Predictive power of the UK biobank for individual health outcome stratification: a multi dimensional machine learning study
Knowing in advance who is at risk of developing a serious illness is one of the most important goals of medical research. It is however a challenging goal, because multiple factors contribute to an individual's risk of developing any specific disease.
The UK Biobank is an excellent resource for trying to address this problem - a wide range of information is available about an enormous number of individuals. Using this information we aim to develop mathematical models to decide how great an individual's risk of developing certain common diseases (type 2 diabetes, cancer, myocardial infarction / stroke, chronic bronchitis/emphysema, and dementia) in the near future is. We will also try to understand what factors make developing common diseases more likely.
In this project in addition to personal information and medical history we will specifically study the effect of physical activity levels and blood test results on predicting future disease.
Once developed our models will be used in a public health education platform where any individual can input their own data and receive an approximate idea of their future health risk