Type 2 diabetes (T2D) is a major public health problem. The prevalence of the disease is growing at an alarming rate and the sharp increase in T2D shows no signs of slowing down. The increased number of T2D cases worldwide has caused a simultaneous increase in the number of T2D related complications. This paper demonstrates how a survival tree based approach can be used to enable predictions to be made concerning when an individual is expected to experience a complication of T2D. A survival tree is used to identify cohorts of individuals with significantly different survival distributions from T2D to complication. By fitting appropriate survival distributions to the individual leaves of the tree, the expected time until complication can be calculated for each group of individuals. Survival trees were built for death, stroke/acute myocardial infarction (AMI) and amputation/coronary revascularisation.
|Name||IEEE 29th International Symposium on Computer-Based Medical Systems (CBMS)|