Journal of Immigrant and Minority Health

, Volume 16, Issue 2, pp 189–194 | Cite as

Healthcare Utilization Among Hispanic Immigrants with Diabetes: Investigating the Effect of US Documentation Status

Original Paper

Abstract

Previous studies have not examined whether documentation status has an effect on healthcare utilization among US Hispanic immigrants with diabetes. A secondary analysis was conducted using data from the Pew Hispanic Center and Robert Johnson Wood Foundation’s 2007 Hispanic Healthcare Survey. Hispanic immigrants diagnosed with diabetes were included in analyses. The association between documentation status and healthcare utilization was assessed using logistic regressions. Of N = 577 Hispanic immigrants with diabetes, 80 % were documented immigrants and 81 % reported having visited a healthcare provider in the last 6 months. Adjusting for confounders, those who were undocumented faced higher odds of having seen a healthcare provider more than 6 months ago or never when compared to those who were documented (OR = 1.79; 95 % CI 1.01, 3.14). Unique opportunities in addressing healthcare disparities can be found in focusing on the Hispanic immigrant population living with diabetes.

Keywords

Healthcare disparities Undocumented immigrants Healthcare utilization Diabetes 

Introduction

Hispanics are one of the nation’s largest and fastest growing minority groups [1]. Currently, they make up about 15 % of the United States (US) population. If population trends continue, it is projected that Hispanics will more than double by 2050 and make up approximately 29 % of the US population. A vast majority of this population will be of Mexican–American and Central or South American descent [2].

Compared to the rest of the US adult population, Hispanic adults tend to have a lower prevalence of many chronic health conditions [3]. However, health population-based studies report that Hispanics are disproportionately affected by diabetes [4, 5] and tend to have worse glycemic control, in comparison to non-Hispanic whites [6, 7]. According to the Centers for Disease Control and Prevention, diabetes is the fifth leading cause of death among Hispanics in the US and is a leading cause of heart disease, stroke, kidney disease, blindness, and amputations [8].

Given that minority immigrant populations often experience health disparities in the US, often as a consequence of access and utilization [9, 10], it is important that these disparities are addressed when identified. Among the Hispanic population in particular, disparities exist in the access to health care for chronic diseases and conditions [11]. Differences in access to quality health care, social and cultural factors, or genetics may explain disparities in diabetes prevalence [8].

Previous literature has acknowledged the barriers to health care access (i.e. insurance coverage [12, 13], English language proficiency [1, 12], and level of acculturation [12, 14]). Mainous et al. [14] showed that Hispanics with low acculturation were more likely to be without a routine place for health care, have no health insurance, and have low levels of education. Research has also identified where the Hispanic population generally seeks health care services (i.e., community or public health centers [1], doctor’s office [1], or emergency room [1, 15]). All of these factors have an effect on healthcare utilization, or the frequency by which individuals go to see a healthcare provider. However, studies have not determined whether healthcare utilization is affected by documentation status (i.e., undocumented and documented immigrants).

The aims of this study are two-fold: (1) to determine if documented and undocumented Hispanic immigrants with diabetes differ by health insurance status and having a usual place of care and (2) to investigate the effect that documentation status has on health care utilization among diabetic Hispanic immigrants. Since health insurance enrollment requires proof of citizenship, undocumented immigrants may be left without health care coverage. This limits where they are able to go for affordable health care services. Thus, we are first hypothesizing that documented Hispanic immigrants with diabetes are more likely to have health insurance and a usual source of care compared to undocumented Hispanic immigrants with diabetes. If we assume that having health insurance and a usual source of care increases the likelihood of having visited a healthcare provider in the last 6 months, we can then hypothesize that undocumented Hispanic immigrants will be less likely to have visited a healthcare provider in the last 6 months, when compared to documented Hispanic immigrants.

Methods

Data Source

A secondary analysis of data from the 2007 Hispanic Healthcare Survey was conducted using a cross-sectional study design. The 2007 Hispanic Healthcare Survey was a public opinion survey conducted by the Pew Hispanic Center and the Robert Wood Johnson Foundation [16]. The survey was designed to provide up-to-date, accurate information on Hispanics and health care in the US. Specifically, it explores how the diverse characteristics of the Hispanic population influence their health information needs, their help-seeking behaviors, their access to health services, their perceptions of the quality of health care, and their level of understanding of chronic disease [3].

