific differences across destinations. All models include year fixed effects, and incorporate NHIS survey weights and adjust for county-level clustering based on IPUMS guidelines for using multiple survey years (IPUMS, 2022).

Results

Health Care Access Differences Among Children: Summary Statistics

We find that Latinx children, no matter their parental nativity, face more health access barriers than their non-Latinx White and Black peers (Fig. 1). Latinx children of immigrants and U.S.-born parents are more likely than non-Latinx White and Black children to be uninsured, experience delayed care, and have no usual place of care. Disparities in health uninsurance and no usual place of care are particularly striking for Latinx children of immigrants. Compared to non-Latinx White and Black children, Latinx children of immigrants are more than 2.5 times more likely to be uninsured (5.6% vs. 5.7% vs. 16.4%, respectively) and have no usual place of care (2.8% vs. 2.3% vs. 9%, respectively).

Fig. 1
figure 1

Disparities in Children's Health Care Access by Parent Nativity and Race/Ethnicity

Full size image

Focusing on the role of place, Table 1 shows summary statistics for children by Latinx settlement destination. Overall, the results reveal health access disadvantages associated with living in established destinations and fast-growing Latinx hubs versus new and minor Latinx destinations, which is contrary to our expectations. For instance, children in established destinations face more health insurance coverage barriers. They are more likely to lack health insurance (10%) than children in new (8%) and minor (6%) destinations. Additionally, children in fast-growing Latinx hubs face more health care utilization barriers than those in established destinations. Children in fast-growing hubs are more likely to experience delayed care (4%) and have no usual place of care (6%) than children in established destinations (3% and 4%, respectively), whereas there is no significant difference in observed levels of delayed care and usual place of care between children in new and established destinations.

Table 1 Summary Statistics of Children (Ages 4–17) by Settlement Destination, Data NHIS 2010–2014
Full size table

Some of these place-based health care access differences may reflect compositional differences between children and families across destinations, as well as county-level differences related to immigrant hostilities and health care resources. For instance, as expected, a much larger share of children living in new (69%) and minor (72%) Latinx destinations are non-Latinx White children—a racial/ethnic group with strong health care access—than in established destinations (19%) and fast-growing hubs (34%). Furthermore, relative to those in established destinations, children in new and minor destinations exhibit a number of factors that could improve health access: a) Their parents have higher levels of education; b) they are more likely to live in an all-citizen household and to have private health insurance; and c) they live in counties with higher proportions of Whites and lower poverty. They also tend to be slightly healthier: Reports of fair/poor health are slightly lower in new and minor destinations (2%) versus established and fast-growing hubs (3%).

Interestingly, we find that the share of children exposed to immigrant/Latinx hostility—as measured by the 287(g) program—is smaller in new (12%) and minor Latinx destinations (6%) than in established destinations (38%) and fast-growing hubs (39%). In contrast, health care resources—as measured by total MDs per capita—are lower for children in new (2.52 MDs per 1,000 residents) and fast-growing hubs (2.60) than in established destinations (3.24).

Lack of Health Insurance Across Destinations: Multivariate Results

We use multivariate logistic regression to control for compositional and contextual differences across settlement destinations and start with an assessment of children’s health insurance coverage. Table 2 presents odds ratios predicting the likelihood that children have no health insurance, and focuses on our primary variables of interest (see Appendix C for full regression model results). Overall, we find that, net of race/ethnicity, children in both new destinations and fast-growing hubs are more likely than children in established destinations to lack health insurance. This disparity between new and established destinations was anticipated, whereas the difference between established destinations and fast-growing hubs was not.

Table 2 Logit Models (Odds Ratios) of Latinx Settlement Destination on No Health Insurance for Children Ages 4–17, NHIS 2010–2014
Full size table

In Model 1, the odds of having no health insurance are 1.37 (p < 0.01) and 1.30 (p < 0.05) greater for children living in new destinations and fast-growing hubs (respectively) than for children in established destinations. Thus, once we account for racial/ethnic demographic differences across destinations—notably, the fact that the majority of children in new Latinx destinations are non-Latinx White (Table 1) and non-Latinx White children have higher health insurance rates (Fig. 1)—we find a health insurance disadvantage associated with living in places with higher recent Latinx growth—in new destinations and fast-growing hubs—relative to living in an established destination.

