Abstract
Background: Depression affects more Hispanics with type 2 diabetes than other ethnic groups. This exploratory, binational study examined the prevalence and correlates of clinical depressive symptoms in Hispanics of Mexican origin with type 2 diabetes living on both sides of the Texas Mexico border.
Methods: Two binational samples, consisting of 172 adult patients of Mexican origin with type 2 diabetes in South Texas and 200 from the Northeastern region of Mexico, were compared. Logistic regression analyses were used to test personal and social correlates to clinical depressive symptoms.
Results: The rate of clinical depressive symptoms was similar in both South Texas and Northeastern Mexico patients (39% and 40.5%, respectively). Gender, education, emergency department visits, and burden of diabetes symptoms were predictors of clinical depressive symptoms in the South Texas sample. Among respondents in the Northeastern Mexico sample, the only statistically significant correlate to clinical depressive symptoms was the burden of diabetes symptoms.
Conclusions: Diabetes and depression must be addressed as priorities in diabetes initiatives at the US Mexico border region. Further research is warranted to examine the extent and impact of involving family practice physicians from both sides of the border in depression screenings among patients with type 2 diabetes.
Depression in patients with type 2 diabetes is associated with increased mortality rates, poor glycemic control, increased diabetes complications, increased functional disability, poorer compliance with treatment recommendations, and higher medical costs.1–4 A recent meta-analysis5 indicates that major depression and elevated depression symptoms are present, respectively, in 11% and 31% of people with diabetes in the United States. Moreover, depression affects more Hispanics with type 2 diabetes than blacks and non-Hispanic whites with the same condition.6–8
Hispanics are the largest and fastest growing minority population in the United States9 and are disproportionately affected by type 2 diabetes and its related complications and comorbidities.10,11 The prevalence of diabetes in this minority group is 2 times higher when compared with whites.11,12 Hispanics with type 2 diabetes are also 2 times more likely to develop diabetes-related complications than non-Hispanic whites.13–16 The burden of diabetes in Hispanics is even more salient in the United States Mexico border area. Almost 16% of border residents suffer from type 2 diabetes, a higher rate than the national rates both in Mexico (7.5%) and the United States (13.9%).17 Along the border, the diabetes death rate for Hispanics living in US counties (46.7 age-adjusted per 100,000 population) is 3 times the rate for non-Hispanic whites (16.3 age-adjusted per 100,000 population).18 On the Mexican side of the border, the diabetes prevalence rate is higher than the national rate in Mexico (8.4% and 7.5%, respectively).19,20
Although depression is 2 times more prevalent in diabetic patients than the general population,2,5,21 the reasons for higher prevalence rates among diabetics are not fully understood. The cause of depression in patients with diabetes is difficult to identify because this condition may have biological roots, may be associated to the burden of complications, or a combination of both.22 A biopsychosocial approach proposed by Fisher et al23 suggests that multiple disease, personal, social, and community factors interact to affect depression in people with diabetes. Previous research about Hispanics with type 2 diabetes in the United Sates found that correlates to depression include age,8 education,8,23 financial stress,23 poorer self-rated health,24 diabetes-related functional impact,23 understanding and knowledge of diabetes,24 lifestyle behaviors,25 and poor glycemic control.26
Research examining depression and diabetes in patients along the United States Mexico border is warranted. High diabetes prevalence rates and related comorbidities in the border region not only affect patients’ quality of life but also result in high economic costs for both the US and Mexico health care systems and communities.19,27 Although Healthy Border 2010 identifies diabetes and mental health issues as border priorities,28 the prevalence of depression in diabetics in the border region is unknown. The purpose of this study was to determine the levels and correlates of clinical depressive symptoms in Hispanic adult patients of Mexican origin with type 2 diabetes residing on both sides of the Texas Mexico border.
