Diabetes is a common and pernicious disease,

Transcription

1 Relationship of Hemoglobin A 1c, Age of Diabetes Diagnosis, and Ethnicity to Clinical Outcomes and Medical Costs in a Computer-Simulated Cohort of Persons With Type 2 Diabetes Gregory de Lissovoy, PhD; Dara A. Ganoczy, MPH; and Nancy F. Ray, MS Abstract Objective: To project the impact of maintaining long-term glycemic control (ie, a sustained reduction in glycosylated hemoglobin (hemoglobin A 1c [HbA 1c ]) on the lifetime incidence and direct medical costs of complications in persons with type 2 diabetes. Study Design, Patients, and Methods: Computer simulation of hypothetical patient cohorts using a published model developed by the National Institutes of Health. Results: Across all HbA 1c levels, Hispanics had the highest and whites had the lowest complication rates. With lower maintained HbA 1c, the absolute decrease in complication rates was greatest and the reduction in direct medical expenditures was highest among Hispanics (18% vs 15% for blacks and 12% for whites). Complication rates and costs were most dramatically reduced when lower levels of HbA 1c were maintained among persons with a younger age at diagnosis. Conclusions: Maintaining long-term glycemic control reduces complication rates and costs for From MEDTAP International, Inc, Bethesda, MD (GL); the Department of Public Health, East Lansing, MI (DAG); and the Medicare Payment Advisory Commission, Washington, DC (NFR). At the time of this study all authors were affiliated with MEDTAP International, Inc. The views expressed here are those of the authors and not the Medicare Payment Advisory Commission. Support for this work was provided by Becton Dickinson Consumer Health Care, Franklin Lakes, NJ. Address correspondence to: Gregory de Lissovoy, PhD, MEDTAP International, Inc, 711 Wisconsin Ave, Suite 6, Bethesda, MD delissovoy@medtap.com. medical care for all ethnic groups regardless of age at diagnosis. Relatively greater benefit is achieved by interventions targeting Hispanics and younger, newly diagnosed persons. (Am J Manag Care 2;6: ) Diabetes is a common and pernicious disease, with more than 1 million persons selfreporting this illness in the United States in Direct medical costs attributable to diabetes mellitus are substantial, exceeding an estimated $44 billion in Management of type 1 and type 2 diabetes has changed significantly during the last 15 years. The Diabetes Control and Complications Trial provided evidence that improved glycemic control reduces the risk of microvascular complications among persons with type 1 diabetes. 2 These findings were confirmed for type 2 diabetes by the United Kingdom Prospective Diabetes Study. 3 During an average 6 years of follow-up, risk of complications was reduced by 35% for each percentage point decrease in hemoglobin A 1c (HbA 1c ). Measurement of glycosylated hemoglobin HbA 1c levels has become an accepted and important indicator of metabolic control and risk of microvascular complications in persons with diabetes. 4 The American Diabetes Association (ADA) has recommended that persons with diabetes maintain a HbA 1c value of less than 7% and that HbA 1c be measured every 3 months to achieve that goal. 5 However, in one recent study, only 2% of health maintenance organization members with diabetes met these widely accepted standards. 6 Effective targeting of scarce resources for diabetes screening and treatment may be hampered by limited knowledge about the relationship of demographic VOL. 6, NO. 5 THE AMERICAN JOURNAL OF MANAGED CARE 573

2 and clinical characteristics (racial/ethnic group, age at diabetes diagnosis) and disease impact (rate of complications, treatment costs) with various levels of glycemic control. Improved understanding of these relationships could enable health plan managers to better allocate resources by implementing interventions where they are most needed and effective. To explore this issue, we used a published model that incorporates findings from numerous epidemiologic studies to estimate the lifetime clinical events and medical costs of persons with type 2 diabetes under a wide range of assumptions. 7 We first describe the structure of the model, including the stages of diabetes-related complications that may occur over a lifetime. We also describe the data and assumptions that provide the basis for model projections. Simulation results are presented by racial/ethnic group (white, black, Hispanic) and age range at diagnosis of diabetes. The effect of these characteristics on the lifetime incidence of complications and costs is examined for several maintained HbA 1c values. We conclude by examining policy implications of the results, along with steps physicians and managed care organizations can take to help patients improve and maintain glycemic control.... METHODS... This study is based on a Markov model that uses Monte Carlo techniques to estimate the lifetime clinical outcomes and direct medical costs of persons with type 2 diabetes. Developed by Eastman and colleagues at the National Institutes of Health (NIH), the model is described in detail elsewhere. 