Language as a Barrier to Access to Healthcare among Vietnamese Immigrants

Abstract: 

This study used the Behavioral Model for Vulnerable Populations to examine language as a barrier to accessing healthcare among Vietnamese immigrants. Thirty-six Vietnamese patients in a tax-supported, safety-net healthcare system in a large urban county in north Texas were interviewed in the fall of 2000. Non-English speakers were significantly less likely to be U.S. citizens or applicants for citizenship, or to work at a paying job, but more likely to need an interpreter during visits to the doctor, to need someone to go with them to the doctor, and to have problems with transportation to doctor visits. Non-English speakers were less healthy than the English speakers, being more likely to be in fair or poor health, limited in the amount of work they can do, anxious, and less likely to feel healthy and energetic. Non-English speakers were also less likely to have a regular source of healthcare. Recommendations for improvement of services included providing more translators, developing health education materials in Vietnamese, and working with community organizations to provide health education in a relevant cultural context.

Table of Contents: 

    Introduction

    The purpose of this study is to examine the disparity in access to healthcare between Vietnamese patients who speak English and those who do not, and to examine and compare predictors of use of healthcare services among Vietnamese patients who speak English and those who do not. Access to healthcare is defined in this study as the use of preventive and primary care. The findings from this study may have implications for larger subgroups of the Asian population.

    Five percent of the total U.S. population consists of Asian Americans. Asian Americans include many groups, the Vietnamese being one of the larger subgroups, along with Koreans, Chinese, and Japanese. There was a large immigration boom following the repeals of the laws in the 1980s that made race a barrier to legal immigration and citizenship, which in turn led to a larger population of first- and second-generation Asian immigrants (U.S. Census Bureau, 2000).

    The Behavioral Model for Vulnerable Populations, developed by Gelberg, Andersen, and Leake (2000), has been adapted for the purpose of understanding the effects of language skills on health services utilization as well as on the predisposing, enabling, and health status variables that have been found to facilitate or hinder access to healthcare. The model, presented in Figure 1, looks at predisposing variables (e.g., race/ethnicity, gender, immigrant status, and language skills); enabling variables (e.g., insurance status, income, transportation, and competing needs for other necessities of life); and health status variables as predictors of use of healthcare services among low-income, uninsured, and vulnerable populations.

    Literature Review

    The American healthcare experience varies greatly from patient to patient. The Behavioral Model for Vulnerable Populations can be used to describe the Vietnamese immigrant population’s access to healthcare. The model that has been adapted includes predisposing, enabling, and need variables that have been found to be predictive of access to healthcare services. Predisposing variables are social and demographic characteristics of persons that exist prior to their need for healthcare services that may influence their health and therefore their need for healthcare. Predisposing variables may also influence their access to enabling resources that allow persons to use the healthcare system when they become ill and need healthcare. These variables are divided into two domains, the traditional and the vulnerable domains. The traditional domain refers to variables that apply to all subgroups in the general population that could affect access to healthcare. Examples in the traditional domain include age and sex among the predisposing variables; health insurance among the enabling variables; and general health status among the need variables. The vulnerable domain includes variables that are especially relevant to low-income populations. Examples include variables such as immigration status among the predisposing variables, needing an interpreter among the enabling variables, and feeling worried or anxious among the need variables.

    Health-Seeking Behaviors of the Vietnamese Population

    There are “significant disparities in access [to healthcare] based on cultural heritage,” according to Andrulis (2003, p. 791). Therefore, before analyzing language barriers among the patients that were under study, it would be beneficial to look at the normative values and customs, but more specifically, the health-seeking behaviors of the Vietnamese population. Vietnamese people exhibit many of those general behaviors that are exhibited by Asian Americans. An observation of health behavior based on prior reading is that Vietnamese patients prefer drug solutions to their ailments rather than lifestyle changes.

    Previous research suggests that minority populations in general are “less likely to receive the same quality of care as those from dominant cultures” (Andrulis, 2003, p. 792). Asian Americans in general are less likely to visit providers for “emerging health problems, referrals, and preventive services compared to whites” (Andrulis, 2003, p. 792). Vietnamese people also put high importance on family and family decision-making in healthcare. These behaviors include scheduling appointments so that both the patient and the patient’s family member who speaks English and can translate for the patient are available to come to the appointment, which increases adherence to scheduled appointments.

    “[Asians who] are amongst the highest growing minority groups in the U.S. and who report relatively high levels of dissatisfaction with health care” (Green et al., 2005, p. 1050). Attitude toward healthcare shapes patients and their medical decisions based on what the patients perceive to be the value of information. Healthcare providers and their ability to satisfy the needs of their patients influence attitudes. Trusting a medical practitioner is crucial but language barriers, intercultural communication, and cross-cultural misunderstandings between patient and doctor can cloud the perception of healthcare and the healthcare experience.

