Journal of Immigrant and Minority Health

, Volume 14, Issue 6, pp 933–940

Prevalence of Chronic Disease and Insurance Coverage among Refugees in the United States


    • Department of PediatricsUniversity of Pennsylvania, Perelman School of Medicine
    • The Children’s Hospital of Philadelphia
  • Elena Fuentes-Afflick
    • Department of PediatricsUniversity of California, San Francisco School of Medicine
  • Mayur M. Desai
    • Department of Chronic Disease EpidemiologyYale School of Public Health
Original Paper

DOI: 10.1007/s10903-012-9618-2

Cite this article as:
Yun, K., Fuentes-Afflick, E. & Desai, M.M. J Immigrant Minority Health (2012) 14: 933. doi:10.1007/s10903-012-9618-2


Little is known about the health status of refugees beyond the immediate post-arrival period in the US. Using data from the 2003 New Immigrant Survey, a nationally representative survey of immigrants who had recently become legal permanent residents, we determined the prevalence of chronic conditions and health insurance coverage among adult refugees who had lived in the US for at least 1 year (n = 490). We compared their health status with that of other immigrants (n = 3,715) using multivariable logistic regression. The median duration of US residency was 5.6 and 8.0 years among refugees and other immigrants, respectively. Refugees were more likely than other immigrants to report at least one chronic condition (24.7 vs. 15.6 %, P < 0.001). After adjusting for sociodemographic differences, the odds of the following conditions remained significantly higher among refugees: arthritis (adjusted odds ratio [AOR] = 1.67, 95 % confidence interval [CI] = 1.07, 2.61), heart disease (AOR = 2.49, 95 % CI = 1.30, 4.74), stroke (AOR = 5.87, 95 % CI = 1.27, 27.25), activity-limitation due to pain (AOR = 1.96, 95 % CI = 1.31, 2.93), and any chronic condition (AOR = 1.37, 95 % CI = 1.03, 1.81). Although similar percentages of refugees (49.0 %) and other immigrants (47.4 %) were uninsured, 46.5 % of refugees with chronic conditions lacked health insurance. Refugees have a high burden of chronic disease and would benefit from expanded insurance coverage for adults with preexisting conditions.


Emigration and immigrationRefugee/chronic diseaseRefugee/health insurance


In the past decade, over one million people have immigrated to the United States (US) because of armed conflict or persecution.1 Among these immigrants, the majority are refugees and asylees, defined as individuals who are unable to return to their home countries due to “persecution or a well-founded fear of persecution on account of race, religion, nationality, membership in a particular social group, or political opinion”[1]. Refugees and asylees differ in that refugees undergo a determination of immigration status before entering the US and asylees undergo a determination of status after entering the US. For simplicity, we use the term “refugee” to refer to all such immigrants.

Historically, US health policies relating to refugees have focused on screening for communicable conditions. However, the prevalence of chronic disease has risen in low- and middle-income countries [2, 3]. In the last decade, more than half of refugees who have resettled in the US originated in nations that have diabetes and hypertension prevalence rates similar to those in the US [46].2 In addition, although US immigration processes may generally favor a relatively young and healthy population of immigrants [7, 8], the events that motivate refugees to migrate may negate these migration-associated selection factors. Despite these concerns, there is limited information about refugees’ health status, access to care, or insurance coverage after the immediate entry period [9].

To date, our understanding of refugee health status in the US pertains to the period immediately following arrival, to infectious diseases, to specific ethnic groups, or to clinic-based samples [1018]. Nationally representative data after the immediate entry period are not available. In Europe, community-based studies of refugees have documented high rates of poor health status (52 %) and hospitalization (12 % in the previous 12 months) [1921]. However, European policies pertaining to public funding for health services for immigrants and to the detention of individuals seeking political asylum are different from US policies [22, 23]. Because health care access and immigration detention policies are likely to be associated with differences in health status, it is difficult to generalize from the European experience to the US.

Additional information on the burden of chronic disease and insurance coverage for refugees in the US would better inform primary care, preventive health, and insurance outreach programs for immigrants. Thus, the two objectives of this study were to: (1) describe the prevalence of chronic disease and insurance coverage for refugees living in the US beyond the immediate post-arrival period; and (2) test the hypothesis that disease prevalence and insurance coverage are less favorable among refugees than among other US immigrants.


