Diabetes mellitus places a significant burden on the U.S. healthcare system. Because of the potential to reduce diabetic complications and costs through intensive management, diabetes has become a primary target for disease management programs. We performed a retrospective analysis of short-term baseline and follow-up clinical, economic, and member and provider satisfaction data from approximately 7,000 people with diabetes being treated through seven managed care plans using Diabetes Treatment Centers of America’s Diabetes NetCareSM, (Nashville, TN), a comprehensive diabetes management program. Our analysis indicates that Diabetes NetCareSM achieved gross economic adjusted savings of $50 per diabetic member per month (12.3%), with gross unadjusted savings of $44 (10.9%) per diabetic member per month. Hospital admissions per 1,000 diabetic member years decreased by 18%, and bed days fell by 21%. Patients with diabetes were more likely to get HbA1c tests, foot exams, eye exams, and cholesterol screenings while enrolled in the program. These data suggest that implementation of a comprehensive healthcare management program for people with diabetes can lead to substantial improvements in costs and clinical outcomes in the short term. It is expected that improvements will increase over time, with continuing improvements in health status and a reduction in the number of future diabetic complications. (J Clin Endocrinol Metab 83: 2635–2642, 1998)
ONE of every seven dollars spent on health care in the United States is spent on behalf of a person with diabetes (1). According to data from the most recent National Health and Nutrition Examination Survey, 5.1% of the U.S. adult population, or 10.2 million Americans, are confirmed diabetics. Moreover, there are believed to be an additional 5.4 million undiagnosed cases (2, 3). People with diabetes are at increased risk of developing numerous micro- and macrovascular complications including renal, ophthalmic, neurological, and circulatory disorders.
In large part because of chronic complications, medical expenditures for people with diabetes are almost four times as high as expenditures for their nondiabetic counterparts (1,2). In 1992, medical expenditures for people with diabetes were estimated to be $105 billion; however, only 16% of direct medical expenditures for diabetics are attributable to diabetes care and acute glycemic events. The remainder of medical expenditures goes to routine medical care, chronic complications of diabetes, and other significant medical conditions including liver disease, malignant neoplasms, gastritis,and affective disorders (2).
The Diabetes Control and Complications Trial (DCCT) was a landmark multicenter trial, sponsored by the National Institutes of Health, that demonstrated that tight control of blood glucose levels lowers Type 1 patients’ risk of developing chronic microvascular complications of diabetes and slows progression of those complications. Over the study period, which averaged 7 years, there was a 50–75% reduction in risk between the intensive treatment group and the standard treatment group in the development of long-term complications of diabetes (4, 5). It is reasonable to expect that therapy achieving such glycemic control will provide similar benefits in terms of reduction in complications to patients with Type 2 diabetes (6).
The two major criticisms of the DCCT treatment protocol are that participants were so highly motivated that their experience cannot be generalized, and that the protocol would be prohibitively expensive to replicate. Additional medical resources of $4,000 to $5,800 per participant were invested annually in the intensively treated population (7). DCCT investigators believe that the reduction in the number of future complications may eventually help defray, but may not offset, the cost of the additional resources consumed in intensive management (4, 8–10).
The organizational and financial structure of managed care organizations (MCOs) makes them well-suited to provide a program of comprehensive preventive care and intensive treatment, similar in approach to that in the DCCT, to people with diabetes. Most plans provide some preventive services to people with diabetes but do not offer comprehensive diabetes management.Arecent Gallup survey found that, of health plans available to employees, only a third cover diabetes education classes, half cover annual eye exams, and half cover quarterly glycosylated hemoglobin tests (11).
Diabetes management strategies are more easily accomplished in the integrated setting of the MCO than in a feefor- service environment. However, because of competitive pressures and member turnover, MCOs must consider shortterm financial returns when making program decisions. The internal organizational changes that would be required to implement a comprehensive diabetes management program are perceived by many MCOs to be too expensive.