7 In a Markov model, a simulated cohort is tracked over time through a series of mutually exclusive health states that characterize the course of disease. Tables of probabilities define the sequence of feasible transitions between health states and the duration in each state. Using Monte Carlo simulation, transitions are randomly determined based on probability distributions. Health states in the model represent common complications affecting people with diabetes: retinopathy, nephropathy, neuropathy, and cardiovascular disease (Figure 1). Probabilities for disease progression and development of vascular and neurologic complications were derived by the NIH group in an exhaustive synthesis of epidemiologic studies of diabetes and clinical trials of therapeutic interventions. For example, the Wisconsin Epidemiologic Study documented the relationship between duration of diabetes, glycemic control, and development of microvascular complications affecting vision. 8 Simulated patients in the NIH model are assumed to receive the standard of care for diabetes that existed before the Diabetes Control and Complications Trial. Resource intensity and costs for this level of care were determined using sources such as the National Health Interview Survey and the 1987 National Medical Expenditure Survey. In the model, progression to a more severe health state within an organ system is independent of the status of other organ systems, so many combinations of health states can occur. Individuals are followed until they die or reach age 95 years. The model assumes that patients developing gross proteinuria received angiotensin-converting enzyme inhibitors for prevention of end-stage renal disease. Reflecting a conservative perspective on the potential benefits of therapy, the model assumes no effect of glycemic control on cardiovascular disease and does not consider the impact of glycemia on progression of proteinuria to renal failure. 7 Model cost parameters include the direct medical costs associated with general diabetes-related care (pharmacologic therapy, physician care, diagnostic testing), and costs for treatment of acute metabolic conditions, microvascular complications, and neurologic complications. The model does not include costs for acute hospitalizations for events such as hypoglycemia. Cost data were obtained from national survey data and clinical trial reports. Model outcomes include the cumulative incidence of each complication and the average lifetime cost of diabetes-related medical care. Because the original NIH model calculated costs in 1994 dollars, costs in our analysis were adjusted to 1997 dollars using the medical care component of the consumer price index. All costs were converted to present value using a 3% discount rate. Indirect costs the value of lost productivity attributable to disease are not included in the analysis. Using this model, we assessed the impact of demographic and clinical characteristics on patient outcomes associated with lifetime maintenance of HbA 1c levels of 6%, 8%, 1%, or 12%. The low end of the range (6%-8%) represents moderate to near-normal glycemic control, and the higher range (1%- 12%) indicates poor glycemic control. The patient population in the NIH model is representative of the age, gender, racial/ethnic group, and age at diabetes diagnosis as determined in national surveys. We focused our analysis on specific subsets of the popu- 574 THE AMERICAN JOURNAL OF MANAGED CARE MAY 2

3 lation simulated by the NIH model: non-hispanic whites, non-hispanic blacks, and Hispanics, with ages at diabetes diagnosis of years, years, and 65 years or older.... RESULTS... Model results demonstrate the benefits of maintaining glycemic control in all persons with type 2 diabetes. Across all race/ethnic groups and regardless of age at diabetes diagnosis, persons with lower maintained HbA 1c values had fewer complications and incurred lower direct medical costs. Although persons with lower HbA 1c values fared better regardless of demographic and clinical characteristics, the degree to which complication rates and costs were No retinopathy reduced with lower HbA 1c values differed by racial/ethnic group and age at diagnosis.... MODELING THE IMPACT OF REDUCED HbA1c LEVELS... Hispanics had the highest rates of complications across HbA 1c levels. Compared with whites with 1% HbA 1c, Hispanics had a