All interviews were conducted using trained interviewers on the Computer Assisted Telephone Interviewing (CATI) system. Sampling weights were applied to yield a statistically representative sample of Latinos within the contiguous US. This survey was administered to both males and females of Hispanic origin or descent from July 16, 2007 to September 23, 2007. The population surveyed included Hispanic immigrants aged 18–97. A total of 4,013 respondents were randomly selected from stratified listings of telephone area codes and surveyed for the 2007 Hispanic Healthcare Survey [17]. Overall, this study had a response rate of 46.3 % [3].

For this secondary analysis, subjects were included if they were Hispanic immigrants who were born in Mexico, the Caribbean, Central America, or South America and were diagnosed with diabetes by a physician or healthcare provider. Only 527 individuals met these eligibility criteria and were included in analyses.

Characteristics of Interest

A variety of variables were hypothesized to affect the relationship between the outcome of interest (health care utilization, as measured by “time since last healthcare visit”) and the risk factor of interest (“documentation status”). The following variables were included in the secondary analysis, as potential confounders: gender, place of origin, perceived health status, having a usual source of healthcare, place of usual care, quality of healthcare received, age, marital status and education status. All variables were categorical, except for age which was a continuous variable.

Data Management and Coding

The outcome variable was determined by the following question, “About how long has it been since you last saw a doctor or another health care provider about your health? Has it been…?” The answers in the original dataset included: “6 months or less”, “more than 6 months, but more than 1 year ago”, “more than 1 year ago, but not more than 3 years ago, more than 3 years”, “never”, “don’t know” and “refused”. These answer choices were re-coded into a dichotomous variable; the two categories were relabeled “6 months ago or less” or “more than 6 months ago, including never”.

The risk factor was documentation status and was coded as either “documented immigrant” or “undocumented immigrant”. For the purposes of this study, an undocumented immigrant was defined as an individual who was neither a citizen nor legal permanent resident of the US. This was determined through a few different questions. Firstly, it was determined if the respondent was born outside of the US, then it was determined whether the respondent was a naturalized citizen or legal resident. If a respondent responded that he/she was foreign born, but neither a (naturalized) citizen nor legal resident then he/she was categorized as an “undocumented immigrant.” If the respondent replied that he/she was born outside of the US and either a (naturalized) citizen or a legal resident, then he/she was categorized as a “documented immigrant”.

Potential confounders were also categorized. Gender was categorized into “male” or “female.” Place of origin was coded as “Mexico” or “other country of origin”, due to sample size needed for analysis. Perceived health status was coded as “excellent, very good or good” versus “fair or poor”. Having a usual place of healthcare was coded as “having at least one place of usual care” or “no usual place of care”. Place of usual care was coded as “community clinic or health center”, “doctor’s office”, “hospital” or “other place”. Age was treated as a continuous variable, ranging from 18 to 97. Marital status was divided into the categories “married or have a partner” or “widowed, divorced, separated or never been married”. Education status was coded as “less than high school education”, “high school diploma or GED”, “some college, including technical/vocational training”, and “college graduate and higher”.

Statistical Analysis

Data analyses were conducted using SAS/STAT® software (version 9.2, SAS Institute Inc, Cary, NC). Simple descriptive characteristics were calculated. A Chi square test was used to determine where differences between documented and undocumented Hispanic immigrants existed. To estimate the association between documentation status and healthcare utilization, logistic regression methods were employed using SAS/STAT procedure SURVEYLOGISTIC. Potential confounders, first identified in the literature and then confirmed in bivariate analyses, were included in the final model. All analyses were performed using weights to allow for generalizability to the greater US population. These weights were already included in the dataset obtained from the Hispanic Healthcare Survey.

