Contents
Contributors
Search

INTRODUCTION

Diabetes is a condition that easily lends itself to a disease management approach. It is characterized by easily measurable and quantifiable outcomes and process measures. When implemented, diabetes disease management programs can significantly improve outcomes, whether compared to the American Diabetes Association (ADA) standards (1) or when assessed based on economic models (2, 3). However, in spite of this, on a population level the data on patients with diabetes show that targets are rarely reached (4-6). This means that these programs are not uniformly available and/or they are implemented poorly. This review will cover the history of diabetes disease management, its current application and goals for the future.

DEFINITION

A disease management program should include four basic components (7). These are: 1. practical treatment guidelines developed using evidence-based models, 2. diabetes self-management education programs for patients and health-care providers, 3. evaluation tools to measure the impact on the program on patient outcomes, and 4. a team approach to the management of the disease condition. Disease management is not suited to the treatment of all conditions, but diabetes fits many of the criteria for being a candidate for disease management. These criteria include a high disease prevalence in the population (8), disproportionally high cost associated with treating the disease (9), complications that are preventable (10, 11) and process and outcomes that are readily measurable (12).

HISTORY

In the past, our efforts to treat patients with diabetes have not resulted in success. If they had, diabetes would not be the leading cause of adult blindness, kidney failure and nontraumatic lower extremity amputation in the United States (13). Looking at A1c levels, one of the major outcomes measures for treating diabetes, the level in patients treated in a fee for service setting was 9.6% (6). Although one of the promises of managed care was that it would provide a systematic approach to improving the distribution of preventive care services at lower costs (14), A1c levels averaged 9.5% in managed care settings. A similar mean A1c level was found in a study using the NHANES III dataset (15). However, these data reflect findings from 1986 - 1996. A more recent analysis shows that some improvement in average A1c levels has occurred, and the level is now down to 8.2% (16). Although this is well above the target of less than 7%, it is an improvement.

In the past few years disease management programs have become increasingly common in health maintenance organizations. By 1998 it was reported that 79% of HMO's offered diabetes disease management programs (17). These programs often vary in definition, scope and availability to patients but clearly they are increasing and offer the opportunity for a coordinated approach to diabetes care.

The publication of the DCCT study (10) and the UKPDS trial (11) showed definitely that patients with both type 1 and type 2 diabetes benefited from improved glycemic control. These data made the need for disease management more convincing and helped fuel the search for developing newer approaches for managing patients. One of the important lessons from the DCCT was the need for intensive team management of the type 1 diabetic patient. Resources expended were high (18) and difficult to apply consistently. After the DCCT Trial was over, patients in both groups were encouraged to perform intensive diabetes management and their care was transitioned to routine practice. The outcome of these patients is being evaluated through the Epidemiology of Diabetes Interventions and Complications (EDIC) Trial (19). After two years the patients in the intensively treated arm of the DCCT had an increase in HbA1c levels from 7.4% at the end of the study to approximately 7.9% and the patients in the conventionally treated arm of the DCCT had a decrease in HbA1c level from 9.1% to 8.0%. After seven years the average A1c level is 8.3% in patients who were in the conventionally treated group and 8.1% in patients in the intensively treated group (20), underscoring the difficulty in maintaining near euglycemia in patients with type 1 diabetes without the structure of a research trial (or perhaps a good disease management program). The UKPDS study used fewer resources, but patients had access to nonphysician health care providers such as dietitians, who gave feedback regarding dietary therapy.

Diabetes is particularly well suited to using a comprehensive approach for managing multiple risk factors because outcomes are easily measured and integration of care is critical to prevent both the microvascular and macrovascular complications of the disease. This has lead to its inclusion in national reporting measures. Criteria for assessing how well diabetes care is being provided was developed as part of the Diabetes Quality Improvement Project (DQIP) (21). The DQIP measure set is shown in Table 1.

DIABETES SELF-MANAGEMENT EDUCATION

Diabetes self-management education (DSME) is an integral component of diabetes disease management. Education can be provided either in an individual or a group format (22). Although it has been difficult to prove that DSME, in isolation, improves outcomes, as part of a treatment program it has been shown to enhance psychosocial and health outcomes (23-25). This education can occur in the setting of a primary care practice (25) and is associated with improved processes of care and better health.