higher incidence of background diabetic retinopathy by 13%, macular edema by 4%, proliferative diabetic retinopathy by 61%, blindness by 46%, microalbuminuria by 51%, gross proteinuria by 53%, end-stage renal disease by 58%, neuropathy by 5%, and lower-extremity amputation by 45%. When 8% rather than 1% HbA 1c was maintained, the percent reduction in the incidence of complica- Figure 1. Health States Associated With Complications of Diabetes Retinopathy Nephropathy Neuropathy CVD No nephropathy No neuropathy No CVD Racial/Ethnic Group Incidence of Complications. Maintaining lower levels of HbA 1c had differential effects on the lifetime rate of complications depending on racial/ethnic group. Projected lifetime incidences for all stages of retinopathy, nephropathy, and neuropathy by racial/ethnic group and maintained HbA 1c value are presented in Table 1. The most debilitating complication for each of these conditions is blindness, end-stage renal disease, and lower-extremity amputation, respectively. Lifetime rates of these extreme health states in newly diagnosed whites, blacks, and Hispanics with age at diagnosis of years and HbA 1c of 6%, 8%, 1%, and 12% are shown in Figures 2A, 2B, and 2C, respectively. Overall, whites had the lowest and Nonproliferative retinopathy Proliferative retinopathy Significant macular edema Visual acuity <2/1 in better eye Microalbuminuria Proteinuria ESRD Symptomatic neuropathy First LEA CVD morbidity CVD = cardiovascular disease; ESRD = end-stage renal disease; LEA = lower-extremity amputation. VOL. 6, NO. 5 THE AMERICAN JOURNAL OF MANAGED CARE 575

4 tions was greatest among whites and lowest among Hispanics, although the absolute percent change in the incidence of complications was greatest for Hispanics and blacks. Reduction in blindness was 62% in whites versus 58% in Hispanics. On the other hand, the incidence of blindness among Hispanics was 36.5% at 1% HbA 1c and 15.3% at 8%, a difference of 21.2 percentage points. For whites, the comparable rates were 19.8% and 7.6%, a difference of only 12.2 percentage points. Cost of Care. The maintained level of HbA 1c also had differential effects on lifetime costs depending on racial/ethnic group. Consistent with their higher rates of complications, Hispanics with 1%-12% maintained HbA 1c also had higher projected lifetime medical costs. Blacks had the lowest costs across all HbA 1c levels because their higher mortality meant that costs were accrued over a relatively shorter lifetime. Figure 3 shows projected lifetime costs for whites, blacks, and Hispanics newly diagnosed at age Table 1. Projected Lifetime Incidence of Diabetes Complications by Racial/Ethnic Group and Percent Maintained HbA 1c in a Newly Diagnosed Cohort of Persons Age Years Percentage of Cohort Experiencing Specified Complication by Maintained HbA 1c Racial/Ethnic Complication Group HbA 1c 6% HbA 1c 8% HbA 1c 1% HbA 1c 12% Background retinopathy White Black Hispanic Macular edema White Black Hispanic Proliferative retinopathy White Black Hispanic Blindness White Black Hispanic Microalbuminuria White Black Hispanic Gross proteinuria White Black Hispanic ESRD White Black Hispanic Neuropathy White Black Hispanic Lower-extremity amputation White Black Hispanic HbA 1c = hemoglobin A1c; ESRD = end-stage renal disease. 576 THE AMERICAN JOURNAL OF MANAGED CARE MAY 2

5 ... MODELING THE IMPACT OF REDUCED HbA1c LEVELS... years with maintained HbA 1c values of 6%, 8%, 1%, and 12%. Total costs among newly diagnosed whites ranged from $61,413 (6% HbA 1c ) to $77,369 (12% HbA 1c ); among blacks, comparable costs ranged from $48,38 to $7,776; and among Hispanics, the range was $58,636 to $89,934. With 6%-8% maintained HbA 1c, whites and Hispanics incurred approximately equal lifetime costs. However, with 1% HbA 1c, lifetime costs for Hispanics were approximately $45 higher than those for whites and $15, higher than those for blacks; this difference was even greater with 12% maintained HbA 1c. Among Hispanics with 1% maintained HbA 1c costs for complications exceeded costs for general diabetes-related care. For blacks these costs were roughly equal because higher mortality tended to shrink costs of long-term complications. Conversely, for whites who experienced relatively lower complication and mortality rates costs of general care comprised more than half the total costs. Hispanics had the highest costs for eye disease, renal disease, and neuropathy/lower-extremity amputation, followed by blacks. Whites, however, had the highest costs for new coronary heart disease: $5789 higher than those for Hispanics and $7164 higher than those for blacks. When HbA 1c was maintained at 8% rather than 1%, the decrease in projected lifetime costs was greater for Hispanics (18%) than