Methods
This was a cross-sectional study based on a survey conducted in 2004 to 2005 with 2 samples in the Lower Rio Grande Valley in South Texas and Reynosa in the Northern Tamaulipas region of Mexico. Using a convenience sampling technique, this study recruited people with type 2 diabetes from clinical settings on both sides of the border (hospitals and physicians’ offices). Inclusion criteria included age of 18 years or older; diagnosis of type 2 diabetes for at least 1 year; and willingness to give informed consent. A total of 199 patients in South Texas and 200 in Northern Tamaulipas agreed to participate in the study. The recruitment sites included hospitals and physicians’ offices. Staff from these facilities identified potential participants and asked them if they would be interested in participating in this study. The research team contacted these potential participants. In the case of the South Texas sample, participants were interviewed at a location of their convenience, usually at the physician's office or Texas A&M headquarters. Participants in Northern Tamaulipas were interviewed at the hospital. In the analysis we only included participants of Mexican origin with complete data, resulting in a sample size of 172 from South Texas and 200 from Northern Tamaulipas. This study was approved by the Texas A&M University Institutional Review Board.
Measures
Depressive symptoms were assessed using the Center for Epidemiologic Studies Depression Scale. This scale is a 20-item self-report depression inventory with possible scores ranging from 0 to 60. Reliability and validity of the scale have been tested in general and clinical populations, yielding very good internal consistency, with an α of 0.85 for the general population and 0.90 for a patient population. Scale scores range from 0 to 60. A cutoff point of 16 and above indicates clinically significant levels of depressive symptoms.29 Although the scale does not provide a diagnosis of clinical depression, it has been shown to predict both current and future clinical depression.30 The Center for Epidemiologic Studies Depression Scale has been validated in patients with diabetes,31 and the Spanish version has high validity and reliability (Cronbach's α, 0.90).32,33
Independent variables included demographic characteristics, health factors, health care use factors, and level of family support. Demographic variables included gender; age; marital status; country of origin (Mexico or United States, only applied to the South Texas sample); education level; socioeconomic status; and acculturation (only applied to the South Texas sample). Socioeconomic status was assessed by asking participants if they were employed or not; employment is an indicator of socioeconomic position.34,35 Household income was not included because of missing values. Acculturation was measured using the Short Acculturation Scale for Hispanics, a 12-item scale that has been validated in Spanish and has an internal reliability of α = 0.92.36 Health factors were measured from self-report information and included the following: self-rated health status; age when diagnosed with diabetes; duration of diabetes; high blood pressure, heart disease, high cholesterol, insulin use, amputations, smoking, vision problems related to reading, driving, and watching TV; sexual function problems; leg discomfort limiting activities around the house, wearing wanted shoes, and sleeping; limited social life because of the need to follow a special diet; hypoglycemic symptoms; diabetes symptoms; patient's beliefs (including “sugar monitoring is painful,” “taking medication or insulin is a waste of time,” and “blood glucose level testing is difficult”). Body mass index was not included in the analyses because of missing values.
The burden of diabetes symptoms variable was measured using the Patient's Questionnaire Diabetes Form 2.1.37 This variable was calculated by summing 15 diabetes symptoms (from 0 to 15).
Use of health care factors included number of emergency department visits, doctor visits, hospital stays, diet counseling sessions, and diabetes education sessions during the past 12 months. Glucose level check was assessed asking the question, How many times do you check your blood sugar in an average week? (daily versus not daily).
Family support was measured using the Family Behavior Checklist, which assesses actions of a relative identified by the participant as the supportive person in their efforts to manage their diabetes. Scale items include the relative's supportive behaviors related to medication, glucose testing, exercise, diet, and in general. A positive summary score (high vs low family support) was obtained by averaging the frequency ratings over all 5 supportive items. Reliability and predictive validity of this scale is between 0.64 and 0.84. The scale also asks the participant to rate the diabetes-related knowledge level of the supportive relative (low versus moderate versus high).38 The scale was translated into Spanish using a back-translation technique.
Data Analyses
Summary statistics were calculated to describe each sample in terms of sociodemographic and health-related variables. Demographic characteristics and health indicators between samples were compared using the Mann-Whitney U test. Logistic regression analyses (unadjusted univariate and adjusted multivariate logistic regression) were used to test the association between the dependent variable (clinical depressive symptoms) and independent variables. The self-rated health status variable was not included in the logistic regression analyses because it was highly correlated with diabetes symptoms (P < .01) in both samples. Including the self-rated health status variable in the multivariate regression may have distorted the results because of the colinearity problem in the regression fitting. Values of odds ratios, 95% CI, and P were reported in the logistic regression analyses. The Hosmer-Lemeshow goodness-of-fit test was conducted to test model fit. P < .05 was considered significant for all statistical tests conducted. The analyses were performed using the computer program SPSS version 13.0 for Windows (SPSS, Inc., Chicago, IL).39
Results
Table 1 shows the distribution of demographic characteristics and health measures for both samples and between-group differences. The majority of participants on both sides of the border were women, married, and had a low socioeconomic status. Almost half of the respondents in South Texas (47.7%) were born in Mexico and the majority (79%) had low levels of acculturation. Respondents in South Texas had a mean age of 57.8 (SD, 14.1) and the Northern Tamaulipas participants of 55.8 (SD, 12.2). Significantly more Northern Tamaulipas respondents were women and had less than a high school education than the South Texas participants.