    Use of Language-Enhanced Resources and Interpreters in Hospitals

    A common belief among healthcare experts is that use of interpreters can compromise communication between the patient and the clinician. However, “the perceived quality of the interpreter is strongly associated with patient’s perceived care overall” (Green et al., 2005, p. 1050). This implies that ad hoc interpreters, such as family members or other patients from a doctor’s waiting room, are the least effective translators because they may compromise the patient’s perceived confidentiality and inhibit discussion of sensitive issues, thus lowering the overall quality of the healthcare experience. Use of ad hoc interpreters is not professional, and should be avoided.

    Issues related to use of interpreters with immigrants include confidentiality, problems understanding instructions for use of prescription medicines, physical misunderstandings, nonverbal connections, and cultural or emotional misunderstanding.

    While office visits with professional interpreters do not necessarily take longer, interpretation can impose time pressures…. Use of interpreters may also compromise rapport between patients and clinicians, and their presence may inhibit patients’ questions, particularly about sensitive topics such as mental health (Green et al., 2005, pp. 1054-1055).

    With the vulnerability of the patient ’s personal status, inadequate healthcare service may lead to decreased use of healthcare services.

    Language concordance occurs when doctors and patients are fluent in the same language. Language discordance occurs when the doctor and the patient are not able to speak to one another fluently in one language. Research suggests that a language concordant doctor is better for lower income and less educated patients because it makes the patient feel the doctor is promising confidentiality (Green et al., 2005). Having a language concordant doctor also cuts down on cultural, ethnic, and nonverbal aspects of communication that can create barriers. A language concordant doctor may not be available in many clinics or a hospital for the over 300 languages spoken in the United States. Within a region that has a large, densely settled population of immigrants, it has been suggested that interpreters should be available on call for these populations (Green et al., 2005).

    Implications of the Health Belief Model: Language as a Perceived Barrier

    Barriers are a two-way street; they are as difficult for the doctors as they are for the patients. The resources available to multi-ethnic groups accessing healthcare range anywhere from no resources, to written materials in the patient’s language of origin, to a translator, to a fluent translator trained in medical translation, up to a doctor who speaks the language of the patient. Language is only one component of the many perceived barriers but others such as transportation, time, location, and money play a compounding role. This can lead to frustration, which can keep cross-cultural patients from following through with cues to action (Poss, 2001).

    Perceptions of the quality of Asian American patients’ communications with their clinicians during their most recent visit were studied by Green et al. (2005). Patient concerns included having “…enough time to explain the reason for their visit, providing understandable information, questions about their care that they wanted to ask but did not, questions about their mental health they wanted to ask but did not, or receiving information about their health and treatment as they wanted” (p. 1054).

    Many of those questions pose problems that can be overcome by recognizing and identifying with the importance of ethnic barriers and being more accessible and patient centered.

    Hypotheses

    Four major hypotheses guided this research. They are as follows:

    1. Hypothesis 1. Non-English-speaking Vietnamese immigrants will be more predisposed to use healthcare services than English-speaking Vietnamese immigrants.
    2. Hypothesis 2. Non-English-speaking Vietnamese immigrants will have fewer enabling resources with which to access healthcare services and experience more barriers to the use of healthcare services than will English-speaking Vietnamese immigrants.
    3. Hypothesis 3. Non-English-speaking Vietnamese immigrants will have greater need for healthcare services as measured by health than will English-speaking Vietnamese immigrants.
    4. Hypothesis 4. Non-English-speaking Vietnamese immigrants will use healthcare services less and will have more difficulty accessing the services than English-speaking Vietnamese immigrants.

    Research Methods

    Data for this study was taken from a larger survey of 2,034 patients who had at least one clinic visit in the JPS Health Network in Tarrant County, Texas, in July and August 2000. Patients were randomly selected from the sampling frame of almost 10,000 eligible patients. There is a small Vietnamese community in Tarrant County and some of the community members use the JPS safety-net, tax-supported, healthcare services network. When the study was conducted the network consisted of a full-service hospital and eight satellite health centers geographically distributed throughout the county. Because the Vietnamese population was such a small proportion of the total population, all the patients with names identified as Vietnamese or Asian were selected for interviews. In all, 36 Vietnamese immigrants were interviewed by translators in Vietnamese or English in a telephone survey conducted by the Survey Research Lab at the University of North Texas in Denton, Texas. Thirteen of the Vietnamese immigrants spoke English well enough to be interviewed in English, while 23 did not. Prior to being called in the survey, patients in the sample were mailed letters explaining the reason for the survey and informing the patients of their rights as survey participants, including the right to refuse to participate and to refuse to answer any question to which they objected. They were also informed that their answers to questions were confidential and that their identities would be protected in the study. When patients were phoned, they were again informed of the nature of the research and of their rights. They were given an opportunity to ask any questions and to refuse to participate in the study. The protocol for this research was approved by the Institutional Review Boards of the University of North Texas in Denton, the University of North Texas Health Science Center in Fort Worth, and the JPS Health Network.