A very important element to DSME is that of medical nutrition therapy (MNT) and exercise. A recent review of studies on MNT have shown significant improvements in weight and A1c levels in patients who participate in the education process provided by a registered dietitian (26). Another review (27) evaluated the effectiveness of dietary interventions in patients with chronic disease as reported in 92 independent studies. Most studies were similarly successful in reducing intake of total and saturated fat (7.3% reduction in the percentage of calories from fat), and increasing fruit and vegetable intake (an average increase of 0.6 servings per day). The two intervention components seemed to be particularly promising in modifying dietary behavior--goal setting and small groups. Therefore, interventions designed to improve dietary behaviors are effective, although the heterogeneity of interventions used make conclusions as to the most effective interventions impossible.

The National Standards for Diabetes Self-Management Education Programs were developed to limit variation in educational interventions (28) and provide basic standards for care. The key elements of the type of self-management education to be provided are presented in Table 2. These standards stress the need for patients with diabetes to acquire both knowledge and skills to manage their disease, which should result in more informed choices and desirable behavior. The focus is shifted from simply providing knowledge to assisting patients in developing strategies for behavior change. Education programs meet the National Standards for Diabetes Self-management Education Programs are recognized by the ADA as programs of excellence (www.diabetes.org).

The National Diabetes Education Program (NDEP) was started in 1997 through a joint effort by the National Institutes of Health and the Centers for Disease Control and Prevention (http://ndep.nih.gov) (29). It is a public/private partnership that now includes involvement of over 200 additional organizations. The goal of the NDEP is to promote early diagnosis and to improve the treatment and outcomes for people with diabetes. Initially the NDEP focused on therapies that improved glucose control, but has recently expanded to include an initiative to promote optimal control of lipids and blood pressure, in addition to optimal control of blood glucose. The audience for the initiative includes people with diabetes and their families (with special emphasis given to those who are members of high risk minority groups), the general public, health care providers and health care payers, purchasers and policy makers.

The NDEP's messages and approaches are consistent with the American Diabetes Association (ADA) and other national guidelines, as well as national outcome-focused programs, such as the Diabetes Quality Improvement Program, and they are designed to improve health care provider practice and patient outcomes. The guidelines are designed to assist health care providers and people with diabetes to:

• practice goal-oriented management of blood glucose, lipids, and blood pressure to help prevent or delay diabetes complications;
• ensure that people with diabetes receive diabetes self-management education, including self-monitoring of blood glucose as necessary;
• ensure proper food intake and physical activity to help achieve target values, including body weight;
• use the A1c test for monitoring blood glucose control and for guiding therapy to achieve blood glucose target levels;
• use combination drug therapy as necessary to achieve and maintain target values;
• use a coordinated team approach to patient care.

Information regarding participating in the NDEP can be found on their website.

EFFECTIVENESS OF DISEASE MANAGEMENT PROGRAMS

Primary care providers and even endocrinologists often don't have the time or resources to treat patients to target levels of A1c, LDL cholesterol and blood pressure (30). Diabetes case managers can bridge the gap between primary care providers and patients by providing focused care regarding one major issue and its complications. Diabetes case managers often function with a team of health care providers-nutritionists, psychologists, pharmacists, social workers and others, who can work together to address each patient's needs. In addition to prescribing medications they can provide diabetes self-management education in both group and individual settings. All of this can be a helpful adjunct to a patient's regular primary care.

Multiple studies now exist to support the benefits of diabetes disease management. Some of the studies are summarized in Table 3. Although outcomes can clearly be improved, the methods used for creating change vary greatly. And in few instances are methods for disease management compared, to determine which are the most cost-effective.

The simplest method of diabetes disease management is that of creating a diabetes registry in order to track patient outcomes and provide feedback to primary care providers. This is further enhanced with decision support tools and nurse case managers who can provide intensive care for patients who need it (31). In a large community internal medicine practice a diabetes registry was created and some of the patients in the registry received direct letters regarding their need for care. Patients who received letters had an improvement in A1c levels and lower LDL cholesterol levels (32).

In a study where 193 patients were stratified based on risk, a disease management intervention was helpful (33). After a year the number of patients in the low-risk category (A1c <7%) increased by 51.1%. Patients at the highest risk for coronary heart disease (LDL >130 mg/dl) decreased from 25.4% at baseline to 20.2%. Patients with a blood pressure <130/85 mmHg increased from 23.8 to 44.6%. Patients and providers expressed significant increases in satisfaction with the program.