for blacks (15%) or whites (12%). The decrease in projected lifetime expenditures with the lower HbA 1c value amounted to approximately $13,5 for Hispanics, $9 for blacks, and $85 for whites. This decrease was due to the lower rate and cost of complications; costs were lower by 36% for Hispanics, 32% for blacks, and 26% for whites. Maintaining this lower level of HbA 1c increased costs for general diabetes-related care across all groups by.5% to 2.7%. Age at Diabetes Onset and Diagnosis Incidence of Complications. Although persons with lower maintained HbA 1c values had fewer lifetime complications and costs for all 3 categories of age at diabetes diagnosis, age had a pronounced impact on outcomes. Rates for all stages of retinopathy, nephropathy, and neuropathy by age at diabetes diagnosis and maintained HbA 1c value are presented in Figure 2. Projected Lifetime Incidence of Blindness (A), End-Stage Renal Disease (B), and Lower-Extremity Amputation (C) by Racial/Ethnic Cohort and Maintained HbA 1c in a Newly Diagnosed Cohort Age Years A B C Incidence of Blindness (%) Incidence of ESRD (%) Incidence of Lower-Extremity Amputation (%) White Black Hispanic Maintained HbA 1c (%) White Black Hispanic Maintained HbA 1c (%) White Black Hispanic Maintained HbA 1c (%) HbA 1c = hemoglobin A 1c ; ESRD = end-stage renal disease. VOL. 6, NO. 5 THE AMERICAN JOURNAL OF MANAGED CARE 577

6 Table 2. As expected, the lifetime incidence of complications decreased with increasing age at diagnosis because major complications develop over a period of many years. Rates for the most severe health states by age category at diagnosis and maintained HbA 1c value are shown in Figures 4A, 4B, and 4C, respectively. Among persons with 1% maintained HbA 1c, those whose diabetes was diagnosed between the ages of 45 and 64 years experienced fewer cases of background diabetic retinopathy (14%), macular edema (33%), proliferative diabetic retinopathy (59%), blindness (61%), microalbuminuria (27%), gross proteinuria (4%), end-stage renal disease (81%), neuropathy (4%), and lower-extremity amputation (6%) compared with persons whose diabetes was diagnosed between the ages of 18 and 44 years. Persons with diagnosis after age 65 had even lower complication rates. When 8% rather than 1% HbA 1c was maintained, the incidence of complications was greatly reduced for all age groups, although the effect was more dramatic among persons with an earlier age at diag- Figure 3. Projected Lifetime Costs for a Newly Diagnosed Cohort Age Years by Racial/Ethnic Group and Maintained HbA 1c 1, 9, 8, 7, 6, Cost ($) 5, 4, 3, 2, 1, w b h w b h w b h w b Racial/Ethnic Group (w = white; b = black; h = Hispanic) Maintained HbA 1c (%) h General diabetes-related care Eye disease Neuropathy/LEA New coronary heart disease Renal disease HbA 1c = hemoglobin A 1c ; LEA = lower-extremity amputation. 578 THE AMERICAN JOURNAL OF MANAGED CARE MAY 2

7 ... MODELING THE IMPACT OF REDUCED HbA1c LEVELS... nosis because their rates of complications were higher. Persons whose diabetes was diagnosed between the ages of 18 and 44 years had a projected lifetime incidence of blindness of 53.2% and 18.2% with 1% and 8% maintained HbA 1c, respectively. By contrast, persons with diagnosis at 65 years or older had cumulative incidence rates of only 2.3% and 1.5% for 1% and 8% maintained HbA 1c, respectively. Costs of Care. Level of glycemic control also had differential effects on projected lifetime costs depending on age at diabetes diagnosis. Because of the lower incidence of complications, lifetime diabetes-related medical costs decreased dramatically with older age of disease onset and diagnosis. Figure 5 shows costs for newly diagnosed cohorts by age category for maintained HbA 1c values of 6%, 8%, 1%, and 12%. With maintained HbA 1c of 1%, total projected lifetime medical costs ranged from $17,862 to $18,953 for persons 65 years or older at diagnosis, compared with costs between $85,476 and $132,253 for persons age years at diagnosis. These cost differentials widened with higher maintained HbA 1c levels. Projected cost was substantially less with 8% rather than 1% maintained HbA 1c among persons Table 2. Projected Lifetime Incidence of Diabetes Complications by Age at Diagnosis and Percent Maintained HbA 1c Percentage of Cohort Experiencing Specified Complication by Maintained HbA 1c Age at Complication Diagnosis (y) HbA 1c 6% HbA 1c 8% HbA 1c 1% HbA 1c 12% Background retinopathy Macular edema Proliferative retinopathy Blindness Microalbuminuria Gross proteinuria ESRD Neuropathy Lower-extremity amputation HbA 1c = hemoglobin A 1c ; ESRD = end-stage renal disease. VOL. 6, NO. 5 THE AMERICAN JOURNAL OF MANAGED CARE 579

