Veterans Deserve Better than VistA
Notwithstanding the nostalgic throwback to vinyl records, there is little chance record albums are going to replace Siri streaming Spotify over Sonos anytime soon. Or that “Hey, Record Player” is going to replace “Hey, Alexa.” You can scream at your record player all you want, and it will not scour the Internet for that Chili Peppers song of your youth. Your record player cannot use AI to make recommendations of other songs or artists you might like. It will not listen to an audio sample and identify the song, band, year of production and where you can buy it. It will not form a playlist based on your past listening patterns. You can’t put your records in your pocket or take them to the gym; and you can’t access them anytime you want from the cloud, on any device. So, as nostalgic as listening to vinyl records may be… they are just old. They are yesterday’s technology.
In computing, old technology does not get better, it just gets older. There is no nostalgia for a Commodore 64 any more than there’s nostalgia for a VCR or Fax machine. Old technology gets more complex. It gets harder and more expensive to maintain. It gets less secure. It becomes less interoperable with modern systems. Less mobility. Less self-service. Fewer features. Less performant. Less analytics, artificial intelligence and machine learning. And no cloud.
Now let’s consider the Department of Veterans Affairs (“VA”) and its current Electronic Health Record Modernization (“EHRM”) project. Most of the VA is still using an Electronic Health Record (“EHR”) system called Veterans Health Information Systems and Technology Architecture (“VistA”). VistA was conceived in the 1970s and deployed in the 1980s. It was developed on the MUMPS database, which itself dates to the 1960s. It is programmed in an obsolete programming language with a limited programmer-base. It is essentially a product of the Carter Administration and is objectively long past its prime. In short, it’s old technology and is all but assured to perpetuate incomplete patient records and fail to meet today’s health challenges facing our veterans. It can’t deliver predictive treatment or care, and it will cost more and deliver less with each passing day.
What’s worse, VistA is not a single system, it is 130 separate systems implemented at autonomous medical centers across the country—that have grown in different directions over decades—i.e., VistA has been repeatedly “forked.” The consequence of “forking” is it is extremely hard to upgrade because there is no single set of source-code from which to work. It is very difficult to secure because of its large vulnerable surface area. It was never designed to outwardly connect to the Internet. And interoperating with other modern systems is very difficult because each “instance” is an island unto itself with its own numerous unique dependencies.
The fact is VistA is objectively at the end of its life. Multiple independent assessments including MITRE’s 2016 Assessment H, the 2016 Commission on Care and GAO reports laid out the need to replace VistA. The 2016 MITRE report made clear that VistA’s ability to deliver new capabilities was stalled and that it was in danger of becoming obsolete.[1] VistA’s security vulnerabilities and concerns that cyber-attackers could impersonate providers or change a patient’s record are well documented.[2]
Further, the VA has tried to modernize VistA several times in the past through efforts such as HealtheVet, iEHR and VistA Evolution. Those efforts have not worked and showcased the challenges of modernizing VistA at just a single site. The fundamental flaw of these previous modernization attempts is that they focused on building on top of a patchwork of variable, undefined, and under documented VistA systems. The truth is there is not enough time or money to ever make VistA work like a modern system.
In 2017, the VA recognized a new approach was needed, and by 2018 it had contracted with Cerner to supply its commercial Millennium EHR to replace VistA. The Cerner system is the same system the Department of Defense—including the Coast Guard—began using a few years earlier, and which will be fully deployed, on schedule and on budget, across DoD at the end of this year.
This is a key point because for decades the DoD and VA attempted to build interoperable but separate EHR systems. Those efforts also failed, which meant that when a service member left active-duty service and began receiving care at VA, their health history did not seamlessly transfer over. Paper was involved. Vague, error-prone memories of past care. Fax machines. Lots of images on old CDs.
Regardless of what system is used, the men and women—whether active duty or retired—who serve the United States deserve a single, interoperable, longitudinal health record that follows them all the way from basic training to assisted living. DoD originally decided to move to Cerner and then the decision was made for the VA to follow suit. The interoperability problem was solved, and a seamless, longitudinal record would be a reality for retired service members receiving care at VA facilities.