The studies mentioned above do not provide long term follow up data or data on the cost-effectiveness of care. A few studies of diabetes disease management have included cost data. One study (34) shows improvement in the frequency of A1c testing through implementation of a diabetes disease management (an increase in the frequency of testing by 44.9%). In addition a drop was seen in every dimension of medical-service utilization. In-patient admissions fell by 391 per 1,000 and total medical costs were also reduced. Another, much larger study (35) was performed from 1992-1997 in a staff model health maintenance organization in Washington State. Data on a total of 4744 patients was collected and characterized as improved if the A1c level fell by at least 1% and was sustained for at least a year. Out of the total, 732 were considered improved. Those did not improve had a much higher baseline A1c level (10%) than those who did (7.7%). Mean total health care costs were $685 to $950 less each year in the improved cohort. Additionally, utilization was consistently lower in the improved cohort. These cost saving were seen within 1 - 2 years of improvement of glycemic control.

A retrospective review was performed of health care claims and other measures of health care use over 2 years among 6,799 patients continuously enrolled in the Geisinger Health Plan (36). Data on patients enrolled in an opt-in disease management program (n=3,118) was compared to those who were not enrolled (n=3,681). Per member per month paid claims averaged $394.62 for program patients compared with $502.48 for nonprogram patients (p <0.05). Program patients had lower inpatient health care use (mean of 0.12 admissions per patient per year and 0.56 inpatient days per patient per year) than in nonprogram patients (0.16 and 0.98, P <0.05 for both measures). Program patients also experienced fewer emergency room visits than nonprogram patients but had a higher number of primary care visits.

Other large studies include a retrospective review of the effect of the Diabetes Treatment Center of America (DTCA) model (37) using data from 7,000 members treated in one of seven managed care plans was evaluated. Compliance with national guidelines for process measures increased-annual rates of assessing A1c levels increased from 34% to 76%, from 23% to 40% for retinal exams and from 2 to 25% for documented foot exams. However, average A1c levels decreased only from 8.9% to 8.5%. Overall, these changes resulted in a total cost decrease of $44 per diabetic member per month, largely due to a reduction in inpatient costs. In another study (38), a computer system was developed to guide and track the care provided to diabetic patients. Of 8,200 diabetic patients, the 30% considered to be at highest risk were entered into the system. Measures of compliance with annual diabetes process and outcomes measures were higher in the computer-managed group. Inpatient utilization also decreased.

In the Midwest the John Deere Health Care Plan provides coverage to 20,000 people with diabetes across 5 states (39). Half of their diabetic population, or 10,000 patients, were enrolled in a DTCA-like program. After one year, frequency of testing for diabetes process measures improved (e.g. 19% for A1c, 20% for lipids, etc), average A1c levels fell by 0.5% and hospitalization rates decreased by 22% and hospital days were reduced by 54%. Overall total medical costs per diabetic patient were reduced by 12%.

Another interesting model is the program at Kaiser-Permanente, Northern California. In this model patients with chronic diseases, such as diabetes, are divided into three levels-the first level means they are stable, the second have poorly controlled conditions and the third level means a patient has complex multi-diagnoses and/or are high use patients (31). Patients in Level 3 are assigned a nurse case manager or medical social worker. Patients in Level 2 are seen by care managers, who range from nurses to pharmacists to dietitians and work intensively with patients for 6 to 15 months until the patient stabilizes and returns to Level 1 (31). Specific outcomes data are not published for diabetes, but the effectiveness of the program in terms of the management of adult asthma have been shown.

In all of these larger population-based studies, improvement was reported in several measures. These data show that diabetes disease management programs can improve specific outcomes. When cost analyses are done, diabetes disease management is usually associated with costs savings or are at least cost-neutral with improved outcomes. However, the average A1c levels in the studies remain above 8%, which is higher than the control arm in the UKPDS Trial and much higher than the American Diabetes Association recommended target of less than 7%. Additionally, in most of these studies data on lipid levels are not available and treatment of macrovascular risk factors is not aggressively pursued. Finally, data on the impact of diabetes disease management on behavioral outcomes are often not reported in conjunction with data on metabolic outcomes and cost analyses.

