Spanish Diabetes Self-Management Program

Diabetes self-management education is increasingly recognized in best practice guidelines and reimbursement policies as a critical ingredient of appropriate health services, especially for the Latino population. Latinos, in addition to carrying a heavy burden of diabetes and related comorbid conditions, encounter limited access to health care and health education due to economic, language and cultural barriers. The proposed project is a health status and health care utilization outcome study. It is unique in that: 1) it will evaluate a low-cost replicable, community-based, diabetes education model, 2) it will study the long-term effects of a low-cost, easily replicable reinforcement system, ATDM calls, and 3) it targets Spanish-speakers, who are at high risk for type 2 diabetes, for diabetes-related complications, and who are under-served by existing programs.

The background for this study is derived from our long-standing programs of research on chronic disease self-management and automated telephone disease management (ATDM) as well as the research of others. This research has demonstrated that: 1) Self-management programs and ATDM programs can improve health-related quality of life, metabolic control, and health care utilization. 2) The effects of these programs on health-related quality of life and health care utilization are mediated partially by enhancement of perceived self-efficacy to cope with the consequences of chronic disease. 3) Self-management education has a potential for achieving substantial financial savings through decreased use of inpatient and outpatient health care.

Information about the outcomes of health self-management programs for Latinos is more fragmentary. However, analyses of our Spanish language Arthritis Self-Management Program and analysis of a pilot of our Spanish Diabetes Self-Management Program suggest that self-management programs can be extended successfully to Latinos. Work by Piette has demonstrated that 1) monolingual Spanish speakers can and will use automated telephone disease management (ATDM) calls as part of their care, 2) Latinos will use ATDM calls to access health education, and 3) ATDM-supported diabetes care can improve health-related quality of life, metabolic control and self-management behaviors.

BACKGROUND AND SIGNIFICANCE

Six factors contribute to the significance of the proposed research. These include 1) the growing population of Latinos in the United States, 2) the disproportional burden of diabetes in the Latino population, 3) the costs of diabetes care coupled with decreasing access to care for the uninsured and underinsured, 4) the lack of well developed and evaluated diabetes education interventions for Latinos, 5) growing evidence that psychosocial interventions for diabetes can impact both health status and health care costs, and 6) evidence that perceived self-efficacy is an important mediator of changes in health related quality of life.

Growing Latino Population

Approximately 22 million Latinos representing 9% of the total population are currently living in the United States1 and Latinos are expected to be the largest minority group early in this century.2 One third of the nation's Latino population lives in California,3 and 60% of this 7.7 million are foreign born.4, 5 By 2007 slightly more than one third of California's population is expected to be Latino.5

Latinos have the lowest level of education of any demographic group in the United States. Only 45% of Latinos in California, compared to 50% nationally, have completed high school.3 Approximately 400,000 Latinos are 65 or older. Between 1983 and 1988 this group grew at a rate of nearly 27%, making it the fastest growing group for any elderly population in the United States.6 California Latinos have twice the poverty rate as whites.5

Burden of Diabetes

Certain factors endemic among Latinos, such as lower socioeconomic status, less formal education, language and cultural barriers, and reduced access to health care, all contribute to the morbidity and mortality associated with diabetes.

Morbidity

Morbidity data suggest that 14% of Latinos suffer from diabetes.7 Data from several studies indicate that the prevalence of diabetes is nearly 3 times greater among Latinos than non-Latinos.5, 7-9 It is also metabolically more severe, placing them at greater risk for medical complications such as diabetic retinopathy, neuropathy, end-stage renal disease and peripheral vascular disease.10-15 The high rate of morbidity is likely to continue in the future as Latinos have more risk factors, especially physical inactivity, unfavorable body fat distribution and obesity for women, and higher blood pressure.16, 17

Mortality National mortality statistics for Latinos, even if they are available, are not reliable. Until recently, neither national data sets nor death certificates contained Latino identifiers.4 Despite this, there have been some comparative studies of mortality rates between Latinos and non-Latino whites by specific causes. While these studies have limitations, they do suggest that mortality rates due to diabetes are higher for both Latino men and women than for non-Latinos.18 These rates are especially high among Mexican and Puerto Rican women, appearing as the fourth leading cause of death for these two groups.19 The single greatest complication of diabetes across Latinos is heart disease. While data are not available on the heart disease mortality that is secondary to diabetes, we do know that cardiovascular disease is the leading cause of death among Mexican Americans, who make up 63% of the U.S. Latino population. Within this group, Spanish speakers have the highest cardiac mortality.16 Although cardiovascular mortality is declining in this country, it is declining more slowly in the Latino population and is still the leading cause of death among older Latinos.20