Data Source

This is a secondary data analysis of the 2003 New Immigrant Survey (NIS) Adult Sample, a nationally representative, cross-sectional, self-reported survey of 8,573 immigrants ages 18 years and older who had recently received legal permanent residency in the US [24]. A detailed description of the NIS is available elsewhere [25]. Briefly, the NIS includes a multistage probability sample of all adults admitted to legal permanent residence from May through November of 2003. The sampling frame was compiled from electronic administrative records held by the US government. The dataset oversampled immigrants who were admitted through employment and diversity programs. Survey instruments were translated into ten languages (Amharic, French, Haitian Creole, Korean, Mandarin, Polish, Russian, Spanish, Tagalog, and Vietnamese) and key concepts were translated into nine additional languages (Arabic, Croatian, Farsi, French, Gujarati, Hindi, Serbian, Ukrainian, and Urdu). Adults were interviewed in-person or on the telephone in the language of their choice within 3 months, on average, after becoming legal permanent residents (LPRs). The overall response rate was 69 %.

Study Sample

For the purposes of this analysis, we limited the NIS dataset to the subgroup of adult respondents who were 18–64 years old at the time of the survey and who had been living in the US for at least 1 year prior to survey administration (n = 4,205). Individuals who met these criteria were then categorized as refugees or non-refugees according to their pre-LPR visa category. Individuals who had been refugees prior to receiving legal permanent residency were categorized as “refugees,” as were immigrants who were admitted to the US as spouses of refugees (n = 490). Individuals who lived in the US as temporary students, workers, or visitors; undocumented immigrants; or immigrant spouses, fiancés, or children of US citizens or LPRs comprised the comparison group of “other immigrants” (n = 3,715). Assuming a prevalence of at least 10 % for any chronic condition, a two-sided alpha of 0.05, and sample sizes as described above [490 refugees and 3,715 non-refugee immigrants], we had at least 80 % power to detect an odds ratio of 1.5 or greater.


Survey respondents answered questions about health status that were based on items from the National Health Interview Survey. Overall health status was assessed using self-rated general health (excellent, very good, good, fair, or poor). Information about the prevalence of chronic disease was determined using self-reported information on previous medical diagnoses of arthritis, behavioral health problems (“any emotional, nervous, or psychiatric problem” or currently taking “tranquilizers, antidepressants, or pills for nerves”), cancer (excluding minor skin cancers), chronic lung disease (such as asthma, chronic obstructive pulmonary disease, or emphysema), diabetes, heart disease (including angina, congestive heart failure, coronary artery disease, or heart attack), hypertension, and stroke. Respondents were also asked if frequent pain made it “difficult for you to do your usual activities such as household chores or work.”

Health insurance status was based on current insurance coverage and categorized as private, public (e.g., Medicaid), or none.

Sociodemographic characteristics associated with health and insurance status were assessed and included: age, time in the US, sex, region of origin (East, South, and South East Asia; Europe and Central Asia; Latin America [Mexico, Central America, South America] and the Caribbean; Middle East and North Africa; Oceania and Canada; Sub-Saharan Africa), current marital status, educational attainment (less than high school, high school, college, postgraduate), self-rated English proficiency (very well/well vs. not well/not at all), household employment status, and income category (using the 2003 federal poverty level [FPL] for a family of four, $18,400: ≤FPL, 101–200 % FPL, 201–300 % FPL, 301–400 % FPL, and >400 % FPL).


Sociodemographic differences between refugees and other immigrants were assessed using t-tests (for continuous variables) and χ2 tests (for categorical variables). Bivariate analyses, using the χ2 test, were performed to compare the prevalence of chronic conditions and insurance status between refugees and other immigrants. Finally, for each dependent variable, we fit a multivariable logistic regression model that included refugee status and controlled for sociodemographic variables significant at the 0.20 level in bivariate analyses. For health status and chronic disease models, covariates comprised: age, time in the US, sex, region of origin, marital status, education, English proficiency, unemployment, and income. For insurance coverage models, covariates comprised: age, time in the US, sex, region of origin, marital status, education, English proficiency, unemployment, income, and presence of any chronic condition. All analyses were performed using SAS Version 9.2 software. The PROC SURVEY procedures and sampling weights were used to account for the complex sample design. PASS 2008 Version 8.0.12 was used for all power calculations. The institutional review board of Yale School of Medicine approved and monitored the conduct of this study, and the Office of Population Research at Princeton University approved the use of the New Immigrant Survey for this analysis.


