Improved understanding of the pathogenesis of type 1 diabetes mellitus has completely changed our view of this disease in the past 25 years-from an acute, fulminant disease, to a chronic, autoimmune process. Information on genetic and serologic markers has increased our ability to identify individuals at risk. Prospectively gathered data indicate that, with a combination of immunologic and metabolic studies, children with a 6-year risk of disease higher than 90% can be identified due to an ongoing immune process. They differ from children with overt disease only in the time it will take for glucose levels to rise above a diagnostic threshold. Therapies to change the progression of beta-cell loss have been tested in patients with newly diagnosed type 1 diabetes. With improved predictive capabilities and agents that can have longer-lasting effects than those tested more than 10 years ago, new prevention studies are underway. These studies are large and costly but the risks posed by such interventions compare favorably with those of developing hyperglycemia and of future complications portended by the diagnosis of diabetes. In this Review we discuss risk-stratification techniques and how they are applied, other diagnostic criteria, and outcomes from diabetes-prevention trials.