Why VA's electronic health record mega-project is failing

The agency's former top technology official explains how management problems, reliance on outmoded development methodology and lack of customer buy-in are combining to sink (for now) a critical modernization project at the Department of Veterans Affairs.

VA facility in Cincinnati 2020 Editorial credit: Jonathan Weiss / Shutterstock.com

The Department of Veterans Affairs currently is reassessing its $16 billion-plus Electronic Health Record Modernization (EHRM) effort. Faced with productivity and patient safety issues in its initial pilot site, further rollouts of the EHRM have been paused by VA Secretary Denis McDonough. And while VA has not yet announced the details of the actions it will take to correct the EHRM program, those actions must include addressing the program's fundamental problems, not just the readily apparent quality and productivity problems surfaced at the pilot site.

First, most government programs of the scale of the EHRM fail, for a variety of reasons that the Government Accountability Office has documented in numerous reports. Even those programs that succeed do so after substantial struggles, often more than doubling their original schedule and budget estimates. In its first three years, the EHRM program is clearly on the path to failure. It has missed nearly every deadline it has set, and its budget estimate has expanded from $16 billion to $21 billion. Externally, there is no indication that VA has studied and implemented the various GAO-recommended governance, organization, accountability or execution measures intended to help agencies avoid large-scale failures, or that VA heeded governance and execution advice from private-sector health care systems that had previously pursued EHR transitions.

Second, VA's track record with large-scale systems development and deployment has been abysmal. Over the past 20 years, it has suffered many failed programs costing taxpayers hundreds of millions of dollars. To try to address this track record, VA's Office of Information and Technology, which is responsible for most of the large systems development programs, adopted disciplined incremental and agile program management approaches about a decade ago, resulting in substantially improved results. Unfortunately, the EHRM program is not managed by OIT. Instead, it is being executed by a separate program management organization in the traditional "waterfall" approach that contributed to so many VA failures.

Third, VA's EHRM program management organization exists outside both the business owner (the Veterans Health Administration) and the department's technology (OIT) organizations. The result is a lack of accountability for the success of the program by the two organizations most impacted by its success or failure. As the initial pilot of the EHRM system held in Spokane, Wash., has shown, the ability of the new EHR to fully support veteran care will be the primary determinant of program success. Yet VHA, the recipient and arbiter of that success, is only tangentially accountable. As part of its program reassessment, VA must determine how to better involve VHA and OIT, including making their leadership directly accountable and responsible for program success or failure.

Fourth, the primary customer lacks buy-in. The primary recipient of the EHRM program is the VHA, one of the nation's largest health care systems. The EHRM program will replace VA's current EHR (CPRS/VistA) with a newer system from Cerner Corporation -- the same system being installed by the Department of Defense. The rational for the investment in the new EHR has been that the current system is old and costly to maintain and that there is a need for better exchange of data with the Department of Defense. Missing from that rationale has been design elements that would benefit VHA and its staff, such as increased quality of care, more efficient work processes, increases in patient safety or reductions in the cost of care. Combined with the lack of accountability caused by the separate program office, this has made VHA a skeptical party to the EHRM implementation.

Fifth, VA has ignored change management realities. Both the 2014 and 2016 Medscape surveys reported that VA's current EHR, CPRS/VistA, was the most highly rated EHR system by clinicians nationwide. In those same surveys, Cerner, the basis for the new EHRM, was rated in the middle of the pack. From a change management perspective, this indicates a huge challenge facing the EHRM program

VHA has over 350,000 staff primarily concerned with providing medical care for veterans. Virtually every medical process used at VA is supported by the current EHR and must be supported by the new EHR. If, from the medical staff's perspective, the new EHR is harder to use, less efficient or more error prone, then working with it will make their job harder, as has been experienced in private-sector EHR implementations for years. Staff must take the time to double-check their work, ensure that the right information is being captured and that the right action is being taken to ensure that the quality of care for each veteran is not adversely affected. As a result, productivity suffers, patient safety is jeopardized with reports of errors and "near misses" and questions are raised about the adequacy of the training that staff has received, as has been seen in the outcomes of the VA EHRM's initial pilot site.

Sixth, VA is hiding from the truth. Much like Kevin Bacon in "Animal House," VA management has been saying, "Remain calm, all is well," for three years, to Congress, to the public and to themselves -- even though internal staff has been well aware of numerous problem indicators. Hiding from the truth has been a cultural problem at VA for many years and has been a cause or a contributor to many of VA's very public failings. In perhaps its most difficult task, VA must correct this problem, so that it can deal with the true state of the EHRM program, if it is to have any chance at success.

Large programs like VA's EHRM are difficult and failure-prone even if everything is done well. VA has to do a realistic assessment of itself and the program. Is VA capable of succeeding at a program of this scale? What are the alternative solutions? What is the likelihood of failure? What will cause the program to fail? What can be done to mitigate demonstrated and potential causes of failure? Who is accountable for program success? What is the benefit to veterans? What is the benefit to staff? What measures can be used to show these benefits are being realized? What is all of this really going to cost?

And please, let's be honest with ourselves, with veterans, with Congress and with the public this time. Even if the answers are bad, dealing with reality is the first step towards getting the program to succeed.