Diabetes NetCareSM is a comprehensive health care management program developed and provided by Diabetes Treatment Centers of America (DTCA, Nashville, TN) to improve the clinical outcomes of people with diabetes, while at the same time addressing the concerns of managed care regarding short-term savings. Diabetes NetCareSM uses a population-based approach that differs from other disease management programs in that it tracks the entire diabetic population within an MCO, integrating and coordinating all aspects of medical care for the population. The program was designed to replicate the essential elements of the DCCT in a manner that generates savings, improves clinical outcomes, and improves member and provider satisfaction in the short and long term.
Diabetes NetCareSM employs a multidisciplinary team that works with plan physicians and their patients to effect behavioral changes and to maintain desirable behaviors in the long term. Profiling tools are used to stratify physicians and hospitals based on the volume, costs, and outcomes of their diabetic patients, thereby allowing provider support coordinators to vary their assistance based on each provider’s needs. Provider support coordinators call on physicians and hospitals to determine what services they require to provide better diabetes care for their patients (for example, continuing medical education, nurse training seminars, and/or patient care conferences). Diabetes NetCareSM also facilitates the formation of a diabetes- focused, medical leadership panel for each plan, composed of primary care physicians and specialists with high volumes of diabetes patients. The panel reviews outcomes and quality issues and provides physician-to-physician interaction to strengthen the message of compliance with accepted practice standards.
To facilitate efficient patient support, Diabetes Net- CareSM stratifies members with diabetes into one of three levels based on the complexity of their disease.Whilemore resources are invested in the complex level 2 and 3 cases, all members are sent reminders about preventive screenings, physicians’ visits, and diagnostic tests, and are encouraged to participate in educational classes and seminars. All patients are also assigned to a diabetes nurse case manager who emphasizes effective self-management behaviors and proactively identifies those at risk for adverse events. Members who have an elevated HbA1c, comorbidities, and/or have been hospitalized (levels 2 and 3) are assigned to a complex case coordinator who works intensively with the physician and patient to alter risk factors including high blood pressure, cholesterol, and HbA1c values.
The nurse case managers are responsible for managing and integrating all the health care needs of their patients, not just those specifically related to diabetes. They manage the inpatient and outpatient resources required in caring for a patient with diabetes who has fractured an arm in a skiing accident, as well as the needs of one hospitalized for ketoacidosis. The member support component of the program provides each member with an advocate, a source of immediate information and support, and a guide through the health care system.
The program is supported by an infrastructure that includes an on-site administrative team, a clinical team, a provider support team, a management team, and an electronic tracking system. The electronic tracking system contains continuously updated patient and provider information. It is used to make resource allocation decisions that support the DTCA goals of quality improvement and efficient use of limited health care resources.
This study analyzes and compares baseline and follow-up clinical, financial, and member and provider satisfaction data from seven sites comprising (at follow-up) approximately 360,000 covered lives and 7,000 diabetic lives (5, 941 diabetic member years) to determine the short-term impact of implementing a comprehensive health care management program for the population with diabetes at managed care plans.
Materials and Methods
The plans participating in the study are all commercial health maintenance organizations (HMOs). For confidentiality reasons, results have been aggregated across plans, and the plans will only be referred to by number. Table 1 provides an overview of the participating plans. Plans are located in the southeast, midwest, south, and mid-atlantic regions of the United States. Each of the plans reimburses primary care physicians on a capitated basis; specialty physicians are paid per service provided. Six of the plans reimburse hospitals per diem, while one uses the diagnosis- related groups system for payment. Plan membership ranged from 30,000 to 87,500 members, and the plans had between 260 and 887 primary care physicians.
The dates of the baseline and follow-up periods for each plan are shown in Table 1. The beginning of each follow-up period corresponds to the implementation of the Diabetes NetCareSM program at each site; there was no phase-in period. The differences in the length of baseline and follow-up periods among the plans were unavoidable due to data availability issues and different program implementation dates. Aggregating the data and evaluating costs and outcomes by member month has minimized the impact of these differences.
