Achieving Data Integration & Interoperability at the VA

As the largest unified healthcare provider in the U.S., Veterans Affairs (VA) manages upwards of 78 billion health records, a volume so massive it requires around 50 terabytes (equivalent to about 850,000 hours of music) of data storage. Furthermore, the data “spans 21 clinical domains of patient health records, which includes lab results; pharmacy prescriptions; inpatient and outpatient diagnoses and procedures; and other Veteran medical data.”

Until quite recently, VA and military patient information was stored across over 560 different data sources. As a result, tracking down a patient’s full medical history was, if not impossible, highly time consuming. However, as the VA continues to lean into its long-anticipated digital transformation, this is now changing. 

Over the past year the VA has launched a number of high profile data integration initiatives:

• The joint health information exchange (HIE). Developed in partnership with the DoD, this project aims to improve interoperability between the two agencies and their healthcare partners by linking together their respective Cerner Enterprise Health Record (EHR) systems. 
• The Veterans Data Integration and Federation Enterprise Platform (VDIF EP). Created in partnership with Intersystems HealthShare, the goal of this project is to provide the VA’s external healthcare partners secure access to its EHR data via a single, unified view. 

As with any healthcare related data analytics initiative, the primary goal is to provide practitioners high quality data that enhances decision making and leads to better patient outcomes. However, data interoperability and integration is only the first, pivotal step in the VA’s ongoing transformation from a healthcare provider to a medical intelligence innovator.

Similar to many public enterprises, the VA seeks to leverage its formidable enterprise data infrastructure to develop new advanced analytics and AI-powered products, services and policies. In fact, in 2019, the VA launched its National Artificial Intelligence Institute (NAII) to spearhead the development of such products.

As of now, the NAII is focused on 5 key initiatives designed to not only democratize data and AI development within the VA, but harness the power of open data sources to drive innovation within the world at large:

• Flagship AI projects. Leverages deep learning, explainable AI, privacy-preserving AI, and AI for multi-scale time series to enhance veteran-facing services and programs. 
• AI tech sprints and partnerships. An ecosystem of public and private datasets that can be accessed and used by federal agencies, industry, academia, and nonprofit organizations to support AI R&D.
• Affiliation hub. Centralized framework for cataloguing and tracking AI projects across the VA. 
• AI@VA Community. A community of researchers, developers and experts dedicated to exploring artificial intelligence, machine learning, and other next gen technologies, and the impacts they will have on Veterans and their families. Members receive a periodic newsletter highlighting impactful AI work across the VA, updates on the NAII, opportunities to collaborate on bleeding-edge projects, and invitations for upcoming AI webinars and meetings.
• AI Education. Develops and delivers AI-related training to a new generation of AI scientists, practitioners, and clinicians. 

However, this is only the beginning. In the upcoming months, the VA expects to share its bigger, long-term data strategy. 

“The vision we're driving toward is using data — managed data, integrated data, quality data — to support and strengthen VA’s journey as a learning enterprise. Both in continuous improvement through operational decision support at every echelon and evidence-based policymaking to continually improve how we serve veterans,” Veterans Affairs Chief Data Officer Kshemendra Paul shared at a recent FedInsider event. 

With that in mind, expect to hear more about the VA’s approach to AI and advanced analytics down the line.