Speaker: Peter Banthorpe, Cathryn Lewis
detailed health and lifestyle data. a study based on half a million lives in the uk biobank cohort
RGA have sponsored a ground-breaking research project at King’s College London. The research focuses on predicting mortality and major morbidity (e.g. Cancer, Heart Disease, Stroke) outcomes in the UK Biobank cohort (UKB) of 500,000 participants aged between 40-69 years, all of whom have undertaken detailed medical history and lifestyle questionnaires, shared complete medical records, provided blood, urine and saliva samples for genotyping and biometric analysis, and agreed to have their health followed indefinitely. The data offers a uniquely powerful resource to answer the following questions: 1.How accurately can the risk of morbidity and major morbidity be estimated using multivariable prediction models based on detailed phenotypic information (medical history, physiology, behavioural and lifestyle risk factors)? 2.Can such prediction models be significantly improved – both in statistical and clinical / absolute terms – by including genotypic data, in the form of polygenic risk scores? Although a positive family history is a risk factor for developing disease, it cannot entirely and exclusively capture an individual’s genetic risk. The polygenic model of human phenotypes has long been theorised, and the results from genome-wide association studies (GWAS) have now revealed that much of the genetic basis for complex traits comprises small effects of hundreds or thousands of variants. Thus, genomic prediction models that consider all genetic variants may more accurately predict risk of complex disease. This approach is known as genomic or polygenic risk scoring and the application of polygenic risk scores (PRS) is now demonstrating that prediction is becoming sufficiently accurate to stratify individuals into very different risk profiles. This research is of sizeable benefit to Actuaries, but also to Clinical and Public Health researchers.