Results

Descriptive Statistics of the Sample Population

Five hundred and twenty-seven Hispanic immigrants from Mexico, the Caribbean, Central America, or South America reported having been diagnosed with diabetes. Approximately half were male and half were female. The mean age of the respondents was 51.7 years (SE = 16.7). The sample consisted of 75.6 % individuals of Mexican origin and 24.4 % of another country of origin. Of these respondents, 80.1 % were documented Hispanic immigrants and 19.9 % were undocumented Hispanic immigrants. In terms of education, 51.7 % had less than a high school education, 26.1 % had a high school diploma or GED, 16.0 % had some college education, and 6.2 % had a college degree or higher (Table 1).
Table 1

Demographic characteristics

Weighted characteristic

Weighted N

Percent (%)

Gender

  

 Male

235.7

49.9

 Female

236.6

50.1

Country of origin

  

 Mexican origin

356.9

75.6

 Other country of origin

115.5

24.4

Current health status

  

 Excellent, very good, or good health

213.6

45.8

 Fair or poor health

254.8

54.2

Usual source of care

  

 At least one place of usual care

418.2

88.5

 No usual place of care

54.2

11.5

Place of usual care

  

 Community clinic or health center

177.6

42.6

 Doctors office

183.8

44.1

 Hospital or other place

55.7

13.3

Quality of health care received

  

 Excellent or very good quality of care

359.6

76.5

 Fair or poor quality of care

94.3

20.1

 No care in the last year

16.1

3.4

Time since last healthcare visit

  

 Visited a healthcare provider in the last 6 months

383.0

81.1

 Visited a healthcare provider 6 months ago or more

89.4

18.9

Documentation status

  

 Documented

245.5

80.1

 Undocumented

61.2

19.9

Highest level of education

  

 Less than high school

238.0

51.7

 High school diploma or GED

120.2

26.1

 Some college

73.5

16.0

 College graduate or higher

28.5

6.2

Health Insurance Status

  

 Has health insurance

335.1

72.0

 Has no health insurance

130.5

28.0

Marital Status

  

 Married or have a partner

286.8

61.3

 Widowed, divorced, separated or never been married

181.0

38.7

 

Mean

Standard error

Age

51.7

16.7

Differences between Documented and Undocumented Hispanic Immigrants with Diabetes

A Chi-square test was performed to determine if documented and undocumented Hispanic immigrants with diabetes were distributed differently across the variables of interest. As Table 2 shows, this test indicated that significant differences in the distribution of documented and undocumented Hispanic immigrants with diabetes existed for the following variables of interest: time since last healthcare visit, country of origin, education status, usual source of healthcare, place of usual care, quality of healthcare received, health insurance status, and marital status. A t test was performed to see if differences existed between documented and undocumented immigrants in relation to age. However, no statistically significant difference was found between the mean age of documented and undocumented Hispanic immigrants with diabetes.
Table 2

Differences between documented and undocumented diabetic hispanic immigrants

Variables

Documented (weighted %)

Undocumented (weighted %)

P value

Time since last healthcare visit

  

<0.0001

 6 months or less

69.2

12.4

 

 More than 6 months ago or never

10.9

7.5

 

Gender

  

0.6460

 Male

39.3

10.6

 

 Female

40.8

9.3

 

Country of origin

  

0.0018

 Mexico

51.4

18.1

 

 Other country of origin

25.6

1.9

 

Education

  

0.0117

 Less than high school

44.0

14.3

 

 High School or GED

21.7

4.3

 

 Some college

21.7

0.9

 

 College graduate or higher

3.6

0.3

 

Current health status

  

0.3837

 Excellent, very good, or good

28.7

8.7

 

 Fair or Poor

51.3

11.3

 

 Usual source of care

  

<0.0001

  At least once place of usual care

75.5

14.2

 

  No usual place of care

4.6

5.7

 

Location of usual source of care

  

<0.0001

 Community clinic or health center

36.7

12.7

 

 Doctors office

35.3

1.8

 

 Hospital or other place

12.2

1.3

 

Quality of health care received

  

<0.0001

 Excellent or very good

60.0

15.5

 

 Fair or poor

18.3

2.2

 

 None received in the last year

1.6

2.4

 

Health insurance

  

<0.0001

 Has health insurance

61.9

9.0

 

 No health insurance

18.5

10.7

 

Marital status

   

 Married or has a partner

56.3

11.6

0.0015

 Widowed, divorced, separated, or never married

23.7

8.4

 
 

Weighted mean (standard error)