8 Figure 4. Projected Lifetime Incidence of Blindness (A), End-Stage Renal Disease (B), and Lower- Extremity Amputation (C) in a Newly Diagnosed Cohort by Age at Diagnosis and Maintained HbA 1c A B C Incidence of Blindness (%) Incidence of ESRD (%) Incidence of Lower-Extremity Amputation (%) years years Maintained HbA 1c (%) years years Maintained HbA 1c (%) years years Maintained HbA 1c (%) HbA 1c = hemoglobin A 1c ; ESRD = end-stage renal disease. whose diabetes onset and diagnosis occurred at a relatively young age. Maintaining 8% HbA 1c decreased the projected lifetime cost by almost $3, for persons with diagnosis at age years, compared with a differential of about $65 for those with diagnosis at age years.... DISCUSSION... Results of the model demonstrate the importance of maintaining glycemic control in persons with type 2 diabetes. For each category of racial/ethnic group and age at diagnosis, the projected rate of complications of diabetes and costs of treatment were substantially lower for HbA 1c levels of 8% and below versus levels of 1% and above. Although every person with diabetes benefits from improved glycemic control, model results suggest that potential clinical benefits and savings in medical care costs are especially significant in 2 of the demographic categories analyzed: Hispanics and people with diabetes onset and diagnosis between the ages of 18 and 44 years. Hispanics are one of the fastest growing ethnic groups in the US population, numbering more than 23 million. Although model results may be unreliable because risk probabilities were drawn from small samples, type 2 diabetes is 3-5 times more prevalent among Hispanics than among whites 9 and roughly 3 times more prevalent than among blacks. 1 Healthcare providers are learning that effective care for diabetes, as well as other chronic conditions, requires sensitivity to the unique culture of the Hispanic community. 11 Programs to screen for undiagnosed type 2 diabetes have typically concentrated on older persons (age 45 years and above) because of the low prevalence of diabetes in younger persons. Now there is evidence that type 2 diabetes is increasingly common at younger ages and is a growing problem among adolescents. Pinhas-Hamiel and colleagues reported a 1-fold increase between 1982 and 1994 in type 2 diabetes incidence among adolescents in greater Cincinnati. 12 Rising juvenile obesity appears to be the major factor. National Health and Nutrition Examination Survey data from indicate that approximately 14% of children and 12% of adolescents are overweight, twice as many compared with data from The cost of treating major complications of diabetes will no longer fall predominantly to Medicare because younger persons with poor glycemic control will experience complications at younger ages. 58 THE AMERICAN JOURNAL OF MANAGED CARE MAY 2

9 ... MODELING THE IMPACT OF REDUCED HbA1c LEVELS... A recent study of the costs of screening for type 2 diabetes concluded that screening may actually be more cost effective among younger people (25-44 years) than at older ages. Even though older persons are at higher risk of diabetes, the benefit of reducing complications among younger persons with diabetes outweighs the added expense of screening. 14 Early identification of persons with dia- Figure 5. Projected Lifetime Costs for a Newly Diagnosed Cohort by Age at Diagnosis and Maintained HbA1c 14, 13, 12, 11, 1, 9, 8, Cost ($) 7, 6, 5, 4, 3, 2, 1, Age at Diagnosis of Diabetes (y) Maintained HbA 1c (%) 65 General diabetes-related care Eye disease Neuropathy/LEA New coronary heart disease Renal disease HbA 1c = hemoglobin A 1c ; LEA = lower-extremity amputation. VOL. 6, NO. 5 THE AMERICAN JOURNAL OF MANAGED CARE 581

10 betes offers greater opportunities for appropriate care, hypoglycemic treatment, and education in self-care. Design of an efficient program to screen young people for diabetes should focus on those with 2 major risk factors, obesity and family history of diabetes. 15 Once persons with diabetes have been identified, the challenge is to maintain long-term glycemic control. The National Committee for Quality Assurance has joined with the ADA and the Health Care Financing Association to establish the Diabetes Quality Improvement Project, which is developing quality of care indicators for management of diabetes. Both the National Committee for Quality Assurance and the ADA advocate using the frequency of HbA 1c measurement to assess quality of care provided to persons with diabetes. Knowing that persons from certain groups are at higher risk for an expensive chronic disease such as diabetes could create an incentive for health plans to discourage their enrollment. That would clearly represent unethical and even illegal policy. Fortunately, virtually all managed care