Sociodemographic characteristics of the sample are summarized in Table 1. On average, refugees were older than other immigrants (mean age 39.1 vs. 35.3 years, P < 0.001) and had lived in the US for a shorter period of time (mean time 78 months vs. 110 months, P < 0.001). The majority of immigrants in both groups had lived in the US for more than 5 years. Approximately two-thirds of refugees originated in Europe and Central Asia (34.1 %) or Latin America and the Caribbean (28.3 %). The comparison group of other immigrants was more likely to be from Latin America and the Caribbean (58.1 %) or Asia (21.4 %); only 10.9 % originated from Europe and Central Asia. Educational attainment differed significantly between refugees and other immigrants (P < 0.001), with refugees being less likely to have completed college or received a postgraduate degree (25.3 vs. 30.2 %). Refugees were also more likely to report speaking English “not well” or “not at all” (48.9 vs. 39.1 %, P < 0.001) and were more likely to be living in poverty at the time of the survey (23.6 vs. 17.7 %, P < 0.001).
Table 1

Characteristics of adult refugees and other immigrants (18 to 64 years), by immigrant subgroup: New Immigrant Survey, 2003


Refugees (N = 490)

Other immigrants (N = 3,715)


Age, years mean (SE)

39.1 (0.5)

35.3 (0.2)


Time in the US, months mean (SE)

78.0 (2.5)

110.0 (1.5)


Female (%)




Region of Origin (%)



 East, South, and South East Asia




 Europe and Central Asia




 Latin America and the Caribbean




 Middle East and North Africa




 Oceania and Canada




 Sub-Saharan Africa




Married (%)




Education (%)



 No Degree




 High School












English proficiency, self-rated (%)


 Not well/not at all




Unemployed (%)




Poor (Income ≤FPL, %)




In comparison to other immigrants, refugees had a higher odds of rating their health status as “fair” or “poor” (14.6 vs. 8.70 %, odds ratio [OR] = 1.80, 95 % confidence interval [CI] = 1.34, 2.41; Table 3) and of having any chronic condition (24.7 vs. 15.6 %, OR = 1.78, 95 % CI = 1.41, 2.26). For specific conditions, refugees reported significantly higher rates of arthritis (OR = 2.36, 95 % CI = 1.53, 3.63), behavioral health problems (OR = 1.71, 95 % CI = 1.06, 2.76), heart disease (OR = 3.39, 95 % CI = 1.79, 6.42), hypertension (OR = 1.98, 95 % CI = 1.43, 2.74), and stroke (OR = 4.55, 95 % CI = 1.08, 19.20). The odds of having pain that limited participation in usual activities were also higher among refugees than other immigrants (OR = 3.18, 95 % CI = 2.25, 4.50). After adjusting for sociodemographic differences between immigrant subgroups, refugees remained more likely to report poor health status (adjusted OR [AOR] = 1.85, 95 % CI = 1.29, 2.65), arthritis (AOR = 1.67, 95 % CI = 1.07, 2.61), heart disease (AOR = 2.49, 95 % CI = 1.30, 4.74), stroke (AOR = 5.87, 95 % CI = 1.27, 27.25), any chronic condition (AOR = 1.37, 95 % CI = 1.03, 1.81), and activity-limiting pain (AOR = 1.96, 95 % CI = 1.31, 2.93).

At the time of survey administration, nearly half of refugees and other immigrants in this sample were uninsured (49.0 and 47.4 %, respectively). Refugees were less likely to have private insurance (34.8 vs. 47.8 %, P < 0.001) and more likely to have public insurance (16.2 vs. 4.8 %, P < 0.001) than other immigrants (Table 2). Nearly one-quarter (23.1 %) of uninsured refugees reported at least one chronic condition (Fig. 1). Furthermore, 46.5 % of refugees with any chronic condition reported being uninsured (data not shown). The majority of uninsured refugees (87.6 %) were from households in which either the primary immigrant or the spouse was employed. In these households, the median household income was $24,500 (data not shown). After adjusting for differences in socioeconomic variables and health status, refugees remained less likely to have private insurance and more likely to have public health insurance (Table 3).
Table 2