This new destination and fast-growing hub health insurance disadvantage relative to established destinations persists across subsequent models, and is not explained by cross-destination demographic differences (including potential differences in children’s self-reported health) or county-level differences. The odds ratios on both new destinations and fast-growing hubs remain relatively unchanged and statistically significant with the addition of child and household (Model 2) and county-level (Model 3) controls. The addition of county-level immigrant hostility and health care resource controls (Model 4) changes the results slightly, but still suggests an overall health insurance disadvantage in new destinations and fast-growing hubs. For fast-growing hubs, the odds ratio remains relatively stable across models (1.29 and 1.27, respectively) but is no longer statistically significant at conventional levels (p = 0.103), which means that the observed health insurance disadvantage (i.e., the 1.27 odds ratio) is only suggestive. For new destinations, the health insurance disadvantage is robust.

Lastly, though county-level immigrant hostility and health care resources shape children’s health insurance coverage, we do not find strong evidence that differences in these factors explain cross-destination differences in children’s health insurance coverage. As expected, we find that immigrant hostility heightens children’s uninsurance risk, while health care resources reduce this risk. For instance, children who live in counties with 287(g) agreements have 34% higher odds of being uninsured than children in counties without these agreements (p < .01). In contrast, children are less likely to be uninsured if they live in a county with more MDs per capita (OR = 0.95; p < 05), which could possibly be due to reverse causation, with MD supply responding to demand from insured patients. The addition of these variables, however, does not explain the observed health insurance disadvantage in new destinations or fast-growing hubs—the odds ratios for both remain relatively robust in Model 4.

Delayed Care and No Usual Place of Care Across Destinations: Multiple Regression Results

Table 3 Parts A and B examine how health care utilization—delayed care (Part A) and no usual place of care (Part B)—differ across Latinx settlement destinations (see Appendix C for full regression results). Because health insurance impacts health care use, we add this variable to the model. The results are largely consistent across outcomes, and suggest three main trends. First, similar to disparities in health insurance across destinations, we find that once we account for demographic differences (including potential differences in children’s self-reported health) across places, children in new destinations and fast-growing hubs have lower, not higher, health care utilization (i.e., more delayed care and/or less likely to have a usual place of care) than children in established destinations. Results in Model 1 show that, net of race/ethnicity and parental nativity, children in new destinations and fast-growing hubs have higher odds of having no usual place of care (OR = 1.57 and 1.75, respectively) than those in established destinations. Children in fast-growing hubs also have higher odds of delaying care (OR = 1.28). These results remain robust with the addition of child and household controls (Model 2).

Table 3 Logit Models (Odds Ratios) of Latinx Settlement Destination on Delayed Care and No Usual Place of Care for Children Ages 4–17, NHIS 2010–2014
Full size table

Second, the results in Model 3, which adds children’s health insurance status, suggest that part of the reason health care utilization rates are lower in new and fast-growing hubs versus established destinations is because health insurance rates are lower in these areas (as seen in Table 2). In Model 3, we find that children who have no health insurance are much more likely to delay care (OR = 10.20; p < 0.001) and to have no usual place of care (OR = 10.19; p < 0.001). Once we control for adjusted place-based disparities in children’s health insurance coverage—which was lower in new destinations and fast-growing hubs (see Table 2)—we find smaller health care utilization disparities across destinations. Between Model 2 and 3, the odds ratios of having no usual place of care for children living in new destinations and fast-growing hubs (versus established destinations) respectively decline by 12% [(1.38–1.56)/1.56*100] and by 9% [(1.64–1.80)/1.80*100]. For fast-growing hubs, the odds ratio for delayed care also declines and becomes non-significant. Thus, part of the reason why children in new destinations and fast-growing hubs have lower health care utilization (i.e., experience more delayed care and/or more likely to have no usual place of care) compared to those in established destinations may be because they are less likely to have health insurance in these areas.