Costs and Access

Over one third of the Latinos in the United States are completely uninsured,21 and the rate of growth of the uninsured Latino population is greater than that of the non-Latino population.22 While 76% of white non-elderly Californians have private health insurance, only 43% of Latinos do.21 Even when covered by insurance, Latinos have lower outpatient utilization rates than whites, especially those who are not English-speaking.2 However, twice as many Latinos report using emergency room services for primary care when compared to whites.23 Latinos also enter hospitals via emergency rooms more than whites, and have longer and more expensive hospital stays.24 This is especially important because hospitalization accounts for 60% of all costs associated with diabetes.25 In addition to being uninsured, language and cultural barriers also contribute to the under utilization of preventive health care among Latinos. Spanish-speaking Latinos encounter obstacles when seeking health care within the complicated health care system, and are less likely to find a regular source of medical care.2 The patterns of health care utilization described above suggest that diabetes self-management education may assist the Latino population in reducing emergency room use and days in hospital.

Diabetes Patient Education Programs for Latinos

While there is a growing epidemiological literature on the health status of Latinos, we have been able to find only two published articles on interventions for U.S. adult Latinos with diabetes. One article reported on the feasibility of a pilot, community-based, diabetes patient education and group support program for Spanish-speaking diabetics, which included the use of group discussions and videos to teach diabetes-specific management skills.26 The other study, by Piette, is discussed in section B.7.2. While there are many diabetes programs targeting Latinos, few of these appear to have been rigorously evaluated.

Evidence Suggesting the Effectiveness of Psychosocial Self-Management Programs for Diabetes

There is growing evidence that psychosocial programs are effective for people with chronic disease in general and diabetes in particular. As documented by Mazzuca in a meta-analysis, a large range of health status changes can be expected. The mean effect size of these changes is approximately .3.27 Programs have been shown to lower blood pressure;28-30 reduce mortality from hypertension,31 melanoma,32 from hematologic malignancies,33 and from breast cancer;34 and, for people with diabetes, increased health-related quality of life and metabolic control.26, 35-43

Several reviews establish the efficacy of psychosocial interventions for people with diabetes.36, 44, 45 From the literature it is clear that psychosocial interventions can help patients improve metabolic control and diabetes self-management behaviors. Some evidence also exists for improvements in health-related quality of life.

Most diabetes education studies are short term. The following are exceptions. Mulrow and colleagues conducted an intervention for low-literacy type 2 diabetics. Initial weight loss was not sustained at 11- month follow-up.46 Rubin and colleagues, however, conducted a coping skills training program that did demonstrate sustained benefits in terms of emotional well-being at one year.47 Studies such as those of Weinberger and Piette used continuing interventions.38, 48 Both of these studies demonstrated sustained improvement in terms of metabolic control and health-related quality of life.

In summary, there is evidence that diabetes educational interventions can improve metabolic control and quality of life. The need for reinforcement is less clear, although it appears that programs with continuing interventions seem to produce better long-term results than do shorter programs.

PRELIMINARY STUDIES

Overview

This proposed study is possible because of the convergence of six lines of inquiry: 1) outcome studies that have demonstrated the influence of chronic disease self-management in improving the health status of both English- and Spanish-speaking patients, 2) outcome studies which have demonstrated the effectiveness of an automated telephone intervention in improving the health status of Latinos with diabetes, 3) a pilot study of a diabetes self-management program for Latinos, 4) methodological studies which have established effective means for recruiting and collecting data from Spanish-speaking people with chronic conditions, 5) the availability of valid self-administered instruments, which have been translated into Spanish and validated with Spanish-speaking individuals throughout the United States and abroad; and 6) theoretical studies that identify mechanisms by which self-management interventions affect health status.

