In 1982 we introduced a program designed to prevent recurrent diabetic ketoacidosis (RDKA), based on the assumption that diabetes education in conjunction with appropriate use of and adherence to insulin therapy should eliminate all RDKA. A hierarchical set of medical, educational, and psychosocial interventions was used. The rate of RDKA was lower after initiation of the program even though patients seen during this period came from lower socioeconomic and more one-parent families (each independently associated with RDKA) than patients seen prior to intervention. In 44 patients with a history of RDKA, insulin omission was documented in 31, inadequate education in 13. Overall, the rate of RDKA decreased from a prereferral mean of 25.2 episodes to a postreferral mean of 2.6 episodes per 100 patient-months (P less than 0.0001). Metabolic control improved after intervention as documented by a decrease in mean hemoglobin A1 from 14.1% to 10.7% (P less than 0.0001). RDKA ceased whether or not psychotherapy was used. Although RDKA is causally related to a variety of social, economic, and family dysfunctions, its prevention requires recognition that its proximate cause is omission of insulin and assurance that a support system exists to ensure adherence.