Recent legislation mandating reporting of undocumented immigrants may instill fear of discovery when they access emergency department (ED) services. The objectives of this study were to: 1) characterize the knowledge and beliefs of undocumented Latino immigrants (UDLI) about health care workers' reporting (or nonreporting) of illegal immigrants in the ED, 2) determine whether UDLI fear discovery when presenting to the ED, and 3) determine the nature and sources of this fear.
This was a cross-sectional study of UDLI and two comparison groups conducted in two California county EDs, from November 2009 to August 2010. The authors interviewed a convenience sample of adult UDLI, Latino legal residents (LLR), and non-Latino legal residents (NLLR) using a structured instrument in their native language. The main outcome was fear of discovery among UDLI and the sources of that fear.
Of 1,224 patients approached, 1,007 (82.3%) were interviewed: 314 UDLI, 373 LLR, and 320 NLLR. The median age was 43 years (interquartile range [IQR] = 31 to 55 years), and 51% were male. UDLIs were less likely to speak English (14%, 95% confidence interval [CI] = 10% to 18%), have health insurance (39%, 95% CI 32% to 44%), or have a regular primary care provider (PCP; 39%, 95% CI = 34% to 45%), compared to LLR (English 56%, 95% CI = 51% to 61%; health insurance 50%, 95% CI = 45% to 55%; regular PCP 51%, 95% = CI 46% to 57%) and NLLR (English 95%, 95% CI = 92% to 97%; health insurance 49%, 95% CI = 43% to 54%; regular PCP 51%, 95% CI = 45% to 56%). Of the 16% of UDLI who stated that nurses and doctors treat undocumented immigrant patients differently than citizens, 41% (95% CI = 29% to 54%) reported less respect given to UDLI by staff. Thirty-two percent of UDLI had heard of Proposition 187, 13% believed hospital staff reported UDLI to immigration authorities, and 9% said they were asked about their citizenship status. Fear of coming to the hospital because of discovery was expressed by 12% (95% CI = 9% to 16%) of UDLI, with 42% (95% CI = 28% to 58%) citing discussions with friends/family, 32% (95% CI = 19% to 47%) citing media and 16% (95% CI = 7% to 30%) citing both as sources of this fear.
One in eight of UDLI presenting to the ED express fear of discovery and consequent deportation. Belief that medical staff report UDLI and recent immigration are risk factors for this fear. Family, friends, and media are the primary sources of these concerns.
El Miedo a Ser Descubierto entre los Inmigrantes Latinos que Acuden al Servicio de Urgencias
La obligación según la legislación actual de denunciar a los inmigrantes indocumentados puede infundirles miedo a ser descubiertos cuando acuden al servicio de urgencias (SU). El objetivo fue caracterizar: 1) el miedo a ser descubierto entre los inmigrantes latinos indocumentados (ILI) que acude al SU; 2) qué pacientes ILI tienen miedo a lo que ocurrirá; y 3) las fuentes de este miedo.
Éste fue un estudio transversal en ILI, y comparativo con otros dos grupos llevado a cabo en dos SU de California desde noviembre de 2009 a agosto de 2010. Se entrevistó a una muestra de conveniencia de adultos ILI, residentes latinos legales (RLL) y residentes no latinos legales (RNLL), mediante un instrumento estructurado en su lengua nativa. El resultado principal fue el miedo a ser descubierto entre los ILI y las fuentes de ese miedo.