Health status and insurance coverage for adult refugees and other immigrants (18 to 64 years), by immigrant subgroup: New Immigrant Survey, 2003



Other immigrants


Health Status (%)

 General health status



  Excellent/very good
















 Behavioral health problem








 Chronic lung disease








 Heart disease












 Any chronic conditiona




 Pain with activity limitationb




Insurance coverage (%)













aChronic conditions include arthritis, behavioral health problems, cancer (excluding minor skin cancers), chronic lung disease, diabetes, heart disease, hypertension, and stroke

bDifficulty participating in household or other work due to pain
Fig. 1

Health status of adult refugees (18 to 64 y), by insurance coverage: New Immigrant Survey, 2003. Fair/poor health: Fair or poor self-reported general health status. Any chronic condition: Chronic conditions include arthritis, behavioral health problems, cancer (excluding minor skin cancers), chronic lung disease, diabetes, heart disease, hypertension, and stroke. Pain with activity limitation: Difficulty participating in household or other work due to pain

Table 3

Logistic regression analysis of health status and insurance coverage for adult refugees in comparison to other immigrants (18 to 64 years): New Immigrant Survey, 2003


Unadjusted OR (95 % CI)a

Adjusted OR (95 % CI)a

Health Statusb

 Fair/poor health

1.80 (1.34, 2.41)

1.85 (1.29, 2.65)


2.36 (1.53, 3.63)

1.67 (1.07, 2.61)

 Behavioral health

1.71 (1.06, 2.76)

1.64 (0.94, 2.85)


1.94 (0.69, 5.50)

2.52 (0.89, 7.29)

 Chronic lung disease

0.87 (0.44, 1.72)

0.79 (0.35, 1.75)


1.52 (0.94, 2.45)

1.44 (0.77, 2.70)

 Heart disease

3.39 (1.79, 6.42)

2.49 (1.30, 4.74)


1.98 (1.43, 2.74)

1.44 (0.95, 2.18)


4.55 (1.08, 19.20)

5.87 (1.27, 27.25)

 Any chronic condition

1.78 (1.41, 2.26)

1.37 (1.03, 1.81)

 Pain with activity limitation

3.18 (2.25, 4.50)

1.96 (1.31, 2.93)

Insurance coveragec


0.59 (0.48, 0.72)

0.73 (0.57, 0.94)


3.73 (2.73, 5.09)

3.44 (2.29, 5.16)


1.04 (0.86, 1.27)

0.87 (0.68, 1.12)

aOR odds ratio, CI confidence interval. All comparisons are of refugees to non-refugee immigrants

bFor health status measures, the multivariable logistic regression models included the following covariates: age, time in the US, sex, region of origin, marital status, education, English proficiency, unemployment, and income

cFor insurance coverage, the multivariable logistic regression models included the following covariates: age, time in the US, sex, region of origin, marital status, education, English proficiency, unemployment, income, and presence of any chronic condition

Bold values indicate p <0.05


In this nationally representative sample of new legal permanent residents in the US, we found that refugees were significantly more likely to report chronic health problems than other immigrants. In addition, half of refugees were uninsured at the time they became LPRs, and one-quarter of uninsured refugees had at least one chronic health condition.

Refugees represent a distinct subgroup of immigrants. As compared with other groups of immigrants, refugees are less likely to be intentional migrants and are more likely to have emigrated because of displacement by violence or persecution [1]. Thus, refugees may be less likely to be subject to the migration selection factors that are thought to contribute to the “healthy immigrant effect” [7, 8, 26]. In fact, US refugee resettlement policy prioritizes “victims of torture or violence; physically or mentally disabled persons; [and] persons in need of urgent medical treatment not available in the first asylum country” [27]. Given these policy priorities, one would expect that refugees would report lower levels of health status than other immigrants. This may account for the persistence of health disparities between refugees and other long-term immigrants after adjusting for demographic and socioeconomic differences.

It is also possible that refugees are more vulnerable to worsening health status after arriving in the US [28, 29]. For example, behavioral and social factors that influence health—such as dietary practices, tobacco use, household structure, and proximity to coethnic communities—may vary between refugees and other groups of immigrants. Furthermore, refugees report lower levels of English proficiency and are more likely to be poor than other immigrants who are eligible for legal permanent residency [30]. As a result, refugees may experience more barriers to accessing and navigating the health system [31].