Minority populations are at particularly high risk for developing type 2 diabetes and many of these patients lack access to high quality medical care. In a small study of pharmacist-managed diabetes care in a free medical clinic, process and outcomes measures were improved in the population followed under the disease management program (40). Data from 89 patients treated by pharmacists following algorithms were compared to 92 diabetic patients receiving routine care. The experimental patients had higher baseline A1c levels (8.8% versus 7.9%) and had more complications. However, during the study there was an increased frequency of A1c and lipid measurements, an increase in compliance with recommended foot and retinal exams and a significant reduction in A1c levels (-0.8% vs -0.05%). In Project Dulce (41), using an staged diabetes management approach in poor Latino patients with type 2 diabetes and a A1c level >9% an improvement over one year was seen (A1c levels fell from 11% to 9%). This program is effective, but only short term data is available and A1c levels still remained well above target.

Clinica Campesina (31) is located in Denver, Colorado, and provides care to a largely uninsured Hispanic population. As part of the diabetes collaborative (42) the clinic used primary care teams, a diabetes registry, physician reminders, diabetes group visits and a focus on diabetes self-management education to improve outcomes. From October 1998 the average A1c level was 10.5% and fell to 8.6% in March 2000. The percentage of patients with self-management goals increased from 3% to 65% and the percentage having retinal examinations increased from 7% to 51%.

In Atlanta, in a large public hospital treating a primarily low income African American population a diabetes treatment program has been implemented (43, 44). The structured care program consists of intensive education in lifestyle modification, self-management training, and diet, coupled with intensification of medical therapy when needed to control hyperglycemia. The overall treatment approach has been shown to result in significantly lower A1c levels with a fall from a baseline of 9.3% to 8.2% at one year (p<0.001).

In Tennesse a statewide Medicaid program (TennCare) was established to provide sercies through capitated managed care organizations (45). A retrospective chart review was performed of 121 patients seen in 1992/93 and then after enrollment in the system (1995/96). Claims data were used to assess baseline characteristics and chart review data were used to assess health services. The average number of clinic visits increased from 6.4 to 8.2. Yearly rates of foot examinations (0.2 vs. 0.5), retinal examinations (0.6 vs. 1.0), and number of LDL and A1c measurements all increased. Average A1c decreased from 10.3 to 8.2. Although hospitalizations and hospital days increased overall, there was no increase in emergency visits, preventable emergency visits, or preventable hospitalizations. Therefore, glycemic outcomes improved but utilization of health care resources increased, probably appropriately in this underserved population with a chronic illness.

Thus, diabetes disease management is effective in low income as well as in higher income populations, although not as well studied in the lower socioeconomic groups.

FINANCING DIABETES DISEASE MANAGEMENT

Diabetes is a costly disease. The 1997 estimated annual economic cost of diagnosed cases of diabetes was $98,152,000 (46). This total includes $44,138,000 in direct medical and treatment costs and $54,014,000 attributed to disability and premature mortality. Total annual health care costs (diabetes and nondiabetes-related) for persons with diabetes were estimated to be $77,700,000. The per capita health care cost for these individuals was about $10,071 compared to a per capita cost of $2,699 for persons without diabetes (46).

The majority of health care costs (62%) are associated with inpatient hospital care; outpatient services account for about 24% of costs. About 9.6% of expenditures are related to pharmacy costs (47). Chronic diabetic complications account for nearly 27% of total health care costs, with 64% of these costs being related to cardiovascular disease. About half of total health care expenditures are related to general medical conditions such as liver disease, septicemia, respiratory failure, and affective disorders. Patients with diabetes have a longer length of stay and a higher readmission rate compared to patients without diabetes (48, 49). Therefore, targeting the inpatient period as well as the post discharge interval can be helpful in reducing the costs and improving outcomes.

When patients with diabetes are hospitalized their diabetes medications are usually substantially changed. Sliding scales are often implemented. Unfortunately, this leads to hyperglycemia. It has been shown that approximately 40% of patients with diabetes will have at least one blood sugar level greater than 300 mg/dl and 23% will experience hypoglycemia during a hospital stay. The risk of hyperglycemia is increased 3 fold by the use of a sliding scale (50). Hyperglycemia is associated with poorer outcomes in stroke (51, 52) and following surgery. In one study of surgical patients, a glucose level of >220 mg/dl on the first postoperative day increased the risk of serious infections sixfold (53). This finding has been confirmed in other studies (54).

