Predictors of liver disease progression and mortality
Principal Investigator: Dr Tom Pembroke
Approved Research ID: 51103
Approval date: October 25th 2019
Liver disease is increasing in the UK and is now a major cause of loss of work and early death. Many liver diseases are linked to alcohol excess, obesity/diabetes, drug use and previous exposure to blood borne viruses and are the result of damage over many years. Current attempts to tackle liver disease in its early phases are limited by a lack of epidemiological data and poor understanding of the natural history of these diseases in the general population. We aim to assess the number of cases of liver disease, liver failure, liver cancers and death in the UKbiobank population using routinely collated GP, hospital and mortality data. A key aim is to establish factors that are associated with the progression of liver disease between these stages. We shall link the progression of liver disease to baseline data from the UKbiobank including factors such as routine blood tests and physical measurements, other significant illnesses and genetic data. Machine learning approaches are advanced statistical models that enable computer programmes to progressively improve their performance to make predictions and optimise the analysis of datasets. We shall apply cutting edge machine learning techniques to identify high risk groups where interventions will have the greatest benefit. This research will improve our understanding of which patients are the key to screen for liver disease and generate new ideas for potential therapies at the optimum time point. It is estimated that this project will take 3 years to complete.