organizations believe that making quality care available to all members of the community is central to their mission. Study Limitations Risk factors in the NIH model such as race and ethnicity are derived from epidemiologic studies. Because these are mostly small-area observational studies, the conclusion that blacks or Hispanics are at intrinsically greater risk for complications of diabetes may be unwarranted. Higher rates of complications may reflect barriers to access or historical patterns of inadequate medical care. Incidence rates observed among members of an ethnic or racial group as observed in one part of the country may not be generalizable to all members of that ethnic group across the country. Measures to achieve glycemic control may involve additional costs for monitoring, office visits, and education beyond those included in the model. However, the model takes a conservative approach to estimating the costs of diabetes care and complications. For example, the model does not consider the potential savings associated with reduction in heart disease that could be achieved through improved glycemic control in conjunction with other measures to reduce coronary artery disease. 16 Because the analysis was performed from a managed care perspective, the substantial economic burden of lost productivity attributable to diabetes did not figure in the analysis. Inclusion of these costs would amplify the magnitude of cost differentials observed in the model, but would not change the underlying conclusions. Although this approach is technically correct, managed care organizations have a strong incentive to track the impact of treatment interventions on indirect costs. For plan members, improved productivity and daily functioning are important outcomes of care and are far more meaningful than abstract clinical measures such as improved glycemic control. Improving Care for Patients With Diabetes Many patients do not receive care as recommended by the National Committee for Quality Assurance and ADA guidelines. Wisdom and colleagues found that fewer than 2% of health maintenance organization patients in their study received diagnostic monitoring as recommended by the ADA. 6 Achieving a sustainable reduction in HbA 1c values among persons with type 2 diabetes is difficult. Possible strategies to improve patient care include provider education programs, provider and patient awareness of HbA 1c values, disease management including prescription of an intensive diabetes therapy regimen, and use of newer technologies for testing HbA 1c. Provider Education. Frequency of HbA 1c testing is low among persons with diabetes in the United States. For example, Weiner and colleagues reported that only 16.3% of Medicare beneficiaries had HbA 1c measurements during a 1-year period from July 199 though June In a later study, the percentage of Medicare beneficiaries with diabetes who were tested for HbA 1c or fructosamine every 6 months increased slightly, from 29.7% between 1992 and 1993 to 36.% between 1994 and Given the underuse of HbA 1c testing, additional programs to disseminate clinical practice guidelines relevant to the primary care of persons with type 2 diabetes to physicians and other medical providers may be warranted. Physician Awareness of Their Patients HbA 1c Values. Because of infrequent testing, many physicians are not aware of their patients HbA 1c values. Physician knowledge of patients HbA 1c values has been demonstrated to promote change in diabetes treatment and subsequent improvement of metabolic control, as indicated by decreased HbA 1c values. In a prospective study of 24 persons with type 1 diabetes, Larsen et al divided patients into 2 groups. 19 In one group, the physician was informed of the patient s current HbA 1c value at the time of each office visit; in the second group, physicians could order the HbA 1c test at the time of the visit, 582 THE AMERICAN JOURNAL OF MANAGED CARE MAY 2

11 ... MODELING THE IMPACT OF REDUCED HbA1c LEVELS... receiving the results later. After 1 year, the mean HbA 1c level had significantly declined from 1.1% to 9.5% among patients whose HbA 1c levels were available to their physician at the time of the clinic visit. Conversely, HbA 1c levels were unchanged in the group of patients whose HbA 1c levels were not immediately available to the physician, with mean initial and final values of 1.% and 1.1%, respectively. Reduction in HbA 1c values was even greater among a subset of patients with initial HbA 1c values exceeding 1%. In this subgroup, mean HbA 1c values decreased from 11.6% to 1.