The overall rate of clinical depressive symptoms was similar for both South Texas and Northern Tamaulipas patients (39% and 40.5%, respectively). There were no significant differences in clinical depressive symptoms between the samples. The majority of patients on both sides of the border reported having hypertension and high cholesterol. The majority in both samples also reported to perceive a high level of family support. Although the main source of family support for the South Texas participants was the spouse, for patients in Northern Tamaulipas it was both their spouse and children.
Compared with respondents in Northern Tamaulipas, significantly more South Texas participants reported better self-rate health status, but they also reported having high cholesterol. Conversely, significantly more patients in Northern Tamaulipas reported fewer physician visits in the past 12 months than did their South Texas counterparts.
Table 2 shows the distribution of diabetes-related characteristics for both samples and between-group differences. The majority of patients on both sides of the border reported being diagnosed with diabetes for less than 10 years and had not received any diabetes education during the past 12 months. Significantly more South Texas than Northern Tamaulipas respondents used insulin and believed that testing glucose level was painful and difficult. Conversely, significantly more patients in Northern Tamaulipas reported less frequent glucose level checks and a higher number of diabetes-related symptoms than their South Texas counterparts. There was also a significant difference between groups in relation to diet counseling, with more Northern Tamaulipas respondents than South Texas patients receiving 2 or more diet sessions during the past 12 months.
Table 3 shows the univariate (unadjusted) logistic regression analyses for each sample. Only variables that were statistically significant are listed in the table. Analyzed singly, the following variables had a statistically significant association with depression in both the South Texas and Northern Tamaulipas samples: gender; emergency department visits; hospital stays; diabetes symptoms (including leg discomfort interfering with sleep, home activities, and wearing preferred shoes); and limited social life because of having a special diet. Among South Texas participants, education level, high cholesterol, visits to the doctor more than once a month, vision problems that interfere with television watching, perception that glucose testing was difficult, and family support also had a statistically significant association with clinical depressive symptoms. Among Northern Tamaulipas participants gender; socioeconomic level; heart disease; vision problems that interfere with reading, driving, and watching television; and perception that glucose testing was difficult were statistically significantly correlated to clinical depressive symptoms.
Table 4 presents the results from the multivariate (adjusted) logistic regression analyses. Only significant associations are presented by sample. Among South Texas respondents, female gender, a low level of education, one or more emergency department visits during the past 12 months, burden of diabetes symptoms, having leg discomfort that interferes with wearing preferred shoes, and the perception that glucose testing was difficult were significantly correlated to clinical depressive symptoms. However, those reporting high cholesterol were less likely to have clinical depressive symptoms.
Among respondents in the Northern Tamaulipas sample the 2 items significantly associated with clinical depressive symptoms in the multivariate logistic model were the burden of diabetes symptoms and leg discomfort interfering with home activities. The estimated odds of having clinical depressive symptoms increased for each additional diabetes symptom by 27% for South Texas participants and 36% for the Northern Tamaulipas respondents.
