Models for risk assessment. TASK LEADER – MSG The concept of a medical risk score is to determine the weight that should be applied to each data variable on a patient to predict likelihood of developing disease (or complications) given their present and historic health status. Although some have faith in a stringent risk assessment approach, research over the years has not shown this to be a particularly productive route, other than reinforcing the obvious or what is known from epidemiological studies. 25xxxx://xxx.xx.xxx.xx/en/Publicationsandstatistics/Publications/PublicationsPolicyAndGuidance/DH_4003246 26xxxx://xxx.xxxxxxxx.xxx.xx/ 27xxxx://xxx.xxxx.xxx..uk/ The purpose of this task is to study existing assessment diabetes models as well as comparative models from the literature such as Framingham and UKPDS for cardiovascular events28. Unfortunately their data comes from epidemiological studies, and it has never really been shown, that the modification of a risk factor (like stopping smoking) really modifies the risk to the extent as calculated by these formulas. The task will survey the different readily available risk assessment scores for persons with diabetes and their evaluations. Based on the published literature, methodology used to develop models will be systematically assessed and accuracy of models for wide spectrum of outcomes in different populations will be systematically assessed. Most promising methods, parameters and models will be identified and provide basis for the further development of multi-parametric risk assessment in the REACTION platform. Predictions from individualized physiology-based in-silico models will be used to rationally (and non-linearly) transform patient “raw data” and will be integrated into to the analysis as additional covariates. A special focus shall lie on physiology-based transformations of patient input like exercise status, food uptake and other relevant varying non-biomarker factors. Most promising models will be implemented, and if needed externally evaluated with additional already published studies and with publicly available data sets like Diabetes Control and Complications Trial (DCCT) for a wide variety of outcomes of interest. During the project, a huge amount of laboratory and clinical data will be collected from different sources and integrated. Those data will together with readily publicly available data build a basis for further development and refinement of the REACTION risk engine. Finally, the task will implement the ...
Models for risk assessment. Study existing assessment diabetes models as well as comparative models. Survey the different readily available risk assessment scores for persons with diabetes and implement the most promising models. The description and material related to the study of the Mechanistic whole-body model of insulin-glucose interaction and kinetics will be included in D1.4.1 Periodic report for the Commission at M12.