De los 1.224 pacientes propuestos, se entrevistó a 1007 (82,3%): 314 ILI, 373 RLL y 320 RNLL. La mediana de edad fue de 43 años (RIC 31–55 años) y el 51% fue de sexo masculino. Los ILI tenían menor probabilidad de hablar inglés (14%, IC 95% = 10% a 18%), de tener seguro sanitario (39%, IC 95% = 32% a 44%) o de tener un médico de atención primaria (MAP) permanente (39%, IC 95% = 34% a 45%) comparado con los RLL (inglés 56%, IC 95% = 51% a 61%; seguro sanitario 50%, IC 95% = 45% a 55%; MAP permanente 51%, IC 95% = 46% a 57%), y los RNLL (inglés 95%, IC 95% = 92% a 97%; seguro sanitario 49%, IC 95% = 43% a 54%; MAP permanente 51%, IC 95% = 45% a 56%). Del 16% de ILI que indicaron que los enfermeros y los médicos trataron a los pacientes inmigrantes indocumentados de forma distinta que a los ciudadanos, el 41% (IC95% = 29% a 54%) señaló un menor respeto hacia los ILI por parte del personal. Un 32% de ILI ha oído hablar de la Proposición 187; un 13% pensó que el personal de hospital informó de su condición de ILI a las autoridades de inmigración; y el 9% dijo que había sido preguntado por su ciudadanía. El miedo a venir al hospital a causa de ser descubierto se expresó por el 12% (95% IC = 9% a 16%) de los ILI, el 42% (IC 95% = 28% a 58%) mencionó conversaciones con familiares/amigos, el 32% (IC95% = 19% a 47%) citó los medios y el 16% (IC 95% = 7% a 30%) mencionó ambos como fuentes de su miedo.
Uno de cada ocho de la condición de ILI que acudieron al SU expresó miedo a ser descubierto y a la consecuente deportación. La creencia que el personal médico informe de ILI y la inmigración reciente son factores de riesgo para este miedo. La familia, los amigos y los medios son las fuentes principales de estas preocupaciones.
Undocumented Latino immigrants (UDLI) face considerable challenges to receiving health care in the United States. Beyond well-described economic and language barriers, a general lack of familiarity with the system and its policies may make it difficult for them to access medical care for even the most basic problems. As it does for other similarly disenfranchised groups, the emergency department (ED) serves as the principal health care safety net for UDLI; with simply no other place to go, they turn to the ED as their primary access point for most of their health care issues.[1, 2]
California's passage of the controversial Proposition 187 and the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) in the mid-1990s, as well as more recent highly publicized national and state legislation in Arizona and Alabama, may have introduced another major deterrent to immigrants' accessing health care in America, that being the fear of discovery and deportation when presenting to the ED.[3-10] After Proposition 187 passed, community health care providers in California reported declines in visits from UDLI, despite the proposition never having taken effect.[6-8, 11-14] Several studies have demonstrated that UDLI use EDs and other ambulatory health care services at significantly lower rates than non-Latinos and U.S.-born Latinos.[1, 4, 15-19] This lower rate of utilization may arise from fear of accessing these services.
Other investigators have examined undocumented immigrants' fear of accessing medical clinics using telephone surveys and neighborhood interviews. While these population-based studies may allow for a more accurate assessment of fear of discovery in the general UDLI population, it is unlikely that such sampling would capture UDLI fear in the setting of decisions to access emergency care. Fear of discovery must be evaluated in real time in a true emergency care setting to accurately capture the feelings of those who are deciding whether or not to access emergency care. Because those with the greatest fear are the least likely to access the ED, the full effect of fear of discovery on UDLI decisions to present for emergency care may be impossible to gauge—this population may remain hidden and avoid emergency care despite serious illness. However, a lower boundary of fear of discovery affecting UDLI emergency care access decisions may be generated by interviewing those who do present to the ED. To establish this lower boundary of fear during emergency care access, the objectives of this prospective study conducted at two county California hospitals were to 1) characterize UDLI knowledge and beliefs about health care workers' reporting (or nonreporting) of illegal immigrants in the ED, 2) determine whether UDLI fear discovery when presenting to the ED, and 3) determine the nature and sources of this fear (if present).
Study Design and Population
This study was approved by the University of California San Francisco (UCSF) Committee on Human Research, as well as the Olive View–UCLA Education & Research Institute. Both institutional review boards approved verbal consent for this study.
We conducted this cross-sectional study in the EDs of two California county hospitals, San Francisco General Hospital (SFGH) in San Francisco County and UCLA Olive View Medical Center (OVMC) in Los Angeles County, which had 57,770 and 60,103 patient visits in 2009, respectively. At SFGH, 22.9% of ED visits were by patients of self-declared Latino origin; OVMC does not record ethnicity information. Between 10 a.m. and 10 p.m. from November 2009 to August 2010, we surveyed a convenience sample of adult (>18 years of age) patients with the following exclusions: 1) trauma patients; 2) patients transferred from another hospital, chronic care facility, nursing home, or hospice; 3) patients unable to participate in an interview because of intoxication, altered mental status, or critical illness; 4) incarcerated patients; 5) patients under psychiatric holds; and 6) Latino patients on visas vacationing in the United States.
