Predicting Risk for Diabetes: Choosing (or Building) the Right Model

Whether to screen for type 2 diabetes mellitus in asymptomatic adults is controversial (1). No trial has established that systematic screening for diabetes and early treatment improve health outcomes more than treatment after routine clinical diagnosis (2). The U.S. Preventive Services Task Force (USPSTF) recently recommended testing adults for diabetes if they have symptoms of diabetes or evidence of diabetic complications, including vascular disease (2). It also recommended screening adults for diabetes if they have sustained blood pressure greater than 135/80 mm Hg, because lowering blood pressure below conventional target values in patients with diabetes reduces the incidence of cardiovascular events and mortality (2). Although the ongoing ADDITION (Anglo-Danish-Dutch Study of Intensive Treatment in People with Screen Detected Diabetes in Primary Care) (3) may settle the benefits of screening and early treatment, the USPSTF currently recommends against screening asymptomatic adults for diabetes, largely because evidence of benefit is lacking (2).

If we do not screen asymptomatic adults for diabetes, we miss an opportunity to identify those who are at increased risk for diabetes because they have dysglycemia (impaired glucose tolerance [IGT] or impaired fasting glucose [IFG]). Randomized, controlled clinical trials from Asia, Europe, and North America have conclusively shown that treating people who have dysglycemia with lifestyle interventions (4–6) or medication (6–8) can delay or prevent type 2 diabetes. The USPSTF recognized that intensive lifestyle modification programs can reduce the incidence of diabetes (2). It concluded, however, that the recommendation to eat a healthful diet, be active, and maintain a healthy weight applies to everyone, not just those at high risk for diabetes (2). To reduce the burden of diabetes, the USPSTF would target entire populations of healthy people rather than persons at especially high risk (2).

Either approach benefits from accurate …