    The data were analyzed using gamma as the measure of association and approximate T as the test of significance in the SPSS package. Although the sample sizes are small, many of the relationships were strong enough to achieve statistical significance while many others approached statistical significance (p > .05 and < .10). Both statistically significant relationships and those that approach significance will be discussed.

    Ethical and Safety Risks

    This project has no ethical or safety risk or any other foreseeable risks. The only issue that could come up is confidentiality and to handle this, the principal investigator of the primary study did not disclose any personally identifiable information on survey participants with the researcher of this study. Only data that were directly relevant to the study were shared. The data were kept on the faculty mentor’s office computer. Names of members of the sample and contact information were kept in a locked file cabinet in the Honors College office at the University of North Texas.

    Variables

    The variables used in this study include the four types of variables described in the model earlier. The predisposing variables included in this model are shown in Table 1 along with their frequencies. Age, sex, and marital status are predisposing variables in the traditional domain while citizenship status is in the vulnerable domain for this immigrant population. The Vietnamese population was fairly evenly distributed across the four age categories with the exception of the 30–39 year old category. Among the Vietnamese, 30.6% fell into the age categories 18–29 years, 40–49 years, and 50–60 years, but only 8.3% were in the 30–39 year old category. The sample was highly skewed toward females, with 83.3% being women. Most of the sample, 61.1%, was currently married and living with a spouse while 38.9% was not. For the vulnerable citizenship category, 35.5% of the sample respondents reported that they were U.S. citizens and 38.2% were applying for citizenship. About 20% reported that they were “not yet qualified” and 5.9% reported that they needed to establish permanent residency first. Roughly one-fourth of the Vietnamese were thus vulnerable in terms of their citizenship status in the United States because noncitizens did not generally have the same rights to welfare benefits as citizens, including healthcare benefits provided by tax-supported safety-net providers.

    The enabling variables used in the study are presented in Table 2. The traditional enabling variable used is the number of hours worked per week. The modal hours worked per week was none (36.1%). Most of the respondents who worked, worked part-time 21 to 37 hours (25.0%) or full-time (27.8%), although 11.1% reported working 45 or more hours a week. There were five variables in the vulnerable domain – needing an interpreter in a healthcare setting in the past year, needing someone to go with the patient to see the doctor, need for the accompanying person to take time off from work to go with the patient to the doctor, whether or not the patient gets sick leave from work, and having had problems with transportation for healthcare visits in the past year. One-third of the Vietnamese respondents reported needing an interpreter in a healthcare setting during the past year. Of the 47.2% of respondents who reported needing a person to accompany them to the doctor in the past year, 31.3% said that person needed to take time off from work to go with them. Only 9.1% of the respondents said they got paid time off when they were sick. Transportation to doctor visits is not much of a problem for this population as only 5.6% reported problems with transportation for doctor visits in the past year.

    The need variables measuring need for healthcare services in terms of health are presented in Table 3. The two traditional variables used were overall rating of health status and whether or not the person was limited in his or her ability to work because of physical or mental problems. The two vulnerable measures used were how often the respondent reported feeling full of energy and how often he or she felt worried, tense, or anxious. More than half (52.8%) of the respondents rated their health as only fair or poor, 27.8% rated their health as good, and only 19.4% thought their health was very good or excellent. Given how poorly they rated their health, surprisingly few reported being limited in their ability to work – only 11.4%. Almost two-thirds (61.8%) reported that they rarely felt full of energy. One in five (20.6%) felt worried, tense, or anxious most days or every day.

    Table 4 contains the frequencies and percentages for the variables that measure healthcare service utilization. There were three traditional measures – having had a usual source of care, having had a personal doctor and, for the women, having had a Pap smear at some point in their lives. The vulnerable measures that were especially relevant for our immigrant sample included having had an emergency department visit in the past year, having had a problem getting needed healthcare in the past year, and having had a problem paying for care from a doctor or hospital in the past year. The majority of the Vietnamese had a usual source of care (91.7%) and reported having had a personal doctor (63.6%), but a majority of the women report that they had never had a Pap smear (53.3%). Among the vulnerable measures, although a majority (63.3%) reported no emergency department visits in the past year, more than a third (36.7%) reported at least one visit, with 11% reporting more than one visit. Emergency visits are an expensive way for this adult population in their prime adult years to receive healthcare. Use of emergency departments may be an indicator of lack of access to healthcare on a routine basis. Other indicators that roughly one-fifth of the sample had difficulty obtaining healthcare included difficulty getting needed care in the past year (17.7%), and problems paying doctor or hospital bills in the past year (17.7%).

    Results

    Predisposing Factors

    The relationship of English-speaking ability to the predisposing variables is presented in Table 5.