We designed our protocol to concomitantly enroll nearly equal numbers of UDLI and two comparison groups: Latino legal residents (LLR) and non-Latino legal residents (NLLR). We initially screened potential subjects by reviewing their triage information (surname, chief complaint, and mode of transport). We verbally consented subjects and interviewed them in private rooms or semiprivate (screen partitioned) gurneys during their waiting times in the ED. We categorized participants into the UDLI, LLR, and NLLR groups through direct interview by their responses to the questions, “What is your primary language?” and “Are you a legal resident of the United States?” Primary language and surnames were considered proxies for ethnicity; however, any uncertainty was clarified by asking the participants if they were of Latino background. We also asked whether participants had a social security number. As a check on their declared residency status, we reviewed participants' registration information, including social security numbers. Whenever we determined that the number of subjects in one of the three groups was 20 fewer than the other groups, we used triage screening information and interviews to enroll more subjects into that group.
Survey Content and Administration
After review of prior work in this field,[1-5, 15, 16, 19] we constructed an instrument consisting of yes/no, multiple choice, free response, and numerical analog questions. After review for content validity by two faculty experts in health care literacy who were not otherwise involved in this study, we pilot-tested the instrument on eight UDLI to confirm test–retest consistency (survey available in English as Data Supplement S1 and in Spanish as Data Supplement S2, available as supporting information in the online version of this paper).
Using a structured instrument, we interviewed each participant in his or her native language and provided each person approached with an information sheet on consent in his or her native language. Research assistants (medical students and postbaccalaureate students) received a 2-hour orientation session and shadowed the principal investigators during interviews to ensure standard survey technique prior to conducting interviews themselves. At the conclusion of each survey, the interviewer assured the subject that EDs maintain complete confidentiality regarding citizenship or noncitizenship and do not report to immigration agencies.
All data were entered into Microsoft Excel (Microsoft Corp., Redmond, WA) using double data entry checking. Statistical tests were performed using Excel and STATA v 9.0 (StataCorp, College Station, TX). Demographic data were summarized and reported in aggregate form. Responses to survey questions were tabulated and reported as frequency percent. The mean plus standard deviation (SD) of analog scale questions were calculated. Mean difference in proportions with 95% confidence intervals (CI) was used to determine statistical significance between comparison groups. Potential associations between predictors and the outcome were first assessed using univariate logistic regression models. We checked 16 variables for univariate association. All variables significant at the p = 0.05 level were then included in a multivariable model.
Of the 1,224 patients approached regarding enrollment during the study period, 217 (17.7%) refused and 1,007 (82.3%) agreed to participate, with 314 (31.2%) participants classified as UDLI, 373 (37%) classified as LLR, and 320 (31.8%) classified as NLLR. Once enrolled, none of the participants withdrew from the study. See Table 1 for a summary of participant characteristics.
Both study sites had similar responses for participants' primary language, English proficiency, housing status, belief that their problem was an emergency, UDLI belief that staff report their legal status, and satisfaction with care received. SFGH UDLI expressed fear of presenting to ED with a greater frequency (14.1% vs. 9.7%) and were less likely to have heard of Proposition 187 (26.5% vs. 39.6%) than OVMC UDLI, respectively.
Collectively, UDLI were less likely to speak English (14%, 95% CI = 10% to 18%), have health insurance (39%, 95% CI 32% to 44%), or have a regular clinic or doctor (39%, 95% CI = 34% to 45%) for their care when compared to LLR (56%, 95% CI = 51% to 61%; 50%, 95% CI = 45% to 55%; and 51%, 95% CI = 46% to 57%) and NLLR (95%, 95% CI = 92% to 97%; 49%, 95% CI = 43% to 54%; and 51%, 95% CI = 45% to 56%). Similar percentages of UDLI, LLR, and NLLR were homeless (6%, 95% CI = 4% to 10%; 6%, 95% CI = 4% to 8%; and 9%, 95% CI = 6% to 12%). On a 0 to 10 scale, UDLI, LLR, and NLLR reported similar levels of satisfaction with their care in the ED (9.0, 8.7, and 8.3, respectively).