With the goal of finally providing this seamless record, the VA undertook what is now the largest health IT modernization project in history. It involves not only updating VA’s EHR system from VistA, but also integrating the new EHR with numerous DoD and VA dependent systems like DEERS which is run by DoD to provide eligibility information. It involves standardizing procedures and workflows that may have been different across 130 VistA implementations at largely autonomous VA medical centers. And it requires training.
The problem with modernization is it doesn’t come with a magic wand and there’s no easy button. Modernization requires change and some short-term pain for the long-term benefits of a modern technology infrastructure, a modern user interface, and a modern set of workflows. Moving from customized, one-off workflows to standard, commercial-off-the-shelf workflows is always hard to implement … and always worth doing. A modernization project of this scale and scope necessarily involves time to untangle the decades of customized processes established in support of VistA, which inevitably involves challenges. This should be a surprise to no one.
And for sure the rollout of the Cerner EHR across the VA has not been without its problems. In hindsight, rolling-out the first deployment in October 2020—at the height of the pandemic—was a mistake. It is understandable that overworked healthcare providers focused intently on patient care, being directed to invest time, effort and training to learn a new EHR system was justifiably frustrating while facing an onslaught of COVID patients.
Yet we need to keep in mind that the same Cerner technology rolled out across the VA is being successfully rolled out across the DoD. And, by the way, this is the same system successfully deployed in thousands of commercial and public hospitals around the world. Search the media for Cerner EHR system problems and you will find precious few. So, if it works at the DoD and at hospitals across the United States, then why can’t it work at the VA? Of course, it can—and it is.
But in June 2022, something else happened … Cerner was acquired by Oracle. Unlike Cerner, Oracle is a hyperscale cloud provider and major enterprise applications vendor. Oracle runs many of the world’s most mission critical enterprise systems. Oracle is an engineering company with 100 percent of our effort focused on system performance and security. Oracle is orders of magnitude larger and more resourced than Cerner alone. So, VA now has essentially two vendors for the price of one—one with extensive clinical expertise and one with extensive engineering expertise. And Oracle has committed to deliver the full program scope within budget and on time.
But there’s more … because Oracle is also committed to delivering an entirely new, modern, next generation web EHR application as a free upgrade—and again within the same budget envelope. We are in the process of a progressive rewrite of the Millennium EHR into a modern, stateless web application. In plain language that means we intend to move Millennium to a cloud-based application that will have a modern web-based user interface, be mobile friendly, allow for voice recognition and include machine learning-based clinical decision support and analytics. Oracle has committed to delivering a beta version of the new EHR in 2023, and again—we will provide this next generation EHR at no extra cost to the VA or DoD, simply as an upgrade to the currently contracted system.
In the meantime, Oracle is hard at work to stabilize and improve performance; make fixes to functionality and design issues; improve training and build a better user experience. Oracle is working on all these items and steadily delivering to VA updates and fixes. We have supplemented Cerner’s engineering team with our deep technical expertise and will continue to add resources as necessary to get the job done. We have a live dashboard tracking Congressional requested items, and are committed to transparency as can be seen by this end-of-year report. A constant dialogue between Oracle and VA is in place in order to prioritize critical work, and much has already been accomplished. For example, we have accelerated and delivered the critical pharmacy modules by two full years because it was a priority.
We are also actively fixing issues like training that have been a source of frustration for VA providers, and we recently announced that we added Accenture consulting to the team to help revamp the training process to make it more efficient, applicable and useful. We have engaged with the Veterans Service Organizations, and we are in direct contact with VA medical centers.
And we are going above and beyond. For example, we have introduced an all-new opioid advisor tool. This tool has already alerted providers in the VA system more than 1,600 times and the DoD more than 17,700 times, representing more than 19,000 prescriptions where a provider could make a better choice for patient safety regarding opioid use.