    Age. The age of the population studied was between 18 and 60 years of age. Non-English speakers were generally older than English speakers. Although the direction of the relationship was in the direction predicted, the results were not statistically significant (T = .300, p = .382) and the relationship was negligible (γ = .086). Among the English speakers, 46.2% were between the ages of 18 and 29, while only 21.6% of non-English speakers were between the ages of 18 and 29. Furthermore, 53.9% of the English speakers were between the ages of 40 and 60 whereas 65.8% of non-English speakers were in this age range.

    Gender. As shown in Table 5, 84.5% of the English speakers were female compared to 82.6% of the non-English speakers. The English- and non-English-speaking groups were not significantly different on gender (T = .157, p = .382).

    Martial Status. Among English speakers, 69.2% were married and living with a spouse at the time of the interview compared to only 56.5% of non-English speakers. These results were not statistically significant (T = .221, p = .349) although the results were in the expected direction with non-English speakers being more disadvantaged on this variable in terms of social capital resources.

    Citizenship. On the vulnerable domain variable of citizenship status, 53.8% of the English speakers were already U.S. citizens and the rest were in the process of applying for citizenship. Among the non-English speakers, only 23.8% were U.S. citizens, 33.3% were in the process of applying, while 42.8% had not begun the application process. The relationship was significant (T = 3.145, p = .001) and very strong (γ = .694).

    Enabling Factors

    The relationship between English-speaking ability and the enabling variables is presented in Table 6. All but one of the relationships with English-speaking ability was significant and in the predicated direction. The relationships were moderate to very strong, providing support for Hypothesis 2 that non-English-speaking immigrants would have fewer enabling resources and face more barriers in accessing healthcare than English-speaking immigrants.

    The English-speaking immigrants were more likely to work and to work more hours than the non-English-speaking immigrants. Whereas 43.5% of the non-English-speaking immigrants did not work, only 23.1% of the English-speaking immigrants were not working. Nearly one-fourth of the English-speaking immigrants worked in excess of 40 hours a week compared to only 4.3% of the non-English-speaking immigrants. This relationship was moderately strong (γ = -.412) and statistically significant (T = 1.721, p = .043).

    Not surprisingly, non-English-speaking Vietnamese immigrants were three times as likely to need an interpreter (43.5%) as the English-speaking immigrants (15.4%). This relationship was statistically significant (T = 1.919, p = .028) and very strong (γ = -.618).

    Non-English-speaking Vietnamese immigrants were much more likely to need someone to go to the doctor with them than English-speaking Vietnamese immigrants. While 23.1% of English-speaking immigrants needed someone to accompany them to the doctor, 60.9% of non-English speakers did so. This relationship was statistically significant (T = 2.374, p = . 009) and very strong (γ = -.677).

    Furthermore, the accompanying person for the non-English-speaking Vietnamese immigrants was much more likely to have to take time off from work to go to the doctor with them than was true for the English-speaking immigrants. None of the English speakers needed to have someone take time off work to go to the doctor with them compared to 35.7% of non-English speakers. The results approached statistical significance (T = -1.491, p = .068) and the relationship was very strong (γ = -1.00).

    Among the working population, non-English-speaking Vietnamese immigrants were more likely to not have paid sick leave from work than the English speakers. One-third (33.3%) of working English speakers were able to take off work with pay when sick compared to only 7.7% of the working non-English speakers. This relationship was in the direction predicted, but was not statistically significant (T = .267, p = .790).

    Non-English-speaking Vietnamese immigrants were more likely to have problems getting transportation for healthcare appointments in the past year than were English-speaking immigrants. None of the English-speaking patients had a problem getting transportation compared to 8.7% of non-English-speaking immigrants. The relationship was statistically significant (T = -1.465, p = .007) and very strong (γ = -1.000).

    Health Factors

    The relationships between English-speaking ability and the measures of health among the Vietnamese immigrant population are shown in Table 7. As predicted, the immigrants who did not speak English were less healthy on all four measures. The relationships were all statistically significant and all were very strong.

    The majority of Vietnamese immigrants that did not speak English (69.9%) rated their overall health status as fair or poor compared to less than one-fourth (23.1%) of the immigrants who speak English. Almost half (46.2%) of the English speakers rated their health as excellent or very good compared to only 4.3% of the non-English speakers. This relationship was statistically significant (T = 3.487, p = .000) and very strong (γ = .774).

    Vietnamese immigrants who did not speak English were more likely to be limited in their ability to work because of physical, mental, or emotional problems than English-speaking immigrants. None of the English-speaking immigrants were limited in their ability to work, while almost one in five (18.2%) of the non-English speakers were limited. This relationship was statistically significant (T = 2.169, p = .015) and very strong (γ = -1.000).

