Diabetes mellitus (DM) affect 5% to 10% of Americans and is estimated to account for $45 billion in direct and $47 billion in indirect costs. The average medical care cost in 1992 for a person with diabetes was $11,157, compared with $2,600 for a person without diabetes. The Diabetes Control and Complications Trial (DCCT) demonstrated that good control of diabetes can delay the onset and slow the progression of many diabetic complications and thereby result in avoidance of costs related to such complications. First Chicago NBD evaluated a worksite-based patient education program for employees with DM. The goal of the program was to determine if such a program could result in improved DM control. A total of 53 employees participated in baseline laboratory testing and met monthly with a diabetic health educator. After 3 months, 45 (85%) employees participated in retesting. One of the 45 employees had an active infection and was not included in the analysis. Of the 44 remaining employees, 43% were on oral agents, 39% on insulin, 2% on combination therapy, and 16% controlled with diet alone. Prior to receiving any laboratory results, 48% of participants rated their level of control as "good" or "very good", while only 9% considered it "poor". After 3 months of educational programs, the subjects' mean fasting blood glucose levels fell from 197.8 mg% to 179.6 mg% (P = 0.12), mean glycohemoglobin declined from 11.5% to 10.1% (P < 0.001), and mean hemoglobin A1C declined from 9.0% to 8.3% (P < 0.001). A worksite-based diabetes disease education program has been shown to significantly improve control of the disease. This should result in lower direct and indirect health care costs and enhanced quality of life.