Of the 16% of UDLI who stated that nurses and doctors treat undocumented immigrant patients differently than other patients (subset n = 51), 41% (95% CI = 29% to 54%) reported less respect or lower standard of care for UDLI, 37% (95% CI = 25% to 51%) reported that language barriers impede proper care and result in less attention for UDLI, and 16% (95% CI = 8% to 28%) reported longer waiting times for UDLI. Nine percent of all UDLI stated they were asked their citizenship status by hospital staff, and 13% believed that staff reported UDLI to immigration authorities (see Table 2).
Table 2. Participants' Perception of Hospital Reporting Practices and Services Rendered
Fear of coming to the hospital because of discovery was expressed by 12% (95% CI = 9% to 16%) of UDLI, with 42% (95% CI = 28% to 58%) of this subset (n = 38) citing discussions with friends/family, 32% (95% CI = 19% to 47%) citing media, and 16% (95% CI = 7% to 30%) citing both friends and media as the source of this fear. None of the UDLIs expressing fear stated that they had a social security number. For the UDLI with fear of presenting to the ED (n = 38), the primary concerns were being reported and deported, 71% (95% CI = 55% to 83%) and not receiving medical care, 16% (95% CI = 7% to 30%). Thirty-two percent (95% CI = 28% to 38%) of all UDLI had heard of Proposition 187 (Figure 1). On multivariable logistic regression, the most significant predictors of fear of accessing the ED were the belief that staff report UDLI and having lived in the United States < 5 years, with odds ratios (OR) of 14.4 (95% CI = 6.2 to 25.1) and 2.6 (95% CI = 1.0 to 6.5), respectively (Table 3 and 4). Of those UDLIs who believed staff report patients' legal status (n = 41), 52.6% were afraid to come to the ED, compared to 7.8% of those who did not believe this (n = 229).
Table 3. Univariate Logistic Regression of Predictors of Fear Among UDLI
SSN = social security number; UCSF = University of California, San Francisco; UDLI = undocumented Latino immigrants.
No subjects who reported being afraid had a SSN; p-value calculated with Fisher's exact test.
Table 4. Multivariable Logistic Regression of Predictors of Fear Among UDLI
UDLI = undocumented Latino immigrants.
Five or more years in United States
Been asked if a citizen? n (%)
Believe staff reports UDLI? n (%)
Although a population-based survey of this hidden population of UDLI may characterize overall background fear of health care access, determining how much fear exists in the real-time ED access of UDLI is essentially impossible, as those who fear the most will not present and therefore cannot be assessed. Nevertheless, our study of UDLI who did present to the ED provides an estimate of the lower boundary of this fear prevalence. We also characterized from where this fear arises, providing a starting point in measures to address and curtail the troubling effects of this fear.
As of March 2010, approximately 9 million UDLI reside in the United States, a number that has tripled since 1990. Although California has the greatest UDLI population, UDLI have spread broadly throughout the country in the past 20 years. Conducted in real time at the primary health care access point for most UDLI, our study demonstrates that one in eight UDLI endorse fear of discovery, and one in six believe that they receive different treatment (primarily less respect and a lower standard of care) when presenting to the ED. The most significant predictors of fear of accessing the ED were the belief that staff report UDLI and having lived fewer than 5 years in the United States. Beyond the other formidable barriers to care they face, these concerns pose a serious threat to the ED's role as the primary medical care safety net for this vulnerable population. Beyond pure medical care, the ED provides emergency subsistence care (food, shelter, and safety) to marginally housed and homeless individuals. Given that we found similar homelessness percentages for UDLI, LLR, and NLLR, fear may also compromise the ED's subsistence safety net role for this group.
The few previous studies examining the effect of fear of discovery on immigrant primary health care access[3-5] have been conducted through telephone surveys and neighborhood interviews. Reviewing data from a Project Hope survey conducted in Latino neighborhoods in Houston, El Paso, Fresno, and Los Angeles, Berk and Schur3 found that 39% of undocumented immigrants stated that they were afraid they might not receive medical care because of their undocumented status. In addition to fear of discovery and deportation, many of the UDLI in our study endorsed a similar fear that they would not receive care.
