Immigration Status and Use of Health Services among Latina Women in the San Francisco Bay Area

Introduction

The provision of healthcare for immigrants has been the source of controversy in California as well as the rest of the country. In the last several years, policies have tightened the documentation requirements for obtaining Medicaid coverage and restrict or forbid the use of public funds for undocumented individuals. ^1,2 The impact of immigration policy issues varies by state, and there is a big impact in California, where the 2005 U.S. Census estimated that >27% of the state's population was foreign born and 57% were noncitizens. ³ Latinos are the largest immigrant group in California, accounting for 45% of all new legal immigrants each year. ⁴ Over 75% of all immigrants in California originate from Mexico, with smaller groups from the Philippines, China, and India. ⁴

Immigration status has been associated with health insurance status ⁵ and access to healthcare. ^6,7 Reported barriers to care include written and spoken language, such legal issues as being undocumented, lack of familiarity in applying for and enrolling in public insurance, fear of payment for services, and long waiting times at healthcare settings. ^{8 –12} The impact of immigration status on health issues may vary by age, particularly within families. A recent study of California families (2001–2005) examined differences in health insurance coverage and access to care according to immigration status. The authors concluded that there were relatively few disparities and some improvement in access to care for children except for undocumented children, who represent a vulnerable group, but there was no concomitant improvement for parents. ¹³ For example, noncitizen parents had poorer access to care than citizens, and there was little change in access over time.

Gender is also an important determinant of health status and use of health services. Studies have shown that gender influences health status, social roles, culturally patterned behavior, and access to healthcare. ^8,14 Women may be eligible for special programs, particularly related to reproductive services, but they may not necessarily access these programs. Therefore, it is crucial to assess the sociodemographic characteristics and health-related needs of immigrant women to understand the challenges they face in obtaining services for maintaining and improving their health and quality of life.

Our knowledge of the relationship among immigration status, access to care, and use of health services in Latinos is limited and evolving, ^15,16 especially for immigrant Latinas. ^8,13,14,17 In particular, the issues that affect subpopulations of immigrants, such as undocumented women, have not been adequately studied. It is difficult to assess the impact of immigration status because it is challenging to obtain information about immigration status. To better understand the healthcare experiences of immigrant Latina women, we undertook a study of Latina women in the San Francisco Bay Area. The objective of our study was to assess the relationship between immigration status and use of health services among Latina women. Our hypothesis was that immigrant Latina women, particularly women with undocumented status, are less likely to use health services than Latina women who are citizens.

Materials and Methods

In 1999–2001, as part of a larger multistate study of welfare reform, ¹⁸ experienced and trained interviewers recruited 1090 Latina women who had just delivered a liveborn infant at three San Francisco Bay Area hospitals: San Francisco General Hospital, Seton Hospital, and St. Luke's Hospital. San Francisco General Hospital is a public hospital in San Francisco County, and Seton Hospital and St. Luke's Hospital are private hospitals in San Mateo and San Francisco counties, respectively. Eligible participants were at least 17 years of age and self-identified as Latin American or Latin Caribbean, with at least one parent or grandparent of Latin American or Caribbean national origin. ¹⁸ This was a consecutive sample where all eligible participants during the 1999–2001 study period were approached for enrollment into the study.

From 2001 to 2004, we recontacted the women and conducted 45-minute, structured interviews in Spanish or English to measure access to care and use of health services at 2 years and 3 years postpartum. The interview contained questions about demographics, health status, and immigration status. For women who completed one follow-up interview, we included all data from that interview. For women who completed both follow-up interviews, we included the data from the 3-year interview because it was the most recent interview. The interviews were conducted in person at San Francisco General Hospital or by telephone, according to each woman's preference. A total of 711 women completed a follow-up interview, and 710 women provided information on immigration status. The Committee on Human Research at the University of California, San Francisco approved this project, and written informed consent was obtained from each woman.

The primary dependent variable was use of preventive and nonpreventive health services. We analyzed two types of preventive health services—preventive health and dental care—and two types of non-preventive services—urgent care and emergency care. Preventive healthcare was defined as a physical examination or checkup, dental care was defined as going to a dentist's office or clinic for care, urgent care was defined as going to an urgent care clinic to receive services, and emergency care was defined as going to an emergency department for care. For each type of care, we asked each woman to report how many times she had used the service during the previous 12 months; the responses were dichotomized as none vs. one or more visits.