P value

Age

52.7 (0.9)

40.9 (1.8)

0.7225

Weighted Multivariate Analyses

The results of the weighted multivariate analyses controlling for potential confounders are shown in Table 3, using documented Hispanic immigrants with diabetes as the reference group. Even after taking confounding factors into account, undocumented Hispanic immigrants with diabetes faced higher odds of having seen a healthcare provider more than 6 months ago, or never, when compared to documented Hispanic immigrants with diabetes (adjusted OR = 1.79; 95 % CI 1.01, 3.18). Undocumented immigrants were also more likely to have no usual place of care (adjusted OR = 5.49; 95 % CI 1.57, 19.16) and lack health insurance (adjusted OR = 3.05; 95 % CI 1.81, 5.17), when compared to documented immigrants. They were however, less likely to consider their health to be “excellent, very good, or good” (adjusted OR = 0.38; 95 % CI 0.18, 0.80) and less likely to be female (adjusted OR = 0.48; 95 % CI 0.27, 0.84), when compared to documented immigrants. The main determinants of healthcare utilization among Hispanic immigrants with diabetes are gender, current health status, usual source of care, and health insurance.
Table 3

Weighted multivariate analysis

Variables

Adjusted odds ratio (95 % CI)

Documentation status

 

 Documented

Reference group

 Undocumented

1.79 (1.01, 3.18)

Gender

 

 Male

Reference group

 Female

0.48 (0.27, 0.84)

Current health status

 

 Excellent, very good, or good

Reference group

 Fair or poor

0.38 (0.18, 0.80)

Usual source of care

 

 At least once place of usual care

Reference group

 No usual place of care

5.49 (1.57, 19.16)

Health insurance

 

 Has health insurance

Reference group

 No health insurance

3.05 (1.81, 5.17)

Discussion

The findings of this study agree with that found in the current literature investigating healthcare utilization within Hispanic immigrants, but also adds to it by examining the effect of documentation status. This study has demonstrated that documentation status, gender, having a usual place of care, and having health insurance have an effect on a Hispanic immigrant with diabetes’ choice to see a healthcare provider more than 6 months ago.

To make sense of the role that documentation status plays on healthcare utilization by Hispanic immigrants with diabetes, we have to acknowledge the differences between these two groups. Although a majority of both documented and undocumented Hispanic immigrants with diabetes in this study indicated their health to be “fair or poor,” documented Hispanic immigrants with diabetes were more likely to have seen a healthcare provider in the last 6 months, when compared to undocumented Hispanic immigrants with diabetes. Reasons for this appear to be related to healthcare access. As hypothesized, undocumented Hispanic immigrants with diabetes lacking a usual place of care and/or health insurance were less likely to have seen a healthcare provider in the last 6 months.

Regarding gender, females were more likely to have seen a healthcare provider in the last 6 months, when compared to males. This is true of diabetes, but also true of women in general, according to current literature. Owens [18] found that women tend to use significantly more services and spend more health care dollars than men. The greatest cost disparity in healthcare service and expenditures between men and women exists in the population aged 45–64 years. Given the risk of diabetes generally increases with age, and that the mean age of Hispanic immigrants with diabetes in this study was 51.7 years, this age group may be a good one to observe in future studies.

All of these findings have public health practice implications tied to economic costs. Previous studies have demonstrated that health care costs can escalate due to a lack of proper management of chronic medical conditions, such as diabetes. Complications with diabetes require greater medical attention, at a greater cost. One study examining glycemic control across different races/ethnicities indicated that when compared to non-Hispanic whites, Hispanics are more likely to be pharmacologically treated and have higher rates of diabetes complications such as end-stage renal disease, retinopathy, blindness, and neuropathy [19]. Many of the diabetes complications were attributed to problems with glycemic control, which was associated with many of the same characteristics examined in this study: namely, a lack of usual care among the Hispanic population with diabetes, a lack of health insurance, and perceived healthcare status.

Since undocumented Hispanic immigrants with diabetes are unlikely to have seen their healthcare provider in the last 6 months, due to a lack of health care and/or a usual place of care, we can assume that at least some members of this population might not be able to manage their diabetes properly. These individuals are at risk of developing diabetes-related complications, which translates to higher healthcare costs. Often, when the cost becomes too high, the burden is placed onto publicly funded health services [20].