Fear may induce delays in ED presentation that may not only lead to poor outcomes in acute disease, such as acute myocardial infarction and stroke, but also may generate greater health care costs and resource use, when patients present in later (higher acuity) stages of disease. Other factors, including patient perception of illness severity, affect decisions to seek emergency care—a higher percentage of UDLI believed that their problem was an emergency.
With regard to fear-induced effects on public health safety-net function, several investigators have described drops in health care visits by immigrants, and public health threats due to immigrants' fear and subsequent delay in seeking care, after the passage of Proposition 187.[11-14] Asch et al. reported that undocumented immigrants reporting fear of discovery were four times as likely to delay getting care for tuberculosis symptoms as those not having such fear. With the ED's emergence as the leading site of new HIV diagnosis[23, 24] and its provision of preventive health care measures, including pneumococcal and influenza immunizations and sexually transmitted disease screening,[25-28] many ED-derived public health measures may be compromised as well.
Given these threats to public health, policy-makers should no longer narrowly consider the care of UDLI (and other immigrants) as merely an immigration issue, and programs to alleviate this fear should be considered. Simple measures within EDs, such as our provision to study subjects of information stating that EDs do not report undocumented immigrants, may be enacted on a broad basis. Other ED- or clinic-based interventions might include the addition of statements at triage and on discharge instructions, confirming hospital nonreporting practices.
Beyond messages at EDs and clinics, community partnerships with local nonprofit organizations or resource centers known to be “safe” or common access points for UDLI may spread the ED safety net message more broadly. Informal quarterly forums, supplemented by educational pamphlets, may foster dialogue between community members and health care providers. Given that many of our subjects reported media sites as the source of their fears, radio and television public awareness campaigns, targeted at communities with large immigrant populations, may also be useful. In a subsequent study, we plan to explore whether some of the above-mentioned interventions aimed at decreasing fear improve access to the ED and whether UDLIs' fear of reporting results in delayed presentations and worse outcomes.
Our results are subject to the well-described limitations inherent in convenience sampling methods, as well as the limitations of interview-based studies, particularly social desirability bias and failure of participants to respond to all questions. We did not elucidate the reasons for the few incomplete survey responses. Although we confirmed social security numbers (or lack thereof) on registration forms, we could not truly verify participants' self-reports of documented status, and some respondents may have been reluctant to admit that they were undocumented—this may have affected the prevalence of fear we noted. However, participants appeared to candidly respond during interviews; some UDLI admitted being undocumented despite having given a social security number at registration.
We conducted this study in two California “sanctuary” cities, where municipal resources are not used to enforce federal immigration laws. We may therefore have underestimated the percentage of UDLI who fear deportation, and there may be even greater concern in nonsanctuary cities. During the design phase of this study, we attempted to include an ED site in Arizona, but the investigator was denied IRB approval—a rare occurrence itself that serves as a testament to the politically and legally charged nature of this topic. We chose not to ask participants' country of origin. It is possible, even likely, that important heterogeneity of attitudes and fears exists among the UDLI lumped into a single group; e.g., Mexican immigrants may feel more or less fear than Honduran immigrants.
Perhaps the most important limitation is the difficulty posed by social desirability and the fears generated when merely trying to bring up this issue in a vulnerable population. Despite techniques to minimize these effects, participants may have given responses that they thought we wanted to hear, or that would not expose them to risk, leading us to underestimate their fear. Given that most UDLI were familiar with ED nonreporting practice, they may have even recognized an ED advocacy role. However, of those UDLI with fear, over two-thirds believed that staff reported UDLI or were unsure of reporting practices. Other sex-based, cultural attitudes also may have affected participants' responses: Latino men, who traditionally value male strength and security, may have been more reluctant than women to admit to a perceived weakness of fear.
One in eight undocumented Latino immigrants presenting to the ED in our sample expressed fear of discovery and consequent deportation. This figure represents a lower limit of fear prevalence; we expect this proportion to be substantially greater among undocumented Latino immigrants who need emergency services but do not present for care. Belief that medical staff report undocumented Latino immigrants and recent immigration are risk factors for this fear, and family, friends, and media are the primary sources of these concerns. When state and national policy-makers draft legislation for health care and immigration reform, they should consider the resultant fear on this vulnerable population and its impact on the ED public health safety net.
The authors thank colleagues at UCLA–Olive View for their collaboration and implementation of the study at their facility.