Increasingly, health care providers and public health officials recognize the importance of early access to primary care and chronic disease screening for refugees [3234]. However, these efforts have focused on the period immediately following arrival in the US, and funding for longer-term health care services is limited [35, 36]. At present, access to health care for refugees after the immediate post-arrival period is often challenging—particularly for uninsured people who have chronic health conditions.

Nearly half of adult refugees and non-refugee immigrants were uninsured when they received legal permanent residency. Although this is consistent with other studies demonstrating low rates of insurance coverage among immigrants [3739], it is somewhat surprising, given that refugees are eligible for Medicaid. It raises concerns about the long-term health and financial security of LPRs [40] and about the ability of safety-net programs to reach vulnerable immigrant groups. In contrast, in the general US population only one-fifth of adults aged 18–64 years were uninsured in 2003 [41], the year in which the NIS was conducted. Furthermore, uninsured refugees are not simply a self-selected group of healthy immigrants who have opted out of the insurance pool. In our sample, one-quarter of uninsured refugees had at least one chronic health problem. Moreover, half of refugees with at least one chronic health problem were uninsured.

In this context, we would expect that refugees would benefit from specific provisions of the Patient Protection and Affordable Care Act [42]. Specifically, premium subsidies, health insurance exchanges, expanded Medicaid eligibility for adults without dependent children, and community rating with guaranteed issue could improve insurance coverage for this subgroup of immigrants. However, refugees are also among those populations that are most likely to be excluded from program expansions due to limited English proficiency or limited health literacy. Proactive policies are needed to ensure that programs with the potential to substantially reduce the number of uninsured are successful in reaching refugees, asylees, and other immigrants.

Study results should be interpreted in light of specific limitations. The sample includes only immigrants who received legal permanent residency after having lived in the US for at least 12 months. While refugees are eligible to apply for legal permanent residency, many other long-term immigrants, including the majority of undocumented immigrants, are not eligible. As a result, undocumented immigrants, particularly highly marginalized undocumented immigrants, are unlikely to be proportionately represented in the NIS. Other low-income legal permanent residents may also be underrepresented in the NIS due to difficulties contacting and enrolling individuals with intermittent or no telephone service. Although the NIS response rate of 69 % is comparable to other surveys of immigrant populations [43] and NIS response rates were similar between sampling strata [44], there may be undetected differences between refugee and non-refugee non-respondents that could have affected the results. This study relied on self-reported health, chronic disease, and insurance status information. Although self-reported health status is widely used in epidemiologic studies [45] the single-item question on self-rated health may rely on perceptions of health that could be influenced by prior experiences with the health system or traumatic events. Chronic disease prevalence may also be underreported by respondents who are uninsured and who have forgone routine medical care. Finally, the US refugee population has changed since 2003, with increasing immigration from Iraq, Africa, and Asia. However, we believe that the NIS is unique among nationally representative surveys because immigration data is rarely included in health surveys. Thus, the present study offers the best available, nationally representative data on chronic disease and insurance coverage for refugees.

Future studies are needed to further explain differences in health status between refugees and other immigrants. Determining the etiology of health disparities among migrants is likely to require life course analyses, as the health of vulnerable, mobile populations may be affected by events prior to, during, and after immigration [4649]. Among highly mobile groups such as refugees, transnational, longitudinal or comparative studies may be necessary in order to fully describe social and epidemiological determinants of health outcomes for migrants following their arrival in the US.

New Contribution to the Literature

There are limited population-level data regarding the health status of refugees in the US beyond the immediate new-arrival period. Research and services for refugees have largely focused on health issues specific to immigrants who are newly arrived in the US, such as infectious disease screening and treatment. This study demonstrates that chronic disease management is an ongoing and important component of health care for refugees. Moreover, despite being eligible for means-tested public benefits, refugees, like other immigrants, are in need of improved access to health insurance.


Data from the Department of Homeland Security’s Immigration Yearbook were used to determine the number of people admitted to the US as refugees, asylees, or humanitarian parolees from 1999 to 2009.


Data from the Department of Homeland Security’s ImmigrationYearbook were used to determine country of origin for all new refugees and asylees from 2000 to 2009.


Copyright information

© Springer Science+Business Media, LLC 2012