The primary purpose of this study is to determine whether the use of a computer assisted decision support (CADS)system by primary care providers (PCPs) for their patients with type 2 diabetes mellitus (T2DM) changes the quality of care relative to a "usual care" group in terms of objective outcome measures of glycemic control (e.g., A1C, mean blood glucose, frequency of hypoglycemic episodes) and in terms of subjective ratings by patients.

Diabetes accounts for an enormous fraction of the cost of health care in the United States and presents a major burden on Military Medical Facilities for care of retirees and dependents. There are insufficient endocrinologists and other diabetes specialists to manage all patients with diabetes mellitus (DM) and a significant fraction of these patients have less than optimal control (hemoglobin A1C's [A1Cs] over 7%). Multiple barriers prevent the necessary improvement in glycemic control that would result in savings in lives and costs. The implementation of a telemedicine and web-based approach for patients to send their blood glucose data which, when combined with relevant laboratory, pharmacy, and A1C targets as set individually for each patient by the Primary Care Physician (PCP), triggers a clinical decision support system (DSS) for the providers can be expected to improve quality of care and efficiency of care. The computer assisted decision support (CADS) system has been integrated with the Comprehensive Diabetes Management Program (CDMP), a web-based, multi-platform, interactive patient and provider tool which is currently operative in the Walter Reed Health Care System (WRHCS), Wilford Hall Medical Center (WHMC) at Lackland Air Force Base (AFB), and five community clinics affiliated with the University of Hawaii (UH). This existing infrastructure permits CADS to be tested in a multiple sites that are geographically diverse with diverse patient populations.

This study will test the safety and efficacy of CADS as used by PCPs in a multi-site, ethnically and geographically diverse study in a 12-month, open, prospective, cluster-randomized, controlled clinical trial. The specific aims of the study are to: (1) monitor the impact of the intervention on: a) measures of glycemic control, b) the number of diabetes -related hospitalizations and emergency room visits, c) the control of co-morbidities, hyperlipidemia and hypertension, d) the number of clinic visits, e) the change in the patients' quality of life as a result of the intervention; and (2) evaluate the PCPs' satisfaction with the technology.

We will employ a cluster-randomized, controlled, clinical trial involving 30 PCPs who will each recruit approximately 19 patients from their respective geographic site. After completion of recruitment, PCPs and their patients will be randomly assigned to 1 of 2 "treatment" categories: CADS, or "Usual Care". Input data for use by the CADS system will come from the electronic medical record (laboratory and pharmacy data) and from the PCP who will set goals for each individual patient's glycemic control. Patients will upload blood glucose data through a modem to a password-protected, secure server at least every 2 weeks and receive modification in their treatment regimen at least every three months from their PCP, based in part on the recommendations provided by the CADS system to the PCP. We will compare quantitative outcome measures of glycemic control (the primary outcome is the change in the patient's A1C), blood pressure, and lipid levels from the two treatment groups. In addition, subjective qualitative data from the patients and providers will be obtained.

Inclusion Criteria:

Providers

1. Non-specialist physician (general internist or family practitioner), nurse practitioner, physicians assistant
2. Absence of orders for deployment or permanent change of station
3. Willingness to recruit up to 19 patients prior to randomization
4. Willingness to deliver "usual care" as defined below in Section 6.3.2.

Patients

• Patients with a diagnosis of Type 2 DM of at least three months duration;
• History of inadequate glycemic control (A1C < 7.0% but ≤ 11%) determined by at least two A1C values within the previous 6 months
• Willingness to test blood glucose four or more times each day at least once a week, and 8 times each day for one day per month;
• Willingness to upload their glucometer to transmit data to a central database every 2 weeks
• Access to a land line telephone in order to upload their glucometer data every 2 weeks
• Not using "prandial" regular or a short-acting insulin (e.g. insulin aspart, lispro, glulisine)
• Not using "premixed insulin" (Novolin Mix, Novolog Mix, Humulin Mix, Humalog Mix)
• 18 years of age or older
• Stable medical status (e.g., no myocardial infarction, stroke, major surgery, or major mental illness during the previous six months)
• Not taking or not expected to be taking any oral glucocorticoids except for replacement therapy for those with adrenal insufficiency, amphetamines, anabolic, or weight-reducing agents during the course of the study
• Not receiving chemotherapy or immunosuppressive therapy access to telephone capable of transmitting data for downloading of a glucose meter

Exclusion Criteria:

Providers

1. Specialist physicians
2. Orders for deployment or permanent change of station
3. Unwillingness to recruit up to 19 patients prior to randomization
4. No prior experience with management of type 2 diabetes in adults
5. Unwillingness to deliver "usual care" as defined below

Patients

1. Patients with Type 1 diabetes or those with Type 2 on prandial rapid-acting insulin or premixed insulins
2. Unwillingness to test blood glucose four or more times a day at least once a week and 8 times a day once a month;
3. Unwillingness or inability to receive training in using the technology and/or upload blood glucose data every 2 weeks.
4. No access to a land line telephone in order to upload their glucometer data every 2 week
5. Inability to communicate in written and spoken English
6. Organ (kidney, pancreas, liver) transplant recipients
7. Severe impairment of dexterity, vision, or intellectual function
8. Pregnancy verified by a urine pregnancy test at baseline in pre-menopausal women. The pregnancy test will be administered by the project officer at each study site.
9. Individuals who are not likely to return for the follow-up because they or their sponsors are likely to have a permanent change of station or termination of service before completion of the protocol