Science

VA has the data and DOE has the technological expertise in Big Data analytics, artificial intelligence and high-performance computing.

One program the two agencies are moving forward together is MVP (Million Veteran Program). MVP, a VA’s innovative genomics program, has already enrolled more than half a million veteran volunteers. These veterans have provided DNA samples; completed surveys about their health, lifestyle and military experiences; and granted secure access to their electronic health records for research purposes. The VA and DOE are also working multiple studies as part of the Big Data Science Initiative. Initial projects are focusing on algorithms to generate highly tailored risk scores for suicide; on prostate cancer and innovative approaches to identify lethal versus nonlethal cancer and require treatment, as opposed to slow-growing cancers that are unlikely to cause any symptoms; and identification of individual risk factors to predict certain forms of cardiovascular disease, to then inform individualized therapy and treatments for patients.

Under this partnership, we are working on many other projects with the VA. One of the projects consist on assisting with aligning the VA clinical analytics roadmap with the Cerner acquisition roadmap and Cerner capabilities. The VA is acquiring the Cerner Corporation (Cerner) Electronic Health Record (EHR) as a modernization of the VA Electronic Health Record. This acquisition will deepen the need for advanced analytics functionality at the VA that can operate on existing data from the VA’s Veterans Health Information Systems and Technology Architecture (VistA), on new clinical data from the future deployment of Cerner, and the secondary archival data in the Corporate Data Warehouse.

Another effort the VA and ORNL are working on is part of a plan to build a comprehensive safety surveillance system within the EHRM Safe HIT framework. We are working together to develop evidence-based models and frameworks to later develop automated, near real-time and real-time systems that will reveal safety concerns and alert users to hazards, both predictable and unpredictable, unintended consequences and opportunities to improve HIT and safety.
Because of the nature of the data hosted by the VA, open source technologies are not always an option. The VA and ORNL are doing research (and developing) an open source test bed, with the goal of a future potential implementation. This required understanding the needs, development of policy and governance, design and implementation of infrastructure, and open source capability evaluation and analysis. The current focus is on Natural Language Processing, but also data manipulation, Artificial Intelligence, and machine learning.