Although we are uncertain exactly what percentage of the nation’s spending on direct medical costs can be attributed to this problem, we do know that the Centers for Disease Control and Prevention has previously reported that the treatment of chronic conditions costs $1.5 trillion a year, or 75 % of the nation’s spending on direct medical costs [21].

Limitations

Like many other studies, this study has its limitations. First, since this study uses cross-sectional data to conduct a secondary analysis, conclusions cannot be reached regarding causation. Similarly, the directionality of the association between the variables cannot be determined. Despite this particular limitation, this study does investigate associations that have not been considered in previously published literature and its ideas can be expanded on in future research.

Secondly, since health care utilization was measured using a single item from the Pew Hispanic Center’s questionnaire, it cannot reflect other dimensions of health care utilization, such as the type of healthcare visit and/or the reason for the healthcare visit. Knowing this information would be useful in future analysis, particularly if researchers are interested in looking at associations between documentation status and healthcare utilization of Hispanic immigrants with diabetes more closely. Furthermore, it is possible that other variables not examined may have additional explanatory value beyond those included in this particular study. Nonetheless, the use of this single item still provided us with important information regarding the relationship between documentation status and healthcare utilization. Examining healthcare utilization from the scope of documentation status is novel, so the amount of information that can be obtained from current survey instruments is limited to what is available.

A third limitation relates to the response rate of the study. Given that the response rate was 46.3 %, it is possible that prevalence estimates may be subject to sampling bias. One source of potential bias could be in the selection of participants. Specifically, self-selection bias may have occurred during data collection, whereby “documented” Hispanic immigrants were more inclined to participate in the study and answer survey questions than “undocumented” Hispanic immigrants—possibly due to fears of legal authorities. To illustrate this point: 80 % of respondents within this study were “documented” Hispanic immigrants with diabetes and 20 % were “undocumented” Hispanic immigrants with diabetes. The reason why this is an important limitation is that it has the potential to affect the accuracy and/or reliability of the responses received. To mitigate this, inclusion and exclusion criteria were applied and respondents were allocated to each group accordingly. However, it is possible that some of the respondents chose to answer differently than is true (i.e., replied that they were citizens when they actually were not), which might have resulted in an undercount of “undocumented” Hispanic immigrants with diabetes. Thus, our findings might underestimate the associations found for “undocumented” Hispanic immigrants with diabetes.

Conclusion

Unique opportunities in addressing healthcare disparities can be found in focusing on the Hispanic immigrant population living with diabetes. If we want to search for strategies to reduce health care expenditures, more attention should be focused on addressing healthcare disparities. As stated earlier, documentation status, gender, having a usual place of care, and having health insurance have an effect the decision to see a healthcare provider by Hispanic immigrants with diabetes. Although we know from this study that healthcare utilization is not the same across documentation status, we are unable to identify causal pathways for this phenomenon due to the use of cross-sectional data. Thus, to further understand what other factors may be influencing the decision to see a healthcare provider, studies that involve primary data collection and qualitative research methods should be conducted.

