Abstract

The revolutions in various scientific fields are making it possible to rethink our approach to how we conduct large-scale trials and observational studies. These include exponential growth in computing power, the availability of electronic health record (EHR) data and low cost genetic and other “omic” testing. We can now use larger amounts of and more complex data to answer questions in population science. The Million Veteran Program (MVP) and our pragmatic trial program are examples of how this is being executed within the VA Healthcare System. With a current enrollment of 575,000 participants, MVP is the largest database of health, lifestyle, military exposure and genetic information in the United States, and is expected to enroll one million veterans. This longitudinal cohort serves as a resource for cutting-edge precision medicine research, such as understanding the genetic risk for diseases and genetic basis for differential response to treatment (pharmacogenomics), and identification of molecular targets for new treatments. Participants provide a blood sample for genomic and other “omic” research, complete health, lifestyle and military exposure surveys, grant access to their electronic medical records and agree to be contacted for potential participation in future research. MVP enrollment is conducted at approximately 60 sites. Three coordinating centers and a toll-free information (call) center support site recruitment. Enrollment is monitored by Genomic Information System for Integrated Science (GenISIS), a secure IT/Informatics infrastructure, which also serves as the data warehouse, and provides a secure scientific computational environment where researchers conduct their analyses. Concurrently, the Veterans Informatics and Computing Infrastructure (VINCI) serves up the VA electronic medical record data for this research.

We are collaborating with the Department of Energy (DOE) to jointly leverage the VA’s vast array of clinical and genomic data with DOE ’s computer science and exa-scale computing expertise and infrastructure. A collaboration with the Department of Defense (DOD) will allow us to enroll recently separated veterans and active duty military personnel in MVP , starting with the Millennium Cohort Study enrollees beginning in 2017. Eight test projects using MVP genoype and clinical data in schizophrenia, bipolar disorder, Gulf War Illness, PTSD , multi-substance abuse, cardiovascular disease, metabolic disease, chronic kidney disease, and age-related macular degeneration are underway. I will present some lessons learned in curation of our extensive data and early GWAS results.

Our pragmatic trial program has also launched a trial comparing two antihypertensive medications without recruitment sites. Participants are identified using EHR data and contacted by phone. If they are willing to participate, they are enrolled by phone. Progress of the study is followed by EHR .These are two of several examples of how big health data are used to conduct large-scale modern epidemiology within a large healthcare system. In collaboration with many partners we will be pursuing innovations in big data analytics, statistics and mathematics, computational biology and computer science to support this endeavor. This will involve interaction with other global leaders working in this new field of big data population science and groups with similar data assets such as the UKBiobank

This talk will be hosted by Professor John Danesh, Professor of Epidemiology and Medicine, Head of Department of Public Health and Primary Care