Cloud: From “Impossible” to “Done”

In an extremely significant and groundbreaking accomplishment, the Department of Veterans Affairs’ (VA) Electronic Health Record system called “VistA” will have been completely migrated to the cloud by July 2024. Moving all of VA’s VistA to the cloud was deemed impossible when the first of VistA’s 130+ on-premises instances moved to VA’s Enterprise Cloud in 2019. By July 2024, it will be done.

What is VA’s VistA?
VistA stands for “Veterans Health Information Systems and Technology Architecture”. VistA is the electronic health record (EHR) system used by VA’s Veterans Health Administration (VHA). With over 400,000 employees, VHA is America’s largest integrated health care system serving the health needs of 9-million Veterans. Because of VistA’s impact to millions of patients and thousands of doctors and nurses, a change to VistA can send tremors through America’s healthcare system. Completing VistA’s cloud migration is a positive earthquake about to happen.

VA’s VistA Cloud Migration Journey
While serving as VA’s Office of Information and Technology (OIT) Deputy Assistant Secretary for DevOps in 2019, I launched the program to begin VistA’s migration to VA’s cloud infrastructure.  Although VistA may sound like a single application, it is in fact a complex ecosystem of over 130 different instances of VistA software comprising 15 million lines of code through nearly 200 integrated applications accessing petabytes of lab, imaging, and text notes data. Day in, day out, 24x7x365, VistA safely and reliably performs hundreds of millions of clinical, administrative, and financial transactions which support hundreds of thousands of Veteran care encounters annually. As complex and critical as VistA is, VA/OIT deserves credit for its sustainment efforts. Very few programs this large operate this smoothly.

Since 2019, VA/OIT has been working closely, carefully, and quietly with VHA to migrate VistA instances out of datacenters and into VA’s Enterprise Cloud. VA’s OIT leadership team, most notably, Reggie Cummings, Kendall Krebs, Dave Catanoso, Sue Speck, Mike Giurbino, Eric Jurasas, and a cross functional team of subject matter experts (government civilian and contractor) like Liz Congdon, Brad Stephens, Tom Spinelli, and many others across all OIT pillars have made this challenge look easy and its unpredictability look rote. By July 2024, all the VistA instances will have migrated to the cloud. This is a BIG deal for VistA, VA, and Veterans.

Opportunities Provided by VA’s VistA Cloud Migration

• Cost avoidance through consolidation of VistA instances. Over the years, VistA has multiplied into 130+ custom variations (instances) to serve Veterans’ needs in different geographic areas and clinicians’ needs with different specialties and patient demographics. These variations have led to an ever-increasing annual cost of maintenance by VA/OIT government civilians and contractor staff to the tune of many hundreds of million dollars per year. Once VistA instances have been “cloudified”, VistA can begin to be stepwise consolidated, thereby reducing VistA’s annual software maintenance/sustainment costs. Given the complexity of VistA’s ecosystem, this stepwise consolidation must be carefully planned and performed by private sector and government experts who know the VistA code, data, workflows, interfaces, and dependent legacy software and SaaS applications.
• Greater standardization of VHA Care delivery. Beyond the opportunity for software cost avoidance, VistA consolidation offers a platform to accelerate standardization of clinical capabilities across VA’s massive 172 Medical Centers and 1,100+ Outpatient Clinics healthcare system.  In addition, starting with the MISSION Act and throughout the COVID pandemic, there has been an ever-growing operational need for VHA to be more interoperable, not only within VA’s own Medical Centers and Clinics, but also with healthcare provided through VA’s external Community Care Network partners.  As the 130+ VistA instances are consolidated, VHA will be able to realize the power and efficiency of a more operationally integrated healthcare system resulting in more efficient care delivery to Veterans both in and out of VA’s network, providing the best possible access to care for our Nation’s Veterans.
• Insights from data analytics & AI. VA has some of the richest, deepest, and most diverse health-related data sets in the world. Unfortunately, when stored in geographically dispersed datacenters, these data are often difficult to locate and access by VA staff, researchers, and other VA software applications. By aggregating VistA and other VA datasets in the cloud, secure seamless access, deduplication, and conflict reconciliation of data will be enabled, regardless of the physical location of the data consumer. VA’s cloud infrastructure provides a secure, authenticated environment for collaborative research, efficient data management, and scalable computational resources thereby enabling VA to get a better return on its research investment. Plus, VistA in VA’s Enterprise Cloud will make it easier for VA to provide accessible data and a flexible computational platform for insightful analytics, innovative artificial intelligence, and machine learning. As an example, at the recent HIMSS24 conference, MITRE Corporation presented its preliminary findings from internally funded research into Open-Source VistA code. The research was titled, “Revolutionizing Software Modernization: The Role of Generative AI in Transforming VistA” by Jacob Zimmer under the guidance of Sarah Corley, MD, FACP with support from MITRE’s VA team led by Dr. Troy Mueller. This was just a glimpse into the potential of Large Language Models applied to software modernization of VistA’s large and very complex application ecosystem.

