EHR migration 2026: what boards should demand before approving
Nine-figure Epic conversions are back on hospital capital agendas. Here is a decision framework for weighing migration against optimization, TEFCA-era exchange, and two to three years of clinical disruption.
EHR migration 2026 has quietly become one of the largest line items on hospital capital agendas. Becker's Hospital Review is now maintaining three overlapping trackers: a running list of roughly 19 hospitals and health systems moving to Epic, a broader list of organizations switching platforms across vendors, and a ranked view of the most expensive EHR projects underway this year. The existence of a cost-ranked list is itself the signal. When conversions are big enough to rank by price tag, they belong in front of the board, not just the IT steering committee.
Why replacement cycles restarted after a decade of consolidation
The current wave is not a single vendor announcement. It is churn across several categories at once. Systems that completed acquisitions during the past three years are now paying the integration bill, and running two or three record systems across a combined footprint is untenable for clinical safety, revenue cycle and reporting. A second group is being pushed by product roadmaps: platforms nearing sunset, or hosting arrangements that no longer fit a system's risk appetite. A third group is migrating for strategic reasons, betting that a larger vendor ecosystem is the fastest route to ambient documentation, coding automation and payer-facing data products.
The distinction matters because it determines what a successful migration looks like. An M&A-driven consolidation is judged on standardization and the retirement of legacy contracts. A roadmap-driven move is judged on capability delivery over five years. Boards that approve a conversion without naming which of these it is end up with a project that cannot be evaluated, because nobody agreed on the scoreboard.
When EHR conversions are big enough to rank by price tag, they belong in front of the board, not just the IT steering committee.
The interoperability floor keeps rising under every migration decision
HHS announced on June 26, 2026 that the TEFCA network had crossed one billion health records exchanged, after officials reported nearly 500 million in March 2026. The Social Security Administration joined TEFCA in February 2026 to speed disability determinations, which extends exchange obligations well beyond provider-to-provider referral traffic. Federal agencies, not just competitors down the road, are now counterparties to a health system's data-exchange posture.
That changes the diligence question. It is no longer enough to ask whether a candidate platform is technically capable of national exchange. Leaders should ask how the system will perform against volumes that doubled in roughly three months, what identity matching and record-location performance looks like in production, and who absorbs the operational load when non-clinical requesters such as benefits agencies join the queue. A conversion decided in 2026 locks in that posture for close to a decade.
Migration versus optimization: the comparison most business cases skip
The weakest EHR business cases compare a new platform against the incumbent as it exists today, unoptimized and under-resourced. That is not a fair fight. Before approving a conversion, boards should require a costed optimization alternative: what would three years of the same capital and analyst capacity buy inside the current system, in build cleanup, workflow redesign, integration work and clinician training?
Sometimes the honest answer is that optimization cannot close the gap, particularly where the incumbent product is being retired or where multi-instance sprawl is the core problem. But the comparison forces precision. It also surfaces the hidden variable in every conversion: internal change-management capacity. A migration consumes clinical governance bandwidth for two to three years, which means every other transformation initiative competing for physician attention gets deferred, whether or not anyone says so out loud.
Budgeting the productivity dip instead of hoping through it
Go-live productivity declines are predictable and routinely underfunded. Ambulatory volumes drop for weeks, documentation time rises, coding accuracy wobbles and denials follow on a lag. Leaders should treat this as a financed cost with an explicit reserve, a schedule assumption for reduced clinic templates, and a communicated recovery curve rather than an operational surprise that lands on service line leaders.
AI adds a newer wrinkle. New EHR builds are where predictive and generative tools get embedded, and ONC's May 2026 data brief on hospital use, evaluation and governance of predictive AI found governance maturity trailing adoption. A conversion is the cheapest moment to install model inventory, local validation and monitoring requirements into the build standard. It is also the moment those requirements are most likely to be traded away for schedule.
Five questions to