Finally, similar to the health insurance models, we find that county-level health-care resources are associated with children’s health care use, but do not explain variation across destinations (Model 5 versus Model 4). Results for county-level immigrant hostilities were not statistically significant. With all controls added, we continue to find a health care use disadvantage for children living in new destinations (usual place of care only) and fast-growing hubs (both delayed care and no usual place of care) relative to children living in established destinations.

Health Care Access Across Destinations: The Role of Race/Ethnicity and Parental Nativity

Next, we examine whether health care access disparities across destinations are stronger for Latinx children, specifically, by adding interactions between our destination and race/ethnicity and parental nativity measures to the full model (all controls included). Because none of the interaction results for delayed care and usual place of care were statistically significant, we focus on the health insurance results (see Appendix D for full results). To aid in interpretation, Fig. 2 displays predicted probabilities calculated from full logit models.

Fig. 2
figure 2

Predicted Probability of Children with No Health Insurance by Settlement Destination, Parent Nativity, and Race/Ethnicity

Full size image

Overall, we find that all Latinx children of immigrants living outside of established destinations face a health insurance disadvantage, and that living in a new destination is particularly detrimental, even for Latinx and Black children of U.S-born parents. Figure 2 shows that the gap in the predicted probability of being uninsured is the widest between Latinx children of immigrants living in new (11%) versus established destinations (6%). A similar, but smaller health insurance gap exists for Latinx children of immigrants living in fast-growing hubs and minor destinations (9% and 8%, respectively, are predicted to be uninsured). Lastly, even among Latinx and Black children of U.S.-born parents, predicted uninsurance rates are slightly higher in new (8% and 6%, respectively) versus established destinations (6% and 4%, respectively).

Sensitivity Analysis

We ran two sensitivity checks on our main model of interest (i.e., our non-interactive, full models from Tables 2 and 3), to ensure the robustness of our results. Our first check demonstrates that our results remain robust using an alternative Latinx destination classification scheme that accounts for potential variation in health care access within new destinations (Appendix E). Following Monnat (2017), we disaggregate our new destination category into early vs. recent new destinations, based on whether rapid Latinx population growth (> 150% or more) occurred between 1990–2000 versus 1990–2010, respectively. Like our main results, we find that children in new destinations—both early and recent—face more health care utilization challenges (i.e., are more likely to have no usual place of care) than children in established destinations, and that children in early new destinations are also more likely to be un-insured. Overall, these results confirm that health access disparities are higher in new destinations and may be particularly stark in early-emerging new destinations.

Our second check demonstrates that our results are robust to the inclusion of state fixed effects, which adjust for potential unobserved state-level differences in health care resources and policies (e.g., Medicaid) and immigrant hostility that may be driving our county-level results (Appendix F). Overall, we find that even after adjusting for state-level differences, children in new destinations and fast-growing hubs still face more health care access disparities (i.e., they are less likely to have insurance and/or more likely to have no usual place of care and delayed care) than children in established destinations. In fact, these results are often stronger in sensitivity analyses.

Discussion and Conclusion

Discussion

Latinx children exhibit major health access challenges. Even after the implementation of the ACA, Latinx children—even those with U.S.-born parents—have some of the highest rates of uninsurance, delayed care, and lack of a usual place of care of any racial/ethnic group (Branch & Conway, 2022; Flores & Tomany-Korman, 2008; Flores et al., 2005; Guzman et al., 2020; Langellier et al., 2016; Ortega et al., 2018; Perreira et al., 2021; Whitener & Corcoran, 2021). As Latinx children increasingly grow up in diverse geographic contexts in the U.S., due to large-scale Latinx and immigrant dispersion to new destinations, a key question is how these different destinations are shaping their health access outcomes, which could also be a broader indicator of their overall potential for incorporation in these communities (Lichter, 2013; Waters & Jiménez, 2005). To address this issue, we examine three indicators of health care access—health insurance status, delayed care, and usual place of care—across four Latinx settlement destinations (established, fast-growing hub, new, and minor destination counties) for children (ages 4–17) overall and among four racial/ethnic and parent nativity groups (Latinx children of immigrants, Latinx children of U.S.-born parents, non-Latinx White children of U.S.-born parents, and non-Latinx Black children of U.S.-born parents). We pay specific attention to how two contextual factors associated with destinations—immigrant/Latinx hostilities and local health-care resources—contribute to these health access disparities across Latinx destinations.