* * * * *
With all of that said—and with the very meaningful actions taken and progress made over the past six months as detailed in our 2022 Year End Report—there are still a few Members of Congress who would inexplicably prefer to pull the plug on the EHRM program and abandon the promise of interoperable systems made to our veterans. The recently introduced bill, H.R. 608, The EHRM Termination Act, directs the VA to abandon the EHRM program and revert to VistA. This proposal fails at every level of analysis and puts our veterans at significant risk. To be clear, there is no state in which the EHRM program is terminated, and the VA does not revert to VistA. This precise question has been subject to endless debate and analysis and there is no third way forward for the VA despite what the proponents of H.R. 608 may believe.
The fact is the VA is already four years and $4 billion into modernizing the VA EHR. Thirty-five years of VA legacy health data including nearly 24 billion health records for 23.5 million patients, has been migrated from VA to the central database running the new EHR. Nine hundred standardized new workflows in the EHR have been established between DoD, VA and the Coast Guard, and more than 11,000 veterans and service members have now utilized both a DoD and VA facility running the new EHR. Add to this Oracle’s promise to leapfrog private care institutions and deploy the first modern, stateless web EHR to the VA for our veterans’ benefit.
Congressional opponents of the EHRM cite instances of patient harm, problems with system stability, or users who don’t like the new system. Don’t get us wrong, even one instance of patient harm is unacceptable. The infamous “unknown queue” is no longer an issue … raise concerns and we will fix them. Issues related to workflows and complexity are being worked through in partnership with the VA. And performance gains are being achieved with each passing day with substantial resources. Already, we have made significant improvements to the system’s capacity and performance, reducing the most severe outage incidents by 67 percent since last June.
As stated, one of the biggest reasons the VA version of Millennium is so unwieldy is because there were 130 different ways of doing things across 130 different instances of VistA. In many cases these divergent workflows were combined—superset—in the EHRM with the hope that there would be something for everyone. Instead, with hindsight, it made the workflows too cumbersome. For example, there are dozens of options for a provider to order a liver enzyme test when in a typical commercial Millennium system there are four or five options for the most used tests with the option to search for additional tests if needed. Leadership at VA is working with us now to simplify and standardize these workflows, which will make the user experience much easier and standard across VA. Issues such as workflow efficiency are to be expected in a modernization effort of this scale, but they are by no means a reason to pull the plug. Let’s just get it right.
H.R. 608 is the wrong approach and will take VA and healthcare for our nation’s veterans backward. VistA cannot be made better; it will just be made older, less secure and more expensive to operate. Worse, it will deny our veterans the many benefits modern health care technology has to offer. It is undeniable that substantial progress is being made, and there is no reason to start over for the sake of starting over and reverting to VistA; yet that is precisely what H.R. 608 proposes to do.
Oracle expects to be held accountable for delivering on our commitment to the men and women who currently serve and have served in the United States Armed Forces. Our vision is straight-forward and achievable. With the ability to carry your EHR in your pocket and interact with your care provider as easily as you interact with a Starbucks barista, bring your own health record (BYOHR) will be as simple as pulling out your mobile device. Care professionals deserve modern voice interfaces which reduce fatigue and human errors. Sick patients deserve the benefits of modern analytics to make sure they are getting the best possible treatment alternatives for their care situation. Predictive technology should be employed to flag risk of suicide and catch opioid addiction. It is unclear what vision perpetuating VistA advances.
We are working collaboratively with our dedicated partners in the VA to drive this effort forward. We are proud of the work we are doing, and we intend to exceed expectations. Our veterans deserve no less. Pulling the plug now ensures snatching defeat from the jaws of victory.
[1] The MITRE Corporation, Independent Assessment of the Health Care Delivery Systems and Management Processes of the Department of Veterans Affairs, Assessment H (Health Information Technology), A-35, accessed March 31, 2016, http://www.va.gov/opa/choiceact/documents/assessments/Assessment_H_Health_Information_Technology.pdf.
[2] A Flaw in the VA’s Medical Records Platform May Put Patients at Risk I WIRED, Aug. 13, 2022; Letter from Hon. Tony Gonzales to VA, Sept. 6, 2022, https://gonzales.house.gov/sites/evo-subsites/gonzales.house.gov/files/evo-media-document/09.06.22%20Veterans%20Affairs%20VistA%20letter.pdf