Electronic tapes containing service utilization, laboratory, and pharmacy claims were collected for the baseline period and for each month of follow-up. Before implementation of the program,members with diabetes were identified from the electronic tapes based on having any of the following characteristics: 1) a record of taking insulin or other oral diabetic agents; or 2) an encounterwith the health care system specifically related to diabetes, as indicated by the presence of a diabetes-specific International Classification of Diseases (ICD)-9-CM code [codes 250 (diabetes mellitus); 250.0 (diabetes mellitus without mention of complication); 250.1 (diabetes with ketoacidosis); 250.3 (diabetes with other coma); 250.4 (diabetes with renal manifestations); 250.5 (diabetes with ophthalmic manifestations); 250.6 (diabetes with neurological manifestations); 250.7 (diabetes with peripheral circulatory disorders); 250.8 (diabetes with other specified manifestations); 250.9 (diabetes with unspecified complications); and 362.0 (diabetic retinopathy)]. Each identified diabetic patient was linked to his or her primary care physician. These “identified” patients make up the group of baseline patients.
In the follow-up period, each identified member’s primary care physician was contacted to confirm that the patient had diabetes. Information for those confirmed as having diabetes was stored in DTCA’s proprietary Electronic Medical Record (EMR) on each site’s computer server. Data stored in the EMR were updated regularly and included case notes from patient contacts, class attendance records, laboratory tests and results, notification of hospital admissions, specialist visits, and emergency room utilization. These “confirmed” patients make up thegroup of follow-up patients, along with all new diabetic plan members.
Chart reviews were performed to collect baseline clinical and laboratory data only for members confirmed as having diabetes. Chart reviews were not performed for patients who were identified as having diabetes during the baseline period, but could not be confirmed in the follow-up period.
Member and provider satisfaction data were collected and analyzed by Solution Point (Nashville, TN), a healthcare technology firm. Approximately 6 months after program initiation, Solution Point surveyed a sample of up to 150 physicians at each site who had treated patients in the program. Just before the program initiation, they surveyed a random sample of 300 plan members with diabetes at 6 sites to obtain baseline information.Asecond random sample of 300 plan members per site was surveyed approximately 9 months after program initiation to obtain follow-up member satisfaction data.
To quantify baseline costs, medical and pharmacy claims from the baseline period were analyzed for all members identified as having diabetes. Complete financial data were not available for Plans 6 and 7,
so they were excluded from the financial analysis. The total number of diabetic member months in the financial analysis is 54,186 in the baseline period and 55,879 in the follow-up period.
All costs have been converted to 1997 dollars using the Medical Care Consumer Price Index (calculated by the Bureau of Labor Statistics). The weighted average cost across plans per member month was calculated using the following formula: diabetic cost per member month!("i(mmi # diabetic costsi)/ "immi, where mmi represents member months in Plan i.
Weighted cost averages were determined in an analogous manner for the following service utilization categories: Inpatient, Outpatient, Physician, Pharmacy, and Other (includes emergency room visits, home health, ambulance, radiology, and laboratory services). Categories were identified by point-of-service codes on the claims data. Physician costs do not include primary care physician visits, which were covered under capitated contracts.
A reinsurance adjustment was made for inpatient stays to prevent the inclusion of costs not actually incurred by the plans. The plans in the financial analysis have reinsurance arrangements that protect them from incurring costs above a fixed ceiling for an individual patient over the course of the year. Any inpatient costs in excess of the ceiling in either the baseline or the follow-up period were excluded from the analysis. The unadjusted costs are reported in the Results section, to quantify the magnitude of excluded costs.
Follow-up costs were ascertained for the confirmed diabetic patients using medical and pharmacy claims data. As in the baseline, the number of diabetic member months was calculated, and the costs by service utilization category were weighted and aggregated across plans to determine overall diabetic cost per member month. A reinsurance adjustment was made for inpatient stays. Baseline and follow-up cost data for the nondiabetic population were evaluated in the same manner as for the diabetic population.
Baseline and follow-up diabetic hospital utilization (admissions and bed days) was tabulated per 1,000 member years. Utilization measures were calculated by tabulating the inpatient utilization at each plan, weighting each plan’s utilization by the number of diabetic member years, then averaging weighted utilization across plans, and multiplying by 1,000 to determine utilization per 1,000 diabetic member years. The
following formula was used for this purpose:
Hospital utilization per 1,000 member years ! 1,000 # "i(myi # utilizationi)/"imyi, where myi represents member years in Plan i. Hospital utilization data were not available for Plan 7, so it is excluded from this analysis. The total number of diabetic member months in the hospital utilization analysis is 66,062 in the baseline period and 59,930 in the follow-up period. All inpatient days were included in these estimates,regardless of whether or not they were covered by reinsurance.
