A tool to predict the risk of mortality in type 2 diabetes has been developed using a cohort of 33,067 patients with type 2 diabetes obtained from the Cleveland Clinical electronic health record. A Cox proportional hazards regression model was created using medication class and 20 variables selected because of their association with mortality. A prediction tool was then developed based on Cox model coefficients. Follow-up of the study ranged from 1 day to 8.2 years (median of 28.6 months) with 3,661 deaths recorded. The concordance index for survival prediction was 0.752, which indicates the tool has good discrimination ability. The calibration curve showed that the prediction did not significantly overestimate or underestimate mortality risk. The authors concluded by suggesting that incorporating medications into mortality prediction charts in patients with type 2 diabetes should improve treatment decisions.
Documentation centre