Big Data

Widespread adoption of electronic healthcare records (EHR), digital technology, and high-throughput laboratory –omic assays, combined with advances in analytic methods (“Big Data”) promise to revolutionize the scale (breadth, depth, and duration) and efficiency of large-scale cardiovascular research. The appropriate use of such methods could transform our understanding of the causes and consequences, prevention and treatment of cardiovascular disease. For such promise to become reality requires new collaborations across academic disciplines and organizational boundaries; new methods for data acquisition, analysis and ethical data sharing; and a new cadre of scientists and skilled in the application of foundational methods of data acquisition and analysis (including machine learning) to cardiovascular research. Critically, this strategy requires access to large, rich biomedical datasets (including routine clinical data and prospective cohort studies) combined with scientific use cases. Oxford already has substantial strengths in all these aspects, and is increasing them further through its new initiatives in the area. 

Enhanced prospective observational studies: Oxford pioneered the establishment of very large cohort studies in diverse populations (UK Biobank; China Kadoorie Biobank; Mexico Prospective Cohort) in order to study a wide range of lifestyle and genetic determinants of cardiovascular disease reliably. These studies increasingly depend on novel Big Data approaches to electronic health care (EHR) records, patient-oriented smartphone technology, remote sensors of behaviour and function, multi-modal imaging, and high-throughput genomic assays. Over the coming years, our work will address the challenge of converting such large multi-dimensional data into meaningful cardiovascular phenotypes. Mendelian randomization, PheWAS, and TreWAS will be used to explore the phenotypic correlates of genetically-determined differences, to identify disease pathways, and to assess their impact on cardiovascular risk. 

21^st Century randomized controlled trials: Oxford’s approach involves using EHR data for more efficient patient recruitment and long-term outcome follow-up, protocol-specific software engineering to enhance protocol adherence, digital technology to capture physical and cognitive function, and central statistical monitoring to enhance site performance. As a consequence, our previous trials (e.g. HPS, SHARP, THRIVE, REVEAL) have been large enough and long enough to assess treatment effects reliably while costing less than one tenth of standard industry prices. We are working closely with NHS Digital to develop even more efficient systems for recruitment in large randomized trials (e.g. the ORION-4 trial of 6 monthly injections of the novel siRNA inclisiran in established atherosclerotic disease is a key exemplar in the UK Government’s Life Sciences Industrial Strategy). Efficient data-enabled trials will only be possible with better Good Clinical Practice (GCP) regulations. Building on our past experience with the Clinical Trial Transformation Initiative, we are working with the European Society of Cardiology, Wellcome Trust, Bill & Melinda Gates Foundation, and Academy of Medical Sciences to develop and establish innovative GCP guidelines based on key scientific principles. 

Big Data Institute (BDI): Oxford’s new BDI provides a central facility for interdisciplinary research and training based on the combination of medical and data sciences. Inter-disciplinary cardiovascular research is enabled by BDI scientists involved in genomics, image analysis, digital phenotyping, and EHR acquisition for research. Oxford is one of 6 substantive Health Data Research UK sites, providing further opportunities for collaborative research, teaching and sharing data, tools, software, and knowledge with the research community. BDI statisticians are also Fellows at the Alan Turing Institute, of which Oxford is a founding partner.  A flexible high-performance facility in the BDI provides shared specialist computing for genomics and machine learning with the adjacent Wellcome Centre for Human Genetics. Ethical expertise is provided by the co-located Wellcome Centre for Ethics and Humanities which promotes proportionate and ethically robust approaches.