The primary independent variable was immigration status, which we obtained by self-report during the study interview ¹⁸ and categorized as undocumented immigrant, documented immigrant, or citizen (referent). During the latter part of the interview, after the woman had developed familiarity with the interviewer and the interview process, we asked a series of questions about immigration status. We began by asking if the woman was a U.S. citizen (U.S. born or naturalized) and ended the series of questions when the woman gave an affirmative response. If the woman was not a citizen, the next question asked if she was a legal resident or had a green card, and the final question asked was if she was undocumented. We aggregated US-born women and naturalized citizens into a single citizen group. ¹⁶ The group of documented immigrants included women with a green card or legal visa, and undocumented women were immigrants who did not have a green card or any other type of legal authorization for residence in the United States.

We assessed four demographic characteristics during the interview. Birthplace was reported by each woman and categorized as Latin America or the Caribbean, Mexico, or United States. Age was categorized as <30, 30–39 (referent), or ≥40 years. Educational attainment was obtained by self-report and categorized as <7, 7–11, or ≥12 years (referent). Marital status was dichotomized as married (married or single and living with partner, referent) or not married (single, divorced, or other).

Each woman reported her health status on a 5-point scale that ranged from excellent to poor; we created two categories for health status: (1) excellent, very good, and good or (2) fair and poor. We analyzed self-reported insurance status at the time of the interview as a measure of access to care and created three categories: uninsured, privately insured (referent), or publicly insured (Medicaid, emergency Medicaid, other public). To assess continuity of care, we asked each woman if she had a primary healthcare provider, which we defined as a personal doctor or nurse, and we categorized the response as a dichotomous variable.

We also analyzed the language spoken during the interview, categorized as Spanish only or Spanish and English vs. English only (referent). Self-reported annual household income was categorized as <$12,001, $12,001–24,000, $24,001–36,000, >$36,001 (referent), and unsure/declined.

We used contingency table analyses to compare selected characteristics across immigration status groups. To assess whether immigration status affected use of health services, we performed multivariable logistic regression analyses to compute odds ratios (OR) and 95% confidence intervals (95% CI). We included immigration status, age, education, marital status, health status, health insurance status, interview language, and annual household income as independent variables in the regression models. We omitted birthplace and primary provider from the multivariable models because these characteristics were strongly associated with immigration status and insurance status, respectively. We also omitted household income from the use of urgent care model because inclusion of income caused a lack of convergence of the model. The sample size for each analysis may vary because of missing data for some study variables. The data analyses were conducted with SAS/STAT software version 9.1 (SAS Institute, Cary, NC).

Results

In the follow-up study, half of all women were undocumented immigrants, one quarter were documented immigrants, and 18% were citizens (Table 1). The distribution of demographic and health status factors varied according to immigration status, and undocumented women were younger, were less educated, had lower income, were most likely to be uninsured, and were least likely to have a personal doctor or nurse. Use of preventive health and dental services during the preceding 12 months varied significantly by immigration status; undocumented immigrants had the lowest rates, followed by documented immigrants. Use of urgent and emergency care did not differ according to immigration status.

To assess the possibility of retention bias, we compared the distribution of selected demographic and reproductive characteristics between women who were retained in and lost to the follow-up study. There was no significant difference (p > 0.05) in the distribution of age, education, birthplace, employment status, marital status, immigration status, or the number of pregnancies when we compared women who were retained in or lost to the study. However, women who remained in the study were less likely to be U.S. born than women who were lost to follow-up (9.6% vs. 17%, p = 0.0006).

After adjusting for potentially confounding characteristics in multivariable logistic regression analyses, we found that immigration status remained independently associated with use of dental services (Table 2). Undocumented and documented immigrants were less likely to have used dental care during the previous year compared with citizens.

Unmarried women were less likely to use preventive health or dental care than married women (Table 2). Women who reported fair or poor health status were less likely to use dental care and more than three times as likely to use urgent and emergency care during the preceding year than women who reported excellent, very good, or good health status.

Health insurance status was independently associated with all four health service outcomes (Table 2). Uninsured women were less likely to use preventive health, dental, or urgent care services compared with privately insured women. In addition, publicly insured women were less likely to use dental care and more likely to use emergency care than privately insured women. Women who did not report a household income were less likely to use preventive healthcare compared to women in the highest income category.