% for patients with HbA 1c values available at the time of visit, compared with a decrease from a mean of 11.7% to 11.4% for patients whose HbA 1c values could be provided at a later date. The study by Larsen et al suggests that, when physicians are well informed about their patient s HbA 1c levels, that information potentially improves patient care and patient outcomes. Patient Awareness of HbA 1c Values. Educational programs stressing self-management for persons with diabetes may result in a better understanding of the illness, with improved glycemic control. 2,21 For example, Snellman and Eckerbom evaluated home monitoring of HbA 1c. 22 Over an 8-year study period, frequent HbA 1c testing resulted in a significant decline in mean HbA 1c values, from 7.5% to 7.1%. Moreover, the average number of clinic visits also declined during this time period, from 4.6 to 3.2 visits per year. The implication of this study is that systems or technologies that facilitate patients knowledge of their HbA 1c values can both improve clinical outcomes and lower resource use. Despite the beneficial impact of patient education, only 35% of individuals with diabetes have taken an educational class or program. 23 One of the major barriers to diabetes education has been the lack of third-party reimbursement for clinicians offering this service. Disease Management Programs. Access to and continuity of medical care are critical components in the management of persons with diabetes. Effective diabetes care requires a high degree of interaction and communication between the patient and the healthcare provider. Recognition and knowledge of appropriate management by physicians and other healthcare providers, including nutritionists and diabetes educators, may influence clinical outcomes of affected patients. Provider practice patterns are known to vary based on the region of the country and the degree to which an area is urban versus rural. 24,25 Characteristics of the healthcare system, such as the availability, organization, and delivery of healthcare services, are additional determinants of access to care. 26,27 Disease management programs have been implemented by many health plans to ensure that patients are regularly monitored, given access to educational programs and support groups, and managed in accordance with recognized clinical guidelines. Quality of care guidelines established by the Diabetes Quality Improvement Project provide a framework for disease management program design that complies with standards of care required for National Committee for Quality Assurance accreditation. 28 New Technology. Several new approaches to diagnostic testing can provide HbA 1c values in a more timely fashion to both medical professionals and persons with diabetes. Point-of-care testing devices designed for clinic use can provide accurate HbA 1c readings in a few minutes. Home test kits make it convenient for patients to draw a blood sample, mail it to a central laboratory, and obtain results by telephone before their office visit. Physicians also receive the results before the patient s visit, giving them more information about their patient and enabling them to provide better care. Innovations such as these may increase adherence to testing at regular intervals such as those recommended by the ADA. This may, in turn, result in sustainable decreases in HbA 1c values among persons with type 2 diabetes.... CONCLUSIONS... Simulation results presented here indicate that reduction in diabetes-related complications can substantially reduce direct medical costs when HbA 1c is maintained at a level of 8% compared with 1%, but there is relatively little further gain when HbA 1c is maintained below 8%. That could be interpreted as a rationale for a relaxed glycemia control strategy, as advocated by some clinicians. 29 However, as shown in Table 2, the incidence of moderate complications such as macular edema and neuropathy is still quite substantial with 8% HbA 1c. Although the associated medical costs may be relatively modest, these conditions exact a huge toll through impaired functional status and reduced quality of life. Strategies for achieving and maintaining good glycemic control are available and feasible. With younger persons now at risk for type 2 diabetes, the cost of poor diabetes care will no longer be borne predominantly by Medicare, but will increasingly shift to providers responsible for the under-65 population. Some healthcare organizations may be VOL. 6, NO. 5 THE AMERICAN JOURNAL OF MANAGED CARE 583