References

  1. 1.
    Doty M. Hispanic patients’ double burden: lack of health insurance and limited English. The Commonwealth Fund; 2003.Google Scholar
  2. 2.
    Passel JS, Cohn D. U.S. population projections: 2005–2050. Washington: PewResearchCenter; 2008, Available at: http://www.pewhispanic.org/files/reports/85.pdf.
  3. 3.
    Livingston G, Minushkin S, Cohn D. Hispanics and health care in the United States: access, information and knowledge. Washington: Pew Hispanic Center/Robert Wood Johnson Foundation; 2008. Available at: http://pewhispanic.org/files/reports/91.pdf.
  4. 4.
    Bonds DE, Zaccaro DJ, Karter AJ, et al. Ethnic and racial differences in diabetes care. Diabetes Care. 2003;26:1040–6.PubMedCrossRefGoogle Scholar
  5. 5.
    Chow EA, Foster H, Gonzalez V, McIver L. The disparate impact of diabetes on racial/ethnic minority populations. Clin Diabetes. 2012;30(3):130–3.CrossRefGoogle Scholar
  6. 6.
    Fernandez A, et al. Language barriers, physician-patient language concordance, and glycemic control among insured Latinos with diabetes: the Diabetes Study of Northern California (DISTANCE). J Gen Intern Med. 2010;26:170–6.PubMedCentralPubMedCrossRefGoogle Scholar
  7. 7.
    Kirk JK, Passmore LV, Bell RA, et al. Disparities in A1C Levels between Hispanic and Non-Hispanic White adults with diabetes. Diabetes Care. 2008;31(2):2240–6.Google Scholar
  8. 8.
    Department of Health and Human Services. Fact sheet: prevalence of diabetes among Hispanics in six U.S. geographic locations. Atlanta: Centers for Disease Control and Prevention; 2011. Available at: http://www.cdc.gov/diabetes/pubs/pdf/hispanic.pdf.
  9. 9.
    Britigan D, Murnam J, Rojas-Guyler L. A qualitative study examining Latino functional health literacy levels and sources of health information. J Commun Health. 2009;34:222–30.CrossRefGoogle Scholar
  10. 10.
    Ascheim J, Bagley B, Doran R, et al. Plan to address health disparities and promote health equity in New Hampshire. New Hampshire: State Plan Advisory Work Group; 2011. Available at: http://www.dhhs.nh.gov/omh/documents/disparities.pdf.
  11. 11.
    Cistancho S, et al. Listening to rural Hispanic immigrants in the Midwest: a community-based participatory assessment of major barriers to health care access and use. Qual Health Res. 2008;19:633–45.CrossRefGoogle Scholar
  12. 12.
    DuBard A, Gizlice Z. Language spoken and differences in health status, access to care, and receipt of preventive services among US Hispanics. Am J Public Health. 2008;98:2021–8.PubMedCentralPubMedCrossRefGoogle Scholar
  13. 13.
    DeVoe JE, Baez A, Angier H, Krois L, et al. Insurance + access ≠ health care: typology of barriers to health care access for low-income families. Ann Fam Med. 2007;5(6):522–8.CrossRefGoogle Scholar
  14. 14.
    Mainous AG, Majeed A, Koopman RJ, Baker R, et al. Acculturation and diabetes among Hispanics: evidence from 1999–2002 National Health and Nutrition Examination Survey. Public Health Rep. 2006;121:60–6.PubMedCentralPubMedGoogle Scholar
  15. 15.
    Derose KP, Escarce JJ, Lurie N. Immigrants and health care: sources of vulnerability. Health Aff. 2007;26(5):1258–68.CrossRefGoogle Scholar
  16. 16.
    Pew Hispanic Center/Robert Wood Johnson Foundation. 2007 Hispanic Healthcare Survey. Washington: Pew Hispanic Center/Robert Wood Johnson Foundation; 2007. Available at: http://www.pewhispanic.org/2007/09/23/2007-hispanic-healthcare-survey/.
  17. 17.
    International Communications Research Group. Pew Hispanic Center Health Survey Topline. Media: International Communications Research; 2007. Available at: http://www.pewhispanic.org/files/2008/08/91.1.pdf.
  18. 18.
    Owens G. Gender differences in health care expenditures, resource utilization, and quality of care. J Manag Care Pharm. 2008;14(3):S2–6.Google Scholar
  19. 19.
    Diaz-Apodaca B, de Cosio FG, Canela-Soler J, et al. Quality of diabetes care: a cross-sectional study of adults of Hispanic origin across and along the United States-Mexico border. Rev Panam Salud Publica. 2010;28(3):207–13.PubMedCrossRefGoogle Scholar
  20. 20.
    Correa-de-Araujo R, McDermott K, Moy E. Gender differences across racial and ethnic groups in the quality of care for diabetes. Women’s Health Issues. 2006;16:56–65.PubMedCrossRefGoogle Scholar
  21. 21.
    Thrall JH. Prevalence and costs of chronic disease in a health care system structured for treatment of acute illness. Radiology. 2005;235:9–12.PubMedCrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media New York 2012

Authors and Affiliations

  1. 1.Department of Social and Behavioral HealthVirginia Commonwealth UniversityRichmondUSA

Personalised recommendations