OBJECTIVE: To measure well-being and treatment satisfaction and their correlates in older people with diabetes. RESEARCH DESIGN AND METHODS: A postal survey was conducted of 1,000 diabetic patients aged > or = 60 years, representing 56% of the resident older diabetic population in an inner-city health district with a largely indigenous population of 230,000 people and a widely varied socioeconomic mix. Well-being and treatment satisfaction were measured with diabetes-specific instruments and correlated with patient data held in a central register. RESULTS: There was an 81% response. The general well-being scores (median [interquartile range]) for patients on diet alone, tablets, and insulin were 54 (44-60), 53 (42-61), and 48 (35-56) (P < 0.001 comparing insulin with diet and tablets) compared with a scale maximum of 66. Treatment satisfaction scores were 35 (31-36), 35 (32-36), and 34 (30-36) (P < 0.001 comparing insulin with diet and tablets), scale maximum 36. Mean HbA1c concentrations were 5.0 +/- 1.4% (for patients on diet alone), 5.8 +/- 1.6% (tablets), and 6.6 +/- 1.7% (insulin) (P < 0.001 for each difference). Neither well-being nor treatment satisfaction correlated with HbA1c. Insulin-treated patients were younger and had been diabetic longer than non-insulin-treated patients; their well-being remained slightly, but significantly, lower when adjusted for age, sex, BMI, and diabetes duration, but treatment satisfaction was no longer significantly different. Women had lower well-being than men. CONCLUSIONS: It has proved possible to measure well-being and treatment satisfaction in a large community-based samples of older people with diabetes. At the level of glycemic control in this population, neither parameter correlated with HbA1c. The lower well-being in insulin-treated patients remained significant in multivariate analysis.