Discussion

In this study of Latina women in the San Francisco Bay Area, more than half of all women were undocumented immigrants, and one quarter were documented immigrants. Although immigration status may be misreported, ¹⁹ the proportion of undocumented Latinas in our study was almost twice as high as the national average (29%). ²⁰ The undocumented women in our study had a unique profile of sociodemographic characteristics relative to documented immigrants and U.S. citizens; more than half were uninsured, and 62% did not have a primary health provider. As hypothesized, and after adjusting for confounding variables, undocumented Latina women in our study were less likely to use dental services than Latina citizens. The findings of our study are consistent with two recent studies, one study of undocumented Mexican-origin adults in New York City and another study of immigrant parents in California, both of which reported low insurance rates and low rates of primary health providers/regular source of care compared with U.S. citizen groups. ^13,21 Although recent political discussions have been contentious and might suggest that Latino immigrants overuse health services, ¹⁷ our findings highlight the vulnerability of undocumented Latina women and underscore the need for programs and policies to address their health needs. Further research is needed to assess the health needs of immigrant Latina women, but policymakers must address the problems of uninsurance and poor use of health services among undocumented Latina women and other vulnerable populations of immigrants.

Latinos have the lowest insurance rates of all racial/ethnic groups in the United States, and lack of insurance is a major issue for immigrants. ²² As expected, we found low rates of insurance among the Latina women in our study, and the proportion of uninsurance varied by immigration status, from a low of 15% among citizens to a high of 52% among undocumented women. The insurance rates in our study were similar to the rates reported among Latina women in the 1982–1984 Hispanic Health and Nutrition Examination Survey, ¹⁷ as well as a multiethnic cohort of childbearing women in Boston ²³ and parents in the California Health Interview Survey of families with children, ¹³ which suggests that insurance status for Latina women has not improved over a 20-year period. In our study, health insurance coverage was particularly important for use of dental care but less so for preventive healthcare, which suggests that women may have used safety net providers for health services but that there is not the same type of dental safety net. From a policy perspective, it is important to address both preventive health and dental needs and implement programs and policies that encourage adults to seek preventive care. In addition, further study is needed to compare the factors associated with use of health and dental services among immigrant Latinas and develop interventions to improve use of dental care.

Preventive healthcare is a cornerstone of public health, and 37% of the Latina women in our study had not received any preventive health visits during the previous year. We hypothesized that undocumented women would have a lower number of preventive health visits than U.S. citizens, but we failed to find a significant difference after adjusting for potential confounders. Our results are consistent with the results of a multiethnic cohort of childbearing women in Boston, where there was no significant difference in use of prenatal care among citizens, documented immigrants, and undocumented immigrants. ²³ On the other hand, previous studies of adults have reported that immigrants were less likely to use preventive care than citizens. ^15,24 Studies of women have also reported that immigrants were less likely than citizens to receive cancer screening tests. ^25,26 Most previous studies, however, have analyzed multiethnic populations, ^15,25 aggregated men and women, ¹⁵ or combined documented and undocumented immigrants, ^15,25,26 which makes it difficult to directly compare our findings with previous studies. Based on our findings, we support the development of uniform, policy-based solutions to address the preventive health and dental needs of Latina women across the life span. Although current policies emphasize reproductive health, which is an important component of women's health, our results underscore the need for policies based on a comprehensive, continuous approach to improving women's health.

Our study findings are subject to several limitations. First, the community of Latinos in the San Francisco Bay Area may not be representative of Latinos in other parts the United States. The Latina women in our study were of Mexican and Latin American origins, so our findings may not be applicable to Latinas of Cuban and Puerto Rican origins, particularly because Cubans are subject to special immigration status guidelines and Puerto Ricans are U.S. citizens. Second, we asked women to report their immigration status, which may be subject to a reporting bias. Two thirds of the women reported that they were undocumented at the time of recruitment into the study, ¹⁸ which is consistent with the rate in the follow-up study (56%). However, the immigration status of Latinos in the San Francisco Bay Area may differ from Latinos in other parts of the country; in Forth Worth, Texas, 41% of adult Latinas were undocumented. ¹⁴ Finally, we do not know if women accurately reported their use of services during the previous year, but we collapsed the number of visits as none or one or more as a means of minimizing reporting error.

Whereas many studies have compared immigrants to nonimmigrants, we directly assessed immigration status in a large cohort of Latina women in the San Francisco Bay Area and found that immigrant Latinas, particularly undocumented women, have problems with use of dental services. Clinical and policy efforts to address the health needs of Latina women must acknowledge the unique risks associated with immigration status, particularly undocumented immigrants.