Disease progression and treatment response in data-driven subgroups of type 2 diabetes compared with models based on simple clinical features: An analysis using clinical trial data
The Lancet Diabetes & Endocrinology May 22, 2019
Dennis JM, et al. - Five diabetes subgroups have been proposed according to diabetes progression differences and complications risk using data-driven cluster analysis, so researchers compared the clinical usefulness of this subgroup-based tactic to predict patient outcomes vs developing models for each outcome using simple patient characteristics. In the ADOPT trial (n=4,351), five clusters were identified using the same data-driven cluster analysis that Ahlqvist and colleagues reported. In the trial data, clusters identified were like those described by Ahlqvist and colleagues in the original study. Investigators found differences in the incidence of chronic kidney disease between clusters, however, they found a better time predictor for chronic kidney disease in estimated glomerular filtration rate at baseline. Data-driven clusters proposed differ in the progression and response of diabetes treatment, but models based on simple continuous clinical features are more useful for stratifying patients. This finding suggests that, rather than assigning patients to subgroups, type 2 diabetes precision medicine will be of most clinical use if it is based on a tactic of using specific phenotypic measures to forecast specific outcomes.
-
Exclusive Write-ups & Webinars by KOLs
-
Daily Quiz by specialty
-
Paid Market Research Surveys
-
Case discussions, News & Journals' summaries