    Non-English-speaking Vietnamese immigrants were more likely to feel healthy and full of energy within 30 days preceding the survey than English speakers. Among the English speakers, 61.6% reported that they felt healthy and full of energy everyday or most days compared to only 23.8% of non-English speakers. The relationship was statistically significant (T = 2.262, p = .012) and very strong (γ = .673).

    Vietnamese immigrants that did not speak English were more likely to report feeling worried, tense, or anxious in the preceding 30 days. Among English speakers, only 7.7% reported they felt worried, tense, or anxious everyday or most days, compared to 28.6% of non-English speakers. This relationship was statistically significant (T = -1.678, p = .047) and very strong (γ = -.655).

    Use of Healthcare Services

    The relationships between English-speaking ability and the use of healthcare services are presented in Table 8. All of these relationships were in the predicted direction and all were moderate to strong relationships. Although only one of the relationships was significant at the .05 level or less, all approach significance (p < .10 and > .05).

    Of the Vietnamese immigrants who spoke English, all of them reported having had a usual source of care. Of the Vietnamese immigrants who did not speak English, 87%, had a usual source of care, but 13% did not. This relationship was statistically significant (T = 1.828, p = .034) and very strong (γ = 1.000).

    Vietnamese immigrants who spoke English were more likely to have had a personal doctor than Vietnamese immigrants who did not speak English. Of the Vietnamese immigrants who spoke English, 76.9% reported having had a personal doctor while only 55% of the non-English-speaking immigrants said they had a personal doctor. This relationship approached statistical significance (T = 1.352, p = .088) and was a moderately strong relationship (γ = .463).

    Among the women, Vietnamese immigrants who spoke English were more likely to have had a Pap smear screening to test for abnormal cells than were Vietnamese immigrants who did not speak English. Of the Vietnamese immigrants who spoke English, 63.6% had received a Pap smear compared to only 36.8% of the non-English-speaking women. This relationship was moderately strong (γ = .500) and approached statistical significance (T = 1.453, p = .073).

    Vietnamese immigrants who spoke English were less likely to have been seen by a doctor in the emergency department than Vietnamese immigrants who did not speak English. Of the Vietnamese immigrants who spoke English, only 23.1% had seen a doctor in the emergency department in the past year while 43.4% of the non-English speakers had had an emergency department visit. This relationship approached statistical significance (T = 1.247, p = .106) and was moderately strong (γ = -.579).

    Vietnamese immigrants who spoke English were less likely to have had a problem getting needed healthcare within the past 12 months than those who did not. Among the Vietnamese immigrants who did speak English, 7.7% reported problems accessing needed healthcare in the past year compared to 22.7% of the immigrants who did not speak English. This relationship was strong (γ = -.558) and approached statistical significance (T = -1.288, p = .099).

    Vietnamese immigrants who did speak English were less likely to have had a problem paying doctor and hospital bills than Vietnamese immigrants who did not speak English. Among the Vietnamese immigrants who spoke English, 7.7% had a problem paying doctors and hospital bills in the past year, compared to 23.8% of the immigrants who did not speak English. This relationship approached statistical significance (T = 1.349, p = .089) and was a strong relationship (γ = -.579).

    Discussion

    This study included data from the population of all Vietnamese immigrant patients seen in a study of a large, urban, safety-net healthcare system in a county in north Texas. In the study, all patients with Asian surnames were selected for inclusion in the telephone interviews because these patients were small in number and the researchers reasoned that they could get the most representative sample of these patients possible by oversampling this population. In all, 36 immigrants of Vietnamese origin were identified.

    The study was designed to examine the effect of English-speaking ability on factors that have been found to be related to use of healthcare services in the United States. The Behavioral Model for Vulnerable Populations was used as the theoretical framework for the study. The hypotheses for this research predicted that lack of ability to speak English would be positively correlated with factors that would predispose immigrants to need to use healthcare services; negatively correlated with factors that would enable them to use healthcare services when they were needed; negatively correlated with measures of health that could trigger actual use of healthcare services; and negatively correlated with use of healthcare services. Even with a very small number of cases, these hypotheses were all supported by the data.

    The most significant predisposing factor was citizenship status. English speakers were much more likely to be citizens or to be applying for citizenship than were non-English speakers. In the safety-net system in which the immigrants were receiving healthcare, only U.S. citizens who were legal residents of the county were eligible to receive reduced rate healthcare on a sliding scale based on income. Noncitizens and nonresidents could receive care in the system only if they paid full price for their care, so lack of citizenship would create a financial barrier for the immigrants.

    The non-English-speaking immigrants were more likely to be disadvantaged on the enabling variables as well. They were less likely to be working, and more likely to need an interpreter and to need someone to go with them to their healthcare appointments. The person who accompanied them was more likely to need to take time off from work to go with them and they were more likely to have problems getting transportation for their healthcare visits. All these factors are likely to create barriers for use of healthcare services.

