Latino Children at the Intersection of Immigration and Health Care Policy: A Mixed-methods Study of Parental Documentation Status, State-level Policy, and Access to Coverage and Care

Abstract

Purpose: Latino children experience the highest uninsurance rate among their peers and those with immigrant parents experience the greatest coverage disparities. Most (60%) of these 10 million children belong to mixed-status families in which parents and children have different documentation statuses that confer differential rights, benefits, and vulnerabilities. Immigrant adults face significant restrictions on public coverage, and barriers created by documentation status suggest that policies intended to restrict access for immigrant adults may ultimately exclude their children, most of whom are U.S.-born citizens. Because of these federal restrictions – which are maintained under the ACA – the bulk of immigrant healthcare policymaking is ultimately delegated to states. My first objective was to estimate gaps in insurance coverage by parental documentation status among a nationally representative sample of citizen children of Latino immigrants. Second, in light of immigrant healthcare policymaking at the state level, I sought to examine whether disparate state-level healthcare policy moderated the relationship between parental documentation status and children’s coverage. My final objective was to explore the mechanisms through which parental documentation status affects children’s access to coverage and care. Methods: My mixed-methods sequential explanatory design integrated secondary data analysis with semi-structured interviews. In AIM 1 I used the Survey of Income & Program Participation (SIPP) to estimate differences in insurance by children’s citizenship and parental documentation status within a nationally-representative cross-section of 4,227 children of Latino immigrants. I pooled a cross-section of 1,260 children of Latino immigrants from the final wave of the 2004 SIPP Panel (August-December 2007) with 2,967 children from the 2nd wave of the 2008 Panel (December 2008-March 2009). I first estimated uninsurance and coverage type by children’s citizenship and parental documentation status. I then estimated binary and multinomial probit models to estimate the marginal effects of children’s citizenship, parental documentation status, and their interaction on the probability of being insured (binary probit) and the probability of being insured by employer-sponsored insurance (ESI), Medicaid/CHIP, and direct-purchase or other private coverage (multinomial probit). In AIM 2, I used multilevel models to examine whether state-level policy on immigrant access to coverage moderated the effect of parental documentation status among a sample of 3,615 citizen children of Latino immigrants in 30 states with sufficient sample size for multilevel modeling. AIM 3 consisted of semi-structured interviews with 14 Latino immigrant parents and 6 key informants in Minnesota, with the objective of uncovering mechanisms to help explain the relationship between parental documentation status and children’s coverage identified in AIMS 1 and 2. Findings: The children of Latino immigrants experienced high uninsurance rates and low rates of ESI. Non-citizen children fared the worst, with an uninsurance rates of 54.1% compared to 28.2% for citizen children (p<.001). Citizen children with at least one undocumented parent had lower rates of insurance than their counterparts (32% vs. 27% for citizen children with citizen/legal permanent resident (LPR) parents, p<.001). These differences were no longer significant after adjusting for age and immigration-related and socioeconomic barriers and facilitators. In adjusted multinomial models, citizen children with undocumented parents were significantly less likely to hold ESI coverage than citizen children with two citizen parents. State-level policy on immigrant access to prenatal coverage moderated the effect of parental documentation status. In states where all (income-eligible) pregnant women are eligible for Medicaid coverage regardless of immigration status, there were no differences in children’s uninsurance rates by parental documentation status. In these states, both children with at least one undocumented parent and their counterparts had an uninsurance rate of 26%. In states where undocumented pregnant women are not eligible for Medicaid prenatal coverage, 45% of citizen children with at least one undocumented parent were uninsured, 17 percentage points higher than children with citizen/LPR parents (p<.001). These differences held up in adjusted models. Finally, Latino immigrant parents for the most part did not feel their own documentation status affected their citizen children’s access to coverage. Two key policies in Minnesota help explain why parental documentation status was not identified as a major barrier to coverage for citizen children. Prenatal coverage is available to all income-eligible pregnant women, regardless of immigration status, and newborns are automatically enrolled in Medicaid/CHIP when their mother is covered by the same at birth. In contrast, undocumented children are restricted from Medicaid/CHIP coverage, and as a result faced the greatest barriers to coverage and care. Conclusions: The bulk of research on coverage disparities for children of immigrants has focused on parental citizenship. Examining parental documentation status – an often masked distinction – provides insight into lack of insurance generally and ESI specifically, and reveals further disparities. The degree of insurance and ESI followed a strong gradient where children with undocumented parents experienced the most vulnerability and children with two citizen parents the least. The gap in ESI– which persisted after adjusting for several parental and family characteristics – appears to be the driving force behind these disparities. State-and local-level analyses provided a more complete picture of coverage disparities by parental documentation status.