Personalized prediction of diabetes outcomes
Principal Investigator: Dr Gerard Lipowski
Approved Research ID: 1781
Approval date: March 3rd 2014
We wish to develop a computational method for personalized risk prediction of diabetes (both prevalent and incident disease) and its evolution over time, in relation to other available health outcomes and life conditions. This project requires data only (not samples) including baseline data on lifestyle factors (e.g. related to nutrition, sleep, mental condition, physical activity including cardiorespiratory fitness, etc.), genetic data (when available) as well as longitudinal follow-up data on the health outcomes (incidence and mortality) as and when sufficient data are available. The first phase of the project will be exploratory in nature and will be quantifying associations between all available physiological, mental, lifestyle, dietary, environmental, socioeconomic (& genetic and biomarkers data when available) and select the variables, which alone or in combinations, most significantly predict incidence and evolution of diabetes outcomes. Confirmation of known associations will be used as control for our methodology. Then, we will analyse which variables contribute most for the risk and outcomes differences between individuals. Depending on the availability of the reassessment data, we will attempt to explore the causality by analysing the time of occurrence of various changes in life style and of diabetes outcome. Finally, the significant associations will be used for personalized prediction of risk factors and protective factors for diabetes outcomes. We expect our methodology to be applicable for prediction of other health outcomes, which will be subject to submission and approval of further applications. Our results will contribute to the development of personalized prevention and diagnostics of chronic diseases, and will benefit the society by providing knowledge for more informed choices for healthy lifestyle.