We find that, net of racial/ethnic and other demographic differences, all children in new Latinx destinations and fast-growing Latinx hubs face more health access barriers than children in established destinations. These children are more likely than children living in established destinations to be uninsured, to experience delayed care, and to have no usual place of care (for fast-growing hubs only). These results largely remain robust to the addition of child, household, and county-level controls. These findings support the idea that destinations with high Latinx growth in recent decades impose barriers to health access for many children, not just Latinx children. Places with high recent Latinx growth, regardless of whether they are new or established areas of settlement, may be struggling to keep pace with the diverse health care needs of their changing child populations.

Part of the reason that children in new Latinx destinations and fast-growing hubs have lower health care utilization may be because they are less likely to have health insurance coverage. In multivariate results, we find that the new destination and fast-growing Latinx hub health insurance disadvantage among children, relative to peers in established destinations, persists in full models, and that children without health insurance coverage are significantly more likely to experience delayed care and to have no usual place of care. Moreover, the addition of health insurance coverage to multivariate models reduces the odds that children in new destinations and fast-growing hubs experience more delayed care and/or have no usual place of care compared to children in established Latinx destinations. Though future research is needed to confirm these results with formal mediation analysis, these results underscore the importance of ensuring that children across Latinx settlement destinations have equitable access to health insurance coverage as a means to utilize local health care services.

We did not find evidence that local immigrant hostilities, operationalized by the existence of 287(g) agreements, and community health resources, operationalized by total MDs per capita, explain disparities in health access across Latinx destinations. However, we do find that these factors shape children’s health care access overall. Children in counties with a 287(g) agreement are less likely to be insured. Conversely, children in counties with more MDs per capita are more likely to be insured and less likely to experience delayed care. These results align with prior evidence that highlight the important role of place-based immigrant hostilities and health care resources in shaping health care access (Cook et al., 2013; Perreira & Pedroza, 2019).

Lastly, our interaction results indicate that new Latinx destination health access disparities—particularly health insurance coverage—are greatest for Latinx children of immigrants, but also affect Latinx and Black children of U.S.-born parents in these areas. These results highlight unique concerns about child health care inequities in new Latinx destinations. No matter their parents’ nativity status, Latinx and Black children living in a new versus established Latinx settlement destination are more likely to be uninsured. After adjusting for these health insurance inequities, we do not find similar disparities in delayed care or usual place of care. These results indicate that new Latinx destinations, many of which are located in the U.S. Southeast, need to develop programs that better support their growing Latinx child population, while also addressing long-standing Black-White racial inequities that hinder Black child health care access (Winders & Smith, 2010).

Our results also reinforce that, similar to their adult Latinx immigrant counterparts, Latinx children of immigrants face unique health care access challenges when they reside outside of established Latinx destinations, particularly in new destinations. All Latinx children of immigrants living outside of an established Latinx destination—new, fast-growing hub, and minor destinations—are less likely to be insured than Latinx children of immigrants in an established destination. However, the gap is greatest for Latinx children of immigrants living in a new destination (a 5-percentage point difference). These results are consistent with prior findings of a disadvantage in health insurance coverage among Mexican American adults and health utilization among Latinx adults in new versus established destinations (Gresenz et al., 2012a; Monnat, 2017). The results are also consistent with several child-based research studies showing worrisome health, residential, and educational outcomes among the Latinx population—especially the immigrant contingent of this population—in new destinations (Ackert, 2017; Ackert et al., 2019; Fischer, 2010; Hall, 2013; Lichter et al., 2010).

The question of why Latinx children of immigrants outside of established destinations have such a health care access disadvantage is not fully answered by this study. The high rates of lacking health insurance among this group, even when netting out other observable factors, suggest that Latinx immigrant families face barriers to health insurance access in new and other non-established destination communities. These health insurance barriers could potentially be related to undocumented status, which is not measured here, but could be assessed in future studies. Another way to look at these results is to consider the relatively more advantaged position of Latinx children of immigrants in established gateways. There may be factors in established destinations that give Latinx children of immigrants a relative health access advantage in these locations. These factors could include co-ethnic support and institutional resources that are tailored towards meeting the needs of immigrant populations (Lichter & Johnson, 2009; Massey & Capoferro, 2008; Zúñiga & Hernández-León, 2005).

