Validation of the CDC-RTI diabetes cost-effectiveness model

Abstract

The purpose of this study was to assess the validity of the CDC-RTI Diabetes Cost-Effectiveness Model by comparing rates of diabetes incidence and complications to existing published results. We performed 47 internal and external validation exercises comparing the model simulated outcomes with the outcomes from 24 published trials. To simulate the outcomes for each published study, we input a cohort with similar baseline characteristics and treatment and then modeled the development of diabetes and its complications for the same follow-up duration as in the trial. Outcomes measured included diabetes incidence, renal disease, neuropathy, retinopathy, cardiovascular disease, and mortality. The results of our model simulations were generally close to published outcomes. To determine how well our model was able to simulate the published outcomes, we ran three sets of simple regressions (actual outcome = b0 + b1 × simulated outcome)— one for the internal validation studies, one for the external validation studies, and one for the external validation studies of diabetes incidence. For the 17 internal validation analyses, the R2 value was 0.992 and the slope of the regression line was 1.001. For the 24 external validation analyses that did not include diabetes incidence, the R2 value was 0.969 and the slope of the regression line was 0.991; the six external validation analyses of diabetes incidence had an R2 value of 0.913. In conclusion, the CDC-RTI Diabetes Cost-Effectiveness Model accurately models the development and progression of diabetes and can be used to evaluate the cost-effectiveness of potential diabetes prevention and treatment programs.