Recommended Next Steps:

• In July, VA should celebrate this VistA cloud migration milestone by recognizing the leadership, persistence, patience, and OIT/VHA collaboration that enabled this achievement (perhaps at the VA Texas Valley Coastal Bend Healthcare System, where it began in June 2019.) VA should reward all the government civilians who contributed to this successful outcome through their application of managerial, technical, organizational, and people skills.
• With the explosion of Veteran Community Care costs (estimated at $30B for FY25), VistA advancements will enable VHA to better track its in-network appointments and be more interoperable with external Community Care Network provider partners.  As such, Veterans will receive better and faster access to care, and VHA will further optimize its clinical capacity, resulting in lower Community Care expenditures.
• Identify savings. Stepwise consolidation and deprecation of VistA instances can save VA millions of dollars per year in sustainment costs. Given VistA’s very complex ecosystem, consolidation must be carefully planned by private sector and VA experts who know VistA code, data, workflows, interfaces, and dependent software applications.
  For example, what are the first two VistA instances to consolidate and why? These factors need to be considered: local VA medical facilities and Veteran population demographics, medical specialties, geography, size, dependencies, functional similarities, external interfaces to SaaS modules and legacy applications, disaster recovery, cybersecurity, observability, change management, timing. Take no giant leaps. Consolidate through very carefully planned baby steps.
• Build interoperability. The principal rationale behind VA Secretary Shulkin’s 2017 EHR decision was to solve “interoperability with DoD but also with our academic affiliates and community partners, many of whom are on different IT platforms”, not to award a sole-source multi-Billion-dollar contract to recreate VistA functionality. Now that VistA’s EHR data will be accessible in the cloud, VA should incentivize the private sector to build secure, open, standards-based, bidirectional interoperability bridges to VA’s VistA Electronic Health Record.
• Virtualize CPRS. CPRS stands for Computerized Patient Record System. It is a graphical front-end user interface to VistA which allows VA healthcare providers to enter, review, and continuously update all information related to any patient including medical history, diagnoses, medications, lab results, etc. CPRS runs in Microsoft Windows on a PC. Updates and patches to CPRS must be pushed out to thousands of PCs in VA. Managing these CPRS deployments is a difficult, expensive, time-consuming task. To simplify this process, VA should accelerate its virtualization of CPRS in Amazon AppStream and possibly Azure Virtual Desktop as well.
• Sponsor innovation. VA should sponsor VistA Tech Sprints to enhance VistA innovation and interoperability. The recent AI Tech Sprints for Ambient Dictation and Document Processing are a great start, but VHA can go farther by sponsoring Tech Sprints focused on VistA. VistA Tech Sprints could make available VA VistA software code and synthetic data sets for research, collaboration, and innovation by industry, academia, and nonprofit organizations with input from VistA experts and clinicians. Winners of these VistA Tech Sprints could be rewarded with SaaS contracts which would enable those innovations to be applied immediately in a controlled VA setting for testing and potentially, adoption across VHA.

VA’s migration of VistA to the cloud is significant.

Achieving this milestone will unlock significant savings for taxpayers and significant benefits for VA healthcare providers and their Veteran patients.  This is an opportunity for VA to switch gears from a VistA maintenance mindset to a VistA innovation and interoperability mindset.