    The non-English-speaking immigrants were also in poorer health than the English-speaking immigrants. The majority of the non-English speakers reported that their health was only fair or poor; they were more likely to be limited in their ability to work because of physical, mental, or emotional problems; they were less likely to feel energetic and healthy; and they were more likely to report feeling worried, tense, and anxious. Finally, in terms of measures of access to healthcare, the non-English speakers also fared worse. They were less likely to have a usual source of care or a personal doctor. The non-English-speaking women were less likely to have ever had a Pap smear. The non-English speakers were also more likely to have been seen in the emergency department, to report a problem getting healthcare in the past year, and to have had problems paying doctor or hospital bills in the past year.

    There are many findings that suggest structural adjustments that could be made in the healthcare system to make it more accommodating to the Vietnamese population. In terms of community involvement, it would be a good idea to promote access to healthcare such as preventive measures and primary care through outlets such as churches and schools. Because the non-English-speaking immigrants have a tendency to be demographically older, it may be a good idea to target them at Vietnamese shopping complexes and community centers. Promoting preventive health through health fairs and community health education could reduce emergency visits through prevention of health problems and for early detection of chronic illnesses. It is also important to educate all immigrants, English-speaking and non-English-speaking, on the benefits they are offered and can receive such as insurance from both public and private outlets. Also, because the Asian population is generally adverse to referrals for mental health problems, it is suggested that there should be a community-wide intervention to promote a greater understanding of these problems and the role that therapy can play in helping people who have these issues. By reducing barriers to language, transportation, and financing, it is possible to reduce the costs of reliance on emergency care, for both healthcare consumers and healthcare facilities.

    Some other specific recommendations include increasing the availability of professional translators regionally. In a county with a large Vietnamese population such as Tarrant County, it would be useful to have a Vietnamese translator available daily. The development of culturally appropriate health education classes, patient education materials, and outreach materials for Vietnamese immigrants would be of help for the people of Tarrant County’s JPS safety-net system.

    Limitations

    In conclusion, it is important to bear in mind that this study included patients in only one safety-net healthcare system. There were few Asians in the overall sample (100 out of a sampling frame of about 10,000 patients) and of the Asians in the sample, only one-third met the criteria to be studied in this paper (36 out of 100). This secondary study included more women and fewer men so that future studies should try to reach more men. Finally, there was no qualitative study of the immigrants in this population so it was not possible to understand the uniquely Vietnamese perspective on the healthcare system and problems the immigrants had accessing it. Future research should include qualitative interviews or focus groups to gain a more in-depth understanding of the Vietnamese perspective.

    References

    • Gelberg, L., Andersen, R. M., & Leake, B. D. (2000). The behavioral model for vulnerable populations: Application for medical care use and outcomes for homeless people. Health Services Research 34(6), 1273–1301.
    • Andrulis, D. P. (2003). Reducing racial and ethnic disparities in disease management to improve health outcomes. Practical Disease Management, 11, 789–800.
    • Campbell, J. L., Ramsay, J., & Green, J. (2001). Age, gender, socioeconomic, and ethnic differences in patient's assessment of primary health care. Quality Health Care, 10, 90–95.
    • Dinh, A., Kemp, C., Rasbridge, L., Massagli, M. P., Phillips, R. S., & Iezzoni, L. I. (2000). Culture and the end of life: Vietnamese health beliefs and practices related to the end of life. Journal of Hospice and Palliative Nursing, 2, 111–117.
    • Eden, J. (1998). Measuring access to care through population-based surveys: Where are we now? Health Services Research, 33, 685–707.
    • Green, A. R., Ngo-Metzger, Q., Legedza, A. T., Massagli, M. P., Phillips, R. S., & Iezzoni, L. I. (2005). Interpreter services, language concordance, and health care quality: Experiences of Asian Americans with limited language proficiency. Journal of General Internal Medicine, 20 (11), 1050–1056.
    • Ngo-Metzger, Q., Massagli, M. P., Clarridge, B. R., Manocchia, M., Davis, R. B., Iezzoni, L. I. et al. (2003). Linguistic and cultural barriers to care: Perspectives of Chinese and Vietnamese immigrants. Journal of General Internal Medicine, 18 (1), 44–52.
    • Poss, J. E., Kemp, C., Rasbridge, L., Massagli, M. P., Phillips, R. S., & Iezzoni, L. I. (2001). Developing a new model for cross-cultural research: Synthesizing the health belief model and the theory of reasoned action. Advances in Nursing Science, 23, 1–15.
    • U.S. Census Bureau (2000). The Asian population: 2000. Retrieved March 2, 2007, from http://www.census.gov/prod/2002pubs/c2kbr01-16.pdf.

