I can recall two awful fumbles:

The former cost the Dallas Cowboys a game. The latter has cost taxpayers billions.

Recently, the VA announced it’ll increase funding again for its ongoing EHR modernization rollout—adding an additional $17B, extending its contract with Oracle Health by up to three years, and bringing the total to $27 billion. I’ve written about this fumble in detail, and I’ll write about it again today:

In this article, I’ll catch you up on the VA’s Electronic Health Record Modernization Program fumble, highlight the recent funding updates, and offer my take from a systems perspective.

The modernization effort began in 2018, when the VA signed a 10-year, $10 billion contract with Cerner (now Oracle) to replace CPRS across the country. The goal was to put the VA and Department of Defense on the same EHR, giving service members one longitudinal health record as they transitioned into veteran care. And to see why we need to move away from CPRS, just watch this training video.

Cerner became Oracle Health after Oracle acquired the company in 2022, inheriting one of the largest and most fumbled health IT projects in the country.

The first VA medical center went live in Spokane, Washington, in October 2020. Four more followed in 2022, and the joint VA–Department of Defense facility in North Chicago went live in 2024.

Progress slowed because the new EHR created serious problems at the sites using it.

By April 2023, the VA paused most new deployments so it could stabilize the system at the existing sites. At that point, the project had already expanded beyond its original budget, clinicians were frustrated, and Congress was questioning whether the rollout should continue (me, too).

The VA resumed deployments in April 2026 with four Michigan medical centers. Four more sites in Ohio and Kentucky followed in June, and three Indiana hospitals went live in August.

That brings the current total to 17 of 164 VA medical centers using the modernized Federal EHR. In other words, roughly 10% of VA medical centers are live, while 147 centers—about 90%—still need to transition.

The next scheduled deployments are the Alaska VA Healthcare System and Louis Stokes Cleveland VA Medical Center in October 2026. The VA plans to add another 26 medical centers in 2027 and complete the nationwide rollout by 2031.

The project is now eight years old. I’ve gone from college graduation through medical school, residency, and into fellowship during that span. Got married. Had a kid. The VA still has 147 medical centers left to convert.

The VA has recently modified its agreement with Oracle Health to add nearly $17 billion to the contract. That brings the contract’s potential value from just under $10 billion in 2018 to roughly $27 billion now (or by the end of 2031, when everything is set to be done). The modification also gives the VA up to three additional one-year ordering periods, potentially extending Oracle’s work through May 2031.

This modification increases how much the agency can spend under the contract as it deploys the EHR to additional medical centers and pays for continued maintenance, updates, infrastructure, and technical support. The $17 billion represents additional spending authority under the contract, rather than an upfront payment to Oracle.

And $27 billion may still understate the program’s eventual cost. A 2024 VA Office of Inspector General report cited an Institute for Defense Analyses estimate that the total could reach $49.8 billion. That projection included $32.7 billion for implementation and another $17.1 billion for sustainment during the rollout and for 15 years after every site goes live.

Note, though, that the $27 billion figure and the $50 billion estimate measure different pieces of the project. The former is the potential ceiling on Oracle’s contract while the latter attempts to capture implementation and long-term operating costs across the broader program. Together, they show how far the financial exposure may extend beyond the current contract headline.

Either way, a $27 billion ceiling makes this one of the largest health IT projects in the country.

The patient-safety concerns of continued use of CPRS as the VA’s standard EHR are easy to understand once you’ve worked inside it, which I have during training. Orders can be entered incorrectly or missed, trends are difficult, nearly impossible, to review, and important information can get buried across disconnected screens. The system relies heavily on clinicians remembering where to look and what to double-check.

Thankfully, the VA has a strong Swiss cheese model around many clinical decisions. A medication order may be reviewed by a pharmacist, nurse, and other members of the care team before it ever reaches the patient. A modernized EHR should strengthen those safeguards by surfacing errors earlier, organizing information clearly, and making the correct action easier for the clinician. For example, in Epic, if I order furosemide for someone who has a serum potassium value of 3.0, I’ll get a “stop” alert asking me if I’m sure I want to order this, as furosemide will likely lower this patient’s potassium even more. The current CPRS does not have this modern safeguard.

One of the greatest advantages of the VA, though, is that it operates as a nationally integrated healthcare system. Veterans can receive care across VA facilities while their clinical data stays within the same broader network. Connecting that shared data to a modern EHR could improve coordination at the bedside, support population-health work, and create an even stronger foundation for research.

A lot of great studies already come from the VA because it cares for a large population over long periods of time. Better structured data could help researchers study clinical outcomes, medication safety, chronic disease, and healthcare delivery at a scale that few other systems can match.

Clinicians will face a learning curve as they move away from CPRS. We handle new clinical protocols, new devices, and new evidence throughout our careers, so we can learn a new EHR too, especially when the training, workflow design, and technical support are built around how care is actually delivered.

In summary, modernizing the VA’s EHR is imperative for patient safety and for the future of veteran care. The rollout has been expensive and poorly executed, but the underlying goal remains worth pursuing.