The Chronic Disease Self-Management Program Study (English)

The Chronic Disease Self-Management Program study was designed to test the general applicability of self-management programs to mixed groups having different chronic diseases. In this study, 1,140 persons 40 years of age or older with at least one of the targeted diseases (heart disease, lung disease, stroke and/or arthritis) were randomized to receive a 7-week (2.5 hours per week) peer-led, community-based self-management program or to usual care control status. Treatment groups included patients with different chronic diseases. After six months, the controls received the Chronic Disease Self-Management Program. Subjects participating in the program were followed for between one and three years. Content of the English Chronic Disease Self-Management Program was based on a literature review and input from 11 focus groups.71, 72 The self-management activities focused on the skills needed to manage the tasks involved in dealing with chronic illness as described by Corbin and Strauss.73 These include: 1) treatment tasks necessitated by the illness such as exercise, symptom management, taking medication(s), and communicating with health care providers, 2) tasks needed to maintain or alter one's life roles in the face of the chronic condition, and 3) emotional tasks such as dealing with anger, depression and an uncertain future. The course development was informed by self-efficacy theory and taught using efficacy-enhancing strategies that included emphasis on skills mastery, modeling, reinterpretation of symptoms, and social persuasion.

In the English Chronic Disease Self-Management Program, participants' mean age was 65 and 65% of participants were female. Subjects had an average of 2.3 chronic diseases and 7% had diabetes. Treatment subjects, when compared with control subjects, demonstrated improvements at 6 months in weekly minutes of exercise, frequency of cognitive symptom management, communication with the physician, self-reported health, health distress, fatigue, disability, and social/role activities limitation. They also had fewer hospitalizations, and fewer days in the hospital (p<.05)74 (publication included in Appendix 1).

The whole study population, including the original intervention group and the control group, receiving the intervention after a six-month wait were followed for three years. Each year for three years compared to baseline, ER/outpatient visits and health distress were reduced (p<.05). Disability increased at the expected rate (p<.05). There were no other significant changes in any of the outcome variables. There was evidence that increased perceived self-efficacy was associated with decreased health distress, which in turn was marginally associated with future reductions in outpatient visits.

Spanish Diabetes Self-Management Pilot

As part of a project supported by the Hospital Council of Santa Clara County, we developed and evaluated a community-based, peer-led diabetes self-management program for Spanish speakers, using a three-month pre-test/post-test design. Like our previous programs, this course was based on a needs assessment conducted with four groups of diabetic Latinos. In addition we conducted three focus groups with nutritionists and diabetes educators working in the Northern California Kaiser Permanente system. The purpose of these later groups was to identify content and key messages to include in the program. The program was then reviewed by diabetes nurse educators, nutritionists, and a diabetologist.

Courses were offered by two different organizations between the summers of 1998 and 1999. Program participants completed human subject's requirements and filled out baseline questionnaires through the mail or by telephone before the first session of the SDSM pilot program. The questionnaires were then sent to Stanford, where the Patient Education Research Center staff collected missing data by telephone. Three-month post-intervention data were again collected either through the mail or by telephone. One hundred and nine of the people participating in the SCSM pilot program completed data collection.

Program participants made significant (p<.05) improvements in health behaviors (exercise, diet, practice of relaxation techniques, examining feet, communication with provider). Similarly, health status improved as measured both by health-related quality of life variables (self-reported health, fatigue, physical discomfort, health distress, and role/activity limitations) and self-reported fasting blood glucose levels (p<.05). No significant improvements were found in health care utilization. There were trends, however, toward less utilization (specifically, less ER use and fewer hospitalizations).

Automated Telephone Disease Management Studies

Piette and colleagues conducted a feasibility study in Department of Veterans Affairs (VA) outpatient clinics to determine whether low-income patients with diabetes were able and willing to use their touch-tone telephones to provide information about their health and self-care.78 During one month, patients completed 216 (83%) of all assessment attempts and reported a number of health problems that might otherwise have gone undetected by their clinicians. Patients overwhelmingly reported that they found the ATDM calls helpful (98%), and that they had no difficulty responding to the calls (98%). The majority also reported that they would like to receive ATDM calls as part of their usual care (89%), and that receiving such calls would make them more satisfied with their VA care (77%).