    Figure 1: Behavior Model for Vulnerable Populations

    Figure 1. Behavior Model for Vulnerable Populations

    Table 1: Frequency Table for Predisposing Variables

    Variable Frequencies Percents
    Age
    18-29 years
    30-39 years
    40-49 years
    50-60 years
       Total

    11
    3
    11
    11
    36

    30.6%
    8.3%
    30.6%
    30.6%
    100.0%
    Sex
    Female
    Male
        Total

    30
    6
    36

    83.3%
    16.7%
    100.0%
    Currently Married, living with spouse
    Yes
    No
        Total

    22
    14
    36

    61.1%
    38.9%
    100.0%
    Citizenship status
    U.S. citizen
    Applying for citizenship
    Not yet qualified
    Need permanent residency
        Total
        Missing data

    12
    13
    7
    2
    34
    2

    35.5%
    38.2%
    20.6%
    5.9%
    100.0%

    Table 2: Frequencies for Enabling Variables

    Variables Frequencies Percents
    Hours worked a week
    None
    1-20 hours a week
    21-37 hours a week
    38-44 hours a week
    45 or more hours a week
        Total

    13
    0
    9
    10
    4
    36

    36.1%
    0.0%
    25.0%
    27.8%
    11.1%
    100.0%
    Needed an interpreter past year
    Yes
    No
        Total

    12
    24
    36

    33.3%
    66.7%
    100.0%
    Someone go with you to doctor
    Yes
    No
        Total

    17
    19
    36

    47.2%
    52.8%
    100.0%

    Person had to take time off
    Yes
    No
        Total
        Missing data


    5
    11
    16
    20

    31.3%
    68.8%
    100.0%
    Gets paid time off when sick
    Yes
    No
        Total
        Missing data

    2
    20
    22
    14

    9.1%
    90.9%
    100.0%
    Problems with transportation past year
    Yes
    No
        Total

    2
    34
    36
    5.6%
    94.4%
    100.0%

    Table 3: Frequencies for Health Variables

    Variables Frequencies Percents
    Overall health
    Excellent, very good
    Good
    Fair, Poor
        Total

    7
    10
    19
    36

    19.4%
    27.8%
    52.8%
    100.0%
    Limited in work because of mental or physical problem
    Yes
    No
        Total
        Missing data


    4
    31
    35
    1


    11.4%
    88.0%
    100.0%
    Full of energy
    Every day, most days
    Some days, never
        Total
        Missing data

    13
    21
    34
    2

    38.2%
    61.8%
    100.0%
    Felt worried
    Every day, most days
    Some days, never
        Total
        Missing data

    7
    27
    34
    2
    20.6%
    79.4%
    100.0%

    Table 4: Frequency Table for Healthcare Utilization Variables

    Variables Frequencies Percents
    Has a usual source of care
    Yes
    No
        Total

    33
    3
    36

    91.7%
    8.3%
    100.0%
    Has personal doctor
    Yes
    No
        Total
        Missing data

    21
    12
    33
    3
    63.6%
    36.4%
    100.0%
    Had Pap smear ever
    Yes
    No
        Total
        Missing data

    14
    16
    30
    6
    46.7%
    53.3%
    100.0%
    Number of emergency department visits in past year
    0 visits
    1 visit
    2 or more visits
        Total


    23
    9
    4
    36


    63.9%
    25.0%
    11.1%
    100.0%
    Problem getting needed care in past year
    Yes
    No
        Total
        Missing data


    6
    29
    35
    1

    17.7%
    82.9%
    100.0%
    Problem paying doctor or hospital bill in past year
    Yes
    No
        Total
        Missing data


    6
    28
    34
    2

    17.6%
    82.4%
    100.0%

    Table 5: Relationship of Ability to Speak English to Predisposing Variables

    Predisposing Variables Speaks English Approx. T
    Approx. sig. level (1-sided)
    Gamma (γ)
    Yes No
    Age
    18–29 yrs
    30–39 yrs
    40–49 yrs
    50–60 yrs

    6 (42.6%)
    0 (00.0%)
    2 (15.4%)
    5 (38.5%)
    n = 13

    5 (21.6%)
    3 (13.0%)
    9 (39.1%)
    6 (26.1%)
    n = 23

    T =.300
    p =.382

    .086
    Sex
    Female
    Male

    11(84.5%)
    2 (15.4%)
    n = 13

    19(82.6%)
    4 (17.4%)
    n = 23

    T = .157
    p = .382

    .073
    Married, live with spouse
    Yes
    No

    9 (69.2%)
    4 (30.8%)
    n = 13

    13 (56.6%)
    10 (43.5%)
    n = 23

    T = .221
    p = .349

    .268
    Citizenship status
    U.S. citizen
    Applying for citizenship
    Not yet qualified
    Need permanent residency

    7 (53.8%)
    6 (46.2%)
    0 (00.0%)
    0 (00.0%)
    n = 13

    5 (23.5%)
    7 (33.3%)
    7 (33.3%)
    2 (9.5%)
    n = 21

    T = 3.145
    p = .001*

    .694

    * p<.05 **p<.10

     