Baseline and follow-up clinical data from the EMR were evaluated to determine the number of eye exams, HbA1c exams, foot exams, and cholesterol exams performed per member year. The numerator for these measures was the unique number of members receiving the exam of interest (e.g. if a member received two eye exams during the time frame, only one eye exam was counted), and the denominator was the number of member years represented in the time frame. For those who received at least two HbA1c tests in the follow-up period, a comparison was made between the numerical results of the first test and the numerical results of the second test. Multiple tests were not analyzed for foot exams, eye exams, or cholesterol screenings, because these tests are generally performed annually. The total number of diabetic member months in the clinical analysis is 50,328 in the baseline period and 71,295 in the follow- up period.
As indicated in Fig. 1, the percentage of members with diabetes per member year receiving at least one HbA1c test rose from 34% in the baseline period to 76% in the follow-up. The American Diabetes Association’s 1998 Standards of Care recommends HbA1c testing at least twice a year in patients who are meeting treatment goals and more frequently (quarterly assessment) in patients whose therapy has changed or who are not meeting glycemic goals (12). The percentage of diabetics per member year receiving an eye exam rose from 23% in the baseline to 40% in the followup. The American Diabetes Association recommends that comprehensive dilated eye and visual examinations be performed annually for all people ages 10 years and older who have had diabetes for 3–5 years, for all those diagnosed after age 30, and for anyone with visual symptoms and/or abnormalities (12). In addition, one of the National Committee for Quality Assurance’s (NCQA) Health Plan Employer Data and Information Set (HEDIS) measures is the percentage of continuously enrolled members with diabetes aged 31 and older who have had a retinal exam in the previous reporting year. According to the National Committee for Quality Assurance (NCQA), the national average for people with diabetes receiving retinal exams in 1996 was 38.4% (13). This number is overstated compared with the findings in this study because of NCQA’s restrictions on age and continuous enrollment.
The percentage of diabetic patients per member year receiving a foot exam rose from 2% in the baseline to 25% in the follow-up. The American Diabetes Association recommends that patients at risk be given a foot exam as part of their quarterly physical evaluation. They advise that patients with diabetes who are not at risk be routinely evaluated (12).
The percentage of diabetic patients receiving cholesterol tests also increased between baseline and follow-up. The percent of patients receiving a yearly cholesterol screening rose from 39% to 63%. The American Diabetes Association recommends that adult patients with diabetes be tested annually for lipid disorders, and that children within accepted risk levels be tested every 5 years (12).
For patients who received at least two HbA1c tests in the follow-up period, a comparison was made between the results of the first test and the results of the second test. As Fig. 2 indicates, the average time between tests was 107 days, and a drop in HbA1c from 8.9% to 8.5% was observed. These results do not account for potential variation in HbA1c values between labs. Given the lack of standardization across laboratories, confidence intervals around these HbA1c measures would not be meaningful.
Hospital utilization was measured via admissions and bed days per 1,000 diabetic member years. Admissions per 1,000 diabetic member years decreased from 239 in the baseline period to 196 in follow-up. Bed days fell from 1,336 days in the baseline to 1,047 days in follow-up. These findings translate into an average length of stay of 5.6 days in the baseline period, and 5.3 days in the follow-up period.
Figure 3 compares per member per month health care costs for the diabetic population in the baseline and the follow-up periods. Total costs are subdivided into five service utilization categories: Inpatient, Outpatient, Physician, Pharmacy, and Other.
Total costs decreased by $44 per diabetic member per month or 10.9%. Using the average plan size in this study of 1,000 people with diabetes, this would translate into gross savings of $528,000 in the first year of operation of the population management program. The largest decrease in costs for people with diabetes was attributable to the reduction in inpatient hospitalizations and bed days. Inpatient hospital costs fell by $47 per diabetic member per month.1
Continue reading and print entire article by clicking "Download/Print PDF" on right, top.