12 reluctant to aggressively implement diabetes management programs because of a perception that benefits will only be observed many years in the future. A recent study demonstrated that, after a period as short as 12 weeks, improved glycemic control can achieve a measurable decrease in healthcare utilization, a reduction in days of restricted activity, and improved quality of life. 3 These findings should motivate healthcare organizations to intensify their efforts to identify persons with diabetes and to help them achieve long-term glycemic control.... REFERENCES American Diabetes Association. Economic consequences of diabetes in the United States in Diabetes Care 1998;21: The Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulindependent diabetes mellitus. N Engl J Med 1993;329: United Kingdom Prospective Diabetes Study Group. United Kingdom Prospective Diabetes Study 24: A 6 year randomized, controlled trial comparing sulfonylurea, insulin, and metformin therapy in patients with newly diagnosed type 2 diabetes that could not be controlled with diet therapy. Ann Intern Med 1998;3: Klein R, Klein BE, Moss SE, Davis MD, DeMets DL. Glycosylated hemoglobin predicts the incidence and progression of diabetic retinopathy. JAMA 1988;26: American Diabetes Association. Standards of medical care for patients with diabetes mellitus. Diabetes Care 1997;2(suppl 1):S5-S Wisdom K, Fryzek JP, Havstad SL, Anderson RM, Dreiling MC, Tilley BC. Comparison of laboratory test frequency and test results between African Americans and Caucasians with diabetes: Opportunity for improvement. Diabetes Care 1997;2: Eastman RC, Javitt JC, Herman WH, et al. Model of complications of NIDDM. I. Model construction and assumptions. Diabetes Care 1997;2: Moss SE, Klein R, Klein BE. Ten year incidence of visual loss in a diabetic population. Ophthalmology 1994;1: Ramirez AG. Hypertension in Hispanic Americans: Overview of the population. Public Health Rep 1996;111(suppl 2): Kenny SJ, Aubert RE, Geiss LS. Prevalence and incidence of non-insulin-dependent diabetes. In: Harris MI, Cowie CC, Stern MP, Boyko EJ, Reiber GE, Bennett PH, eds. Diabetes in America, 2nd ed. Bethesda, MD: National Institutes of Health, National Institute of Digestive and Kidney Diseases; NIH publication Caudle P. Providing culturally sensitive health care to Hispanic clients. Nurs Pract 1993;18: Pinhas-Hamiel O, Dolan LM, Daniels SR, Standiford D, Khoury PR, Zeitler P. Increased incidence of non-insulindependent diabetes mellitus among adolescents. J Pediatr 1996;128: Centers for Disease Control. Update: Prevalence of overweight among children, adolescents, and adults United States, MMWR Wkly Rep 1997;46: CDC Diabetes Cost-Effectiveness Study Group. The costeffectiveness of screening for type 2 diabetes. JAMA 1998;28: Cowie CC, Harris MI, Eberhardt MS. Frequency and determinants of screening for diabetes in the US. Diabetes Care 1994;17: Turner RC, Millns H, Neil HAW, et al for the United Kingdom Prospective Diabetes Study Group. Risk factors for coronary artery disease in non-insulin dependent diabetes mellitus: United Kingdom Prospective Diabetes Study (UKPDS 23). BMJ 1998;316: Weiner JP, Parente ST, Garnick DW, et al. Variation in office-based quality. A claims-based profile of care provided to Medicare patients with diabetes. JAMA 1995;273: Hays K, Hogan C, Docteur E. Monitoring Access of Medicare Beneficiaries. Washington, DC: Physician Payment Review Commission; Publication Larsen ML, Hørder M, Mogensen EF. Effect of long-term monitoring of glycosylated hemoglobin levels in insulindependent diabetes mellitus. N Engl J Med 199;323: Harris MI, Cowie CC, Howie LJ. Self-monitoring of blood glucose by adults with diabetes in the US population. Diabetes Care 1993;16: Padgett D, Mumford E, Hynes M, et al. Meta-analysis of the effects of educational and psychosocial interventions on the management of diabetes mellitus. J Clin Epidemiol 1988;41: Snellman K, Eckerbom S. Possibilities and advantages with home sampling of HbA 1c : Eight years experience. Diabet Med 1997;14: Coonrod BA, Betschart J, Harris MI. Frequency and determination of diabetes patient education among adults in the US population. Diabetes Care 1994;17: Wennberg JE, Freeman JL, Culp WJ. Are hospital services rationed in New Haven or over-utilized in Boston? Lancet 1987;1: Wennberg JE, Gittelsohn A. Variations in medical care among small areas. Sci Am 1982;246: Weissman JS, Epstein AM. Falling Through the Safety Net. Baltimore, MD: Johns Hopkins University Press; Aday L, Andersen R. Equity of access to medical care: A conceptual and empirical overview. Med Care 1981;19(12 suppl): Health Care Financing Administration. Quality of care information: Project activities. Diabetes Quality Improvement Project (DQIP). Available at: Accessed April 2, Vijan S, Hofer TP, Hayward RA. Estimated benefits of glycemic control in microvascular complications in type 2 diabetes. Ann Intern Med 1997;9: Testa MA, Simonson DC. Health economic benefits and quality of life during improved glycemic control in patients with type 2 diabetes mellitus. JAMA 1998;28: THE AMERICAN JOURNAL OF MANAGED CARE MAY 2