Two follow-up randomized trials were conducted. The first study determined the impact of ATDM-supported diabetes care for 280 low-income English- and Spanish-speaking patients recruited from indigent-care clinics. Intervention patients received 12-months of biweekly ATDM calls with follow-up by a diabetes nurse educator. Control patients received usual care. The second randomized trial evaluated a similar intervention tailored to VA care. In this study, 301 patients were enrolled and randomized to 12-months of ATDM-supported telephone care or usual care.

Patients in these studies completed ATDM assessments consistently over the twelve months of their participation and were satisfied with their experience79 (publication included in Appendix 1). Across both studies, patients successfully completed more than 4,000 bi-weekly ATDM assessments or 71% of all assessment attempts. Only 4% of attempts were terminated because the patient hung up the telephone prematurely. Patients reported self-monitored blood glucose levels during 70% of successful assessments. In their 12-month follow-up survey, 87% of patients reported that they were "moderately" or "very" satisfied with the ATDM calls. Eighty-six percent said that receiving such calls would make them more satisfied with their health care, and 79% said that they personally would choose to receive similar calls in the future. Spanish-speaking patients responded at least as consistently as English-speakers and were more satisfied with the system. This is despite the fact that more than half of all the Spanish-speakers reported no formal education.

Data from these studies indicate that patients with diabetes are interested in receiving self-care education via ATDM calls. On average, Spanish-speaking patients selected optional self-care tips during 64% of their ATDM calls compared to 36% for English speakers (p < .0001). Spanish speakers also selected dietary education modules more often (52% versus 28%, p < .0001). After 12 months, most Spanish speakers and roughly one fourth of English speakers continued to select each message type and received a substantial amount of education as a result.

Data from both randomized trials suggest that ATDM-supported diabetes care can improve self-care, glycemic control, and patient-centered outcomes. Among participants in the county clinic trial, outcome data were collected at 12-months for 248 (89%) of all enrollees. Compared to the usual care group, intervention patients at follow-up reported more frequent glucose self-monitoring, foot inspection, and weight monitoring, and fewer medication adherence problems (all p < .03).48 The publication is included in Appendix 1. Mean follow-up HbA1c values among intervention and control patients were 8.1% and 8.4% (p = .01), and more than twice as many intervention patients were within the normal range for HbA1c (p =.04). Intervention patients also had lower mean serum glucose levels (180mg/dl versus 221 mg/dl) and reported better glycemic control (both p < .005). On average, intervention patients reported fewer diabetic symptoms than usual care patients (4.0 versus 5.4) (p< .0001) including fewer symptoms of hyperglycemia (p< .0005) and hypoglycemia (p < .001).

Importantly, impacts within the Spanish-speaking subgroup were greater than those seen in the sample overall. The average HbA1c among Spanish speakers in the intervention group was 1.1% lower (95% confidence interval (CI) 0.2% to 1.9%) than among those in the usual care group, and six times as many Spanish-speaking intervention patients had normal end point HbA1c levels (18% versus 3%, p = .05). In addition, the mean serum glucose level at follow-up was 71 mg/dl lower (95% CI 13 to 129 mg/dl) among Spanish-speaking patients in the intervention group, and intervention patients had 1.6 fewer symptoms than Spanish-speaking controls (95% CI = 0.0 to 3.2).

ATDM-supported diabetes care also had positive impacts on patient-centered outcomes81 (publication included in Appendix 1). Compared to control patients, intervention patients reported greater satisfaction with care, especially with the technical quality of the services they received, their choice of providers and continuity of care, their communication with providers, and the quality of their health outcomes (all p< .04). There also were improvements in patients' self-efficacy to perform self-care activities (p = .006), and perceived access to care. Compared to patients receiving usual care, those receiving the intervention reported fewer symptoms of depression at follow-up (p = .02) and fewer days in bed due to illness (p< .03).

Data from the VA randomized trial currently are being analyzed. Preliminary findings indicate that impacts on patients' self-care, glycemic control, and satisfaction with care are similar to those just described.82

Summary

In summary, the substantial work on the diabetes self-management program development and results from randomized studies of outcomes and mediating processes provide a strong foundation for extending the self-management program to diabetes. These studies have also produced a set of validated outcome instruments, and verified strategies for data collection within the targeted population. This program of research is conducted within a conceptual framework that informs the implementation of the intervention.