An increasing body of literature suggests that aggressive treatment of hyperglycemia can improve outcomes in hospitalized patients (55). In one trial, use of IV insulin and glucose followed by subcutaneous insulin in patients with an acute myocardial infarction lowered mortality at one year (56). Results of intensive glucose control following coronary artery bypass surgery have been positive. IV insulin has been shown to reduce in hospital mortality (57) and lowers rates of sternal wounds (58, 59). In an ICU setting, an insulin infusion was given to patients to maintain blood glucose levels in the range of 80 - 110 mg/dl. The intensively treated group had a reduction in both morbidity and mortality compared to patients who were conventionally treated (60).

When treating hospitalized patients with diabetic ketoacidosis, endocrinologists have a shorter length of stay, use fewer resources and have a lower readmission rate (61). In another study looking at the impact of a diabetes team on the treatment of diabetes in hospitalized patients, if patients had a primary diagnosis of diabetes the length of stay was 5.0 versus 7.5 days in control group. There was no change in the length of stay in patients with a secondary diagnosis of diabetes (62). However, when admission rates after three months were measured, they were 15% in the patients followed by the diabetes team compared to 32% in the control group.

These improvements in outcomes when discharged patients are followed closely has been seen in other groups, such as elderly patients, as well (63). Focusing on reducing inpatient costs, initially by providing focused care and eventually by lowering rates of complications will clearly save money. However, this cost savings is often difficult to track and not all providers and groups are equally at risk for outpatient and inpatient costs. None-the-less this remains an important area to explore in terms of lowering the total cost of diabetes care and improving the quality of life for patients with diabetes.

Work disability and quality of life are two areas that can be substantially affected by adequate diabetes treatment. Rates of disability in patients with diabetes has been high-rates are 25.6% in patients with diabetes compared to 7.8% in patients without diabetes (64). Annually this translates to a loss of nearly 88 million disability days with 74,927 workers reported to be permanently disabled (46). On average, persons age 18 to 64 with diabetes lost 8.3 days from work as compared with 1.7 days for persons without diabetes.

Work disability can be reduced if patients with diabetes are appropriately treated (65). This is clearly an important issue to employers who often pay for health insurance. In a randomized, controlled study of 569 patients with type 2 diabetes glycemic control was improved control using long acting glipizide (glipizide GITS)(65). As expected, after 12 weeks, glycemic parameters were better in the treated patients. In addition, however, quality of life was improved as well as were a number of health economic outcomes. In the glipizide GITS group there was higher retained employment, less absenteeism, fewer bed-days and less disability.

Other areas of diabetes treatment are also important and are associated with cost effective care. For instance, patients with type 2 diabetes are at very high risk for cardiovascular disease (66). In a study examining the subset in the Scandianavian Simvastatin Survival Study (4S) with impaired fasting glucose or diabetes, use of simvastatin reduced cardiovascular disease-related hospitalizations. Cardiovascular disease related hospital days were reduced by 23% by simvastatin in the normal fasting glucose group and by 40% in diabetic subjects (67, 68). Average length of stay was reduced by 2.4 days in treated diabetic subjects compared to those on placebo (67). Use of drugs such as angiotension converting enzyme inhibitors and angiotension receptor blockers may lower rates of progression to renal failure in patients with diabetes, as well as decrease risk of developing congestive heart failure or even death (69-72).

Studies such as the UKPDS trial have shown that although pharmacy costs increase in intensively treated patients (73), this costs is more than offset by the lower costs of treating chronic complications (largely due to hospitalizations) found with stricter glycemic control. The UKPDS trial did not include aggressive lipid lowering in its study design, which would likely further reduce the costs of chronic complications. Studies such as the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial, which is combining intensive glycemic and hypertension control and aggressive lipid lowering should provide us with more data over time.