Discussion
Our findings indicate that the prevalence of clinical depressive symptoms in both South Texas and Northern Mexico border residents with type 2 diabetes (39% and 40.5%, respectively) is higher than what is reported in the literature of clinical diabetic samples in the United States (32%).5
Specifically, the prevalence rate of clinical depressive symptoms among South Texas respondents in our study was also higher than those reported in studies conducted with older Mexican Americans in Southwestern states (25.6% to 31.1%)40,41 or Hispanic adult patients (31.6%),6,23 but similar to a study with Hispanic women in South Florida (40.6%).24 In the case of our border sample from the Mexican side, the clinical depressive symptoms prevalence rate found was similar to previous studies conducted in Central Mexico (37.91% to 39%)25,42 but inconsistent with another study from Mexico City (46%).43
Our study found that the strongest predictor of clinical depressive symptoms in type 2 diabetes patients on both sides of the Texas Mexico border was the burden of diabetes symptoms. Among respondents in Mexico this was the only predictor of depression. Our findings are in line with previous research in the United States.44–47 Our study, however, did not establish causation. Further research is needed to identify a causal pathway to determine how depression and the burden of diabetes symptoms interact. In addition, future research should investigate if identifying and treating depression in diabetic patients may result in a decreased burden of diabetes symptoms.
Additional significant correlates to clinical depressive symptoms in the South Texas sample were female gender and low levels of education, which is consistent with previous studies.8,23,42,43,48–50 Research suggests that low access to education affects health literacy and English proficiency in Hispanic patients with type 2, leading to distress and poorer diabetes self-care.8,51
Emergency department visits were also associated with higher levels of depressive symptoms in US respondents. The finding agrees with previous research52–54 and underlines the importance of recognizing depression as a risk factor for increasing health care expenditures.
Our study shows that the South Texas respondents who perceived that blood glucose testing was difficult were more likely to report clinical depressive symptoms, which is consistent with previous research.55 Diabetes interventions in the border region focusing on self-efficacy and the improvement of skills in self-monitoring of blood glucose may result in positive diabetes-related outcomes.56–58
An intriguing finding in our study was that the South Texas participants who reported having high cholesterol were less likely to have clinical depressive symptoms. Although this study did not examine medication intake by respondents, it is possible that this finding might be related to the type of cholesterol medications taken by South Texas patients. Research studying the association between statins therapy and depression is still controversial,59–61 but some studies indicate that cholesterol-lowering drugs are associated with a reduced risk of depression.62,63 Further research is warranted to better understand the effects of statins-based medications on depression among diabetic patients in both the US and Mexican health care systems. Additional research should also investigate if hyperlipidemia should be considered an early marker of depression.
This exploratory study has several limitations. The convenience sample of diagnosed patients with type 2 diabetes limits generalizations and thereby causal inferences cannot be made. In addition, both dependent and independent variables were measured using a self-report instrument, which carries possible intrinsic respondent biases and measurement errors. Another limitation is that this study did not explore if participants were taking antidepressants, which may have introduced respondent biases.
Despite its limitations, this preliminary study makes a significant contribution to the literature by assessing the prevalence and correlates of clinical depressive symptoms in Hispanic patients with type 2 diabetes living on both sides of the Texas Mexico border. To the best of our knowledge this is the first binational study addressing depression and diabetes among border residents. Our findings may have health policy implications. Healthy Border 2010 identifies diabetes and mental health issues as border health priorities. Screening diabetics for depression is of great importance because there is evidence that treatment of depression improves glycemic control64 and increases a patient's quality of life.65 Healthy Border 2010, however, recognizes that access to mental health services along the US Mexico border is problematic because of a shortage of specialty mental health providers and services.28 In addition, differences between both the US and the Mexican health care systems and practices pose barriers to meaningful cooperative health programs that, if successful, could contribute to the health improvement of the border population.66 Further research is warranted to examine the feasibility and impact of binational programs involving family practice physicians on both sides of the border in depression screenings among patients with type 2 diabetes.
Acknowledgments
The authors wish to extend their gratitude to Maria Alen, MD, from TAM Health Science Center; Diana Garcia, BS, and Pama Ellis, RN, from the Diabetes Management Center Rio Grande Regional Hospital; Josefa Lopez, MD, and Carolina Rivera, RN, from Hospital General de Reynosa; Marcel Twahira, MD; Juan Campos, MD; Grace Lawson from El Milagro Clinic, for their assistance and insightful input during the design and implementation of this study.
Notes
This article was externally peer reviewed.
Funding: This research was supported by the Texas Department of State Health Services and by Health Services Research Program (a collaborative research venture involving the TA&M HSC School of Rural Public Health and the Scott and White Hospital and Clinic and the Scott and White Health Plan).
Conflict of interest: none declared.
- Received for publication November 13, 2007.
- Revision received February 18, 2008.
- Accepted for publication February 22, 2008.