This research points to several areas for future inquiry. Future work, both quantitative and qualitative, should explore household and contextual factors that may be contributing to a heightened risk of not having health insurance among children outside of established destinations, particularly Latinx children of immigrants in new destinations. This research could examine variation in health insurance options available to immigrant families in these locations, as well as knowledge of insurance options and barriers to accessing insurance. Similarly, future work could discern why children of immigrants in established destinations are not as disadvantaged in their health access outcomes relative to those in new and fast-growing hub destinations, examining factors such as policy contexts, knowledge flow about policies among the co-ethnic community, and other institutional supports.

These future studies should also consider the changing nature of Latinx settlement destinations. New and established destinations are not stagnant, nor uniform. There are important differences both between (e.g., new vs. established) and within (e.g., established vs. Latinx hub) each Latinx settlement destination. Our study demonstrates that places with high Latinx growth, no matter if they are a new destination or more established Latinx hub, may be struggling to keep pace with changing child population health needs. Future studies should consider other important destination differences (e.g., recency of Latinx growth, origin country differences) that are likely to lead to differential child health access both between and within Latinx settlement destinations.

Limitations

This study has limitations that must be acknowledged. As a quantitative study, this work can unveil associations between destinations and Latinx child health access outcomes, but cannot speak fully to the mechanisms driving these associations. In particular, this study does not address the roles of social connections and interactions in contributing to these patterns, and qualitative, community-based research would be best-situated to examine these issues. Additionally, it is likely that geographic variation in unauthorized status contributes to health access differences across destinations, but we are unable to measure unauthorized status at the individual and household level in this study. It is also possible that geographic differences in other child health indicators besides parental reports of child health explain health care access and use patterns, such as the potential for children in new and fast-growing Latinx hubs to be healthier than children in established destinations, and thus to have less need for care. There are also several community-level factors that might contribute to cross-destination differences in Latinx child health access which are not examined here, including other types of hostilities towards immigrant communities besides the adoption of a 287(g) program (e.g., negative public attitudes), the nuances of health insurance and care markets within communities, and other health care resource indicators, such as the number of active nurse practitioners and physician assistants in a community, which may act as substitutes for MDs in some contexts.

Finally, because we are looking at the period 2010 to 2014, we used more recent data on child health insurance status from the American Community Survey (ACS 5-year estimates, 2015–2019, from IPUMS; Ruggles et al., 2022) to assess whether similar health disparities between new and established Latinx destinations persisted into the Trump era. With this robustness check, we find that children, especially Latinx children of immigrants, living in new destinations and fast-growing hubs are less likely to be insured than those in established destinations in this later period (Appendix G). Though the ACS does not have data on health care utilization, this robustness check suggests that similar geographic disparities in children’s health care utilization have persisted after the 2010–2014 period. Future work is needed to verify and extend this evidence, particularly into the Biden, COVID-19, and post-COVID-19 eras.

Conclusion

Overall, this study provides a greater understanding of health access among the growing and geographically diversified U.S. Latinx child population, which faces unique health access barriers. We document clear disadvantages in health access among Latinx children in new destinations, especially in obtaining health insurance coverage, but also point to barriers for Latinx children in fast-growing hubs, indicating that communities with high recent Latinx growth, regardless of historical Latinx presence, may struggle to meet the health needs of Latinx families. These findings for new destinations are worrisome when put in context with other research showing disadvantaged health, education, residential, and well-being outcomes among Latinx populations in new destinations (Ackert, 2017; Ackert et al., 2019; Fischer, 2010; Gresenz et al., 2012; Hall, 2013; Lichter et al., 2010; Monnat, 2017). Thus, Latino/a families in new destination areas may be important targets for interventions to support the health and well-being of the country’s fastest-growing racial/ethnic group.