THE FUTURE

New improvements will continue to improve diabetes management and should be integrated into disease management systems. Some of these technologies will be easily integrated. For instance, continuous monitoring of blood glucose levels through subcutaneous sensors will provide more data on day long glucose profiles than has ever been available before (74, 75). However, analysis of the data will require that patients and health care providers learn to interpret it differently. It will be very helpful to patients to have alarms that can alert them to blood sugar levels that are too high or too low. It will provide a feeling of safety that has been lacking for patients taking insulin. But patients and providers will need to interpret glucose levels based on trend analysis and algorithims will need to be developed to help patients adjust to rising and falling glucose values. A glucose level of 80 mg/dl will require a different action if it is increasing toward 100 mg/dl or falling to 60 mg/dl.

Already more than 200 possible combinations of agents for treating type 2 diabetes (76). This will only increase over time and increase the complexity of diabetes management. In a system of diabetes care it is possible to develop clear algorithms for using drugs in sequence, increasing, decreasing and adding drugs based on efficacy and side effects. We know that monotherapy fails in most people over time (11). However, recently the glycohemoglobin assessment program was undertaken to obtain data on the current quality of diabetes care being provided to patients with type 2 diabetes in the Western United States (16). The study was a noncomparative multi-center epidemiological survey. Data was collected from consecutively enrolled patients at 9 separate practice sites. Patients who were included were diagnosed as having type 2 diabetes treated with pharmacologic therapy. 588 patients were included in the final analysis. The overall mean A1c level was 8.2%. However, only 20.4% of patients achieved a A1c level of <7.0%. On the other extreme, only 18.9% of patients had a A1c level of >9.5%. Oral antidiabetic agents, either alone or in combination were more successful at enabling patients to reach treatment goals compared to therapies involving insulin. 59% of patients were treated with combination therapies and only 12% were treated with insulin alone. The mean HbA1c level in this study is lower than in prior surveys. Use of combination therapy for the treatment of type 2 diabetes has increased and the use of insulin as monotherapy has decreased. Although more patients need to reach the ADA target of a A1c level of <7.0%, fewer have extremely high A1c levels than they did in the past. This suggests that the quality of diabetes care is improving which is trend that will hopefully continue.

Unfortunately, type 2 diabetes is occurring in younger and younger age groups (77). Strategies for disease management will increasingly need to include detection of diabetes and diabetes risk factors in children and adolescents with an attempt to reduce the number of people who develop the disease. The Diabetes Prevention Program (DPP) has shown that consistent adherence to medical nutrition therapy and physical activity can prevent the progression of impaired glucose tolerance to type 2 diabetes (78). Unfortunately, it is difficult for many patients to adhere long term to these lifestyle recommendations. In the DPP use of medications to prevent diabetes was also tested and it was found that metformin helped to preventing diabetes, as well, particularly in younger, overweight individuals. In a smaller study, troglitazone reduced rates of the development of type 2 diabetes in women who are at high risk for getting the disease (79). Therefore, emerging data suggests that pharmacologic therapy may be a useful adjunct to MNT and physical activity to delay or prevent progression to type 2 diabetes. Integration of techniques for disease prevention will become important elements in diabetes disease management because forestalling the development of the disease will hopefully decrease the duration of exposure and risk for complications.

CONCLUSIONS

Diabetes disease management is a concept that is here to stay. On both a small and a large scale it has been shown to improve the coordination of care provided to diabetic patients and improve outcomes. Many organizations have developed or are in the process of developing these programs. Implementation of these programs remains more difficult, in part due to economic constraints as well as due to a lack of uniform computerized health care data on patients. However, disease management is evolving. With the advent of increasing numbers of drugs essential for treating diabetes and their proven benefits it is helpful for physicians to have a system for starting and tracking the necessary therapies to provide optimal care. For instance, most patients with type 2 diabetes will end up on one or more medication for treating their hyperglycemia, a drug or two for treating diabetic dyslipidemia, an ACE-inhibitor or ARB, another drug or so for treating hypertension and an anti-platelet agent plus any other medication they might require. Equally important is the diabetes self-management education component of disease management, which helps patients understand what they must do to improve their health and quality of life.

Cost savings are now being realized at the level of inpatient care. Ultimately fewer chronic complications should occur. Employers will hopefully begin to realize an improvement in the job performance of their diabetic patients as less time is lost due to disability. Overall patients, employers, physicians and health care organizations will benefit from the systematic utilization of diabetes disease management programs. They have proven their effectiveness. Now they just need to be implemented.