Phenome-wide association study and prediction of two-years mortality
Professor Erik Ingelsson
WARNING: the interactive features of this website use CSS3, which your browser does not support. To use the full features of this website, please update your browser.
Several epidemiological studies have examined the association between risk factors and overall mortality. Most of these studies, however, focus on a risk factor at a time. The UK Biobank project represents a unique opportunity for an unbiased comprehensive assessment of risk factors for short-term mortality in the general population.
This research project aims to investigate the association between a variety of exposures assessed in the UK Biobank and two-years mortality. Moreover, we aim to build and validate a risk score for prediction of two-years mortality using data mining techniques. To maximize the chance to discover new predictors, we would like to include as many data points as possible, both from in-person testing measurements, as well as questionnaires. However, no genetic or biochemistry assessment would be included, as these variables will not become available until 2015. This also implies that no biological samples are analysed, but we will only use information already collected in the UK Biobank.
The primary analyses will be performed on all-cause mortality. In secondary analyses, we will also test the score on specific causes of death (cancer mortality; cardiovascular mortality; other causes). The full UK Biobank cohort is going to be used for this project, independently of the health status of the participants. However, health information at baseline (e.g. if the participants have been admitted to hospital or have experienced a major health disorder) will be included in the analysis to improve the ability to obtain accurate predictions.
We hope to find unexpected predictors leading to new biological knowledge and new avenues for primary prevention, and to improve the possibilities to target individuals at highest risk for medical intervention. We plan to include the full cohort in our analysis