    Table 6: Relationship of Ability to Speak English to Enabling Variables

    Enabling Variables Speaks English Approx. T
    Approx. sig. level (1-sided)
    Gamma (γ)
    Yes No
    Hours work per week
    None
    20–37 hrs
    38–44 hrs
    45–96 hrs

    3 (23.1%)
    3 (23.1%)
    4 (30.8%)
    3 (23.1%)
    n = 13

    10 (43.5%)
    6 (26.1%)
    6 (26.1%)
    1 (4.3%)
    n = 13

    T = -1.721
    p = .043*

    -.412
    Needed interpreter last year
    Yes
    No

    2 (15.4%)
    11(84.6%)
    n = 13

    10 (43.5%)
    13 (56.5%)
    n = 23

    T = -1.919
    p = .028*

    -.618
    Need someone with you to go to doctor
    Yes
    No

    3 (23.1%)
    10 (76.9%)
    n = 13

    14 (60.9%)
    9 (39.1%)
    n = 23

    T = -2.374
    p = .009*

    -.677
    Person takes time off work
    Yes
    No

    0 (00.0%)
    2 (100.0%)
    n = 2

    5 (35.7%)
    9 (64.3%)
    n = 14

    T = -1.491
    p = .068**

    -1.000
    Paid sick leave
    Yes
    No

    1 (11.1%)
    8 (88.9%)
    n = 9

    1 (7.7%)
    12 (92.3%)
    n = 13

    T = .267
    p = .790

    .713
    Problems with transportation past year
    Yes
    No

    0 (00.0%)
    13 (100.0%)
    n = 13

    2 (8.7%)
    21 (91.3%)
    n = 23

    T = -1.465
    p = .007*

    -1.000

    * p<.05 **p<.10

    Table 7: Relationship of Ability to Speak English to Need Variables

    Need Variables Speaks English Approx. T
    Approx. sig. level (1-sided)
    Gamma (γ)
    Yes No
    Overall health
    Excellent, very good
    Good
    Fair, poor

    6 (46.2%)
    4 (30.8%)
    3 (23.1%)
    n = 13

    1 (4.3%)
    6 (26.2%)
    16 (69.9%)
    n = 23

    T = 3.487
    p = .000*

    .774
    Limited in work for physical/mental problem
    Yes
    No

    0 (00.0%)
    13(100.0%)
    n = 13

    4 (18.2%)
    18 (81.8%)
    n = 23

    T = -2.169
    p = .015*

    -1.000
    Feel healthy, full of energy
    Every day, most days
    Some days, never

    8 (61.5%)
    5 (38.5%)
    n = 13

    5 (23.8%)
    16 (76.2%)
    n = 21

    T = 2.262
    p = .012*

    .673
    Number of days worried, tense, anxious
    Every day, most days
    Some days, never

    1 (7.7%)
    12 (92.3%)
    n = 13

    6 (28.6%)
    15 (71.4%)
    n = 21

    T = -1.678
    p = .047*

    -.655

    *p<.05 **p<.10

    Table 8: Relationship of Ability to Speak English to Utilization Variables

    Utilization Variables Speaks English Approx. T Approx. sig. level (1-sided) Gamma (γ)
    Yes No
    Has usual source of care
    Yes
    No

    13(100.0%)
    0 (00.0%)
    n = 13

    20 (87.0%)
    3 (13.0%)
    n = 23

    T = 1.828
    p = .034*

    1.000
    Has personal doctor
    Yes
    No

    10 (76.9%)
    3 (23.1%)
    n = 13

    11 (55.0%)
    9 (45.0%)
    n = 20

    T = 1.352
    p = .088**

    .463
    Ever had a Pap smear
    Yes
    No

    7 (63.6%)
    4 (36.4%)
    n = 11

    7 (36.8%)
    12 (63.2%)
    n = 19

    T = 1.453
    p = .073**

    .500
    Number times seen in emergency department past year
    None
    1 time
    2+ times

    10 (76.9%)
    2 (15.4%)
    1 (7.7%)
    n = 13

    13 (56.5%)

    7 (30.4%)
    3 (13.0%)
    n = 23


    T = 1.247
    p = .106**

    .395
    Problem getting needed healthcare in past year
    Yes
    No

    1 (7.7%)
    12 (92.3%)
    n = 13

    5 (22.7%)
    17 (77.3%)
    n = 22

    T = -1.288
    p = .099**

    -.558
    Problems paying doctor or hospital bill last year
    Yes
    No

    1 (7.7%)
    12 (92.3%)
    n = 13

    5 (23.8%)
    16 (76.2%)
    n = 21

    T = -1.349
    p = .089**

    -.579

    * p<.05 **p<.10