OBJECTIVE: A new diabetes risk score (DRS) can accurately and precisely predict development of type 2 diabetes (T2D)and may improve patient outcome if routinely used as part of population screening for at risk patients. The DRS uses algorithms with panels of 5-9 serum biomarkers detected by an ultrasensitive molecular counting technology. We sought to evaluate the cost-effectiveness of DRS. METHODS: Using data from a five-year prospective validation study, we developed a ten year cost-effectiveness model of DRS. Dependent upon the subjects score, we modeled testing and treatment patterns, costs and outcomes of the fraction of the population who converted to T2D, those that did not convert and those that died. All comparisons were made to a no-testing strategy. Cost inputs including hypothetical reimbursement for the test, preventive treatment, costs of disease for T2D, preference weights and mortality were obtained from the literature. A United States health care system perspective was taken. Univariate sensitivity analyses and scenario analyses were performed. RESULTS: We simulated a cohort of patients over age 40 with a BMI of 25 or greater. We assumed 12.5% of tested patients would develop T2D with no intervention over five years. Annual mortality for T2D is 1.08%. Base case simulation showed that the DRS is cost effective at five years ($9000/QALY) and at ten years ($400/QALY) compared to a no-testing strategy. Sensitivity analyses suggest that the incremental cost-effectiveness ratio is most sensitive to the cost and effectiveness of preventive treatment and the pre-test risk of converting to T2D. Using DRS to identify patients at high risk was cost effective up to a reimbursement level of $1100 per test. CONCLUSION: Under a variety of scenarios, early prediction of type 2 diabetes with a serum biomarker panel is cost-effective given certain assumptions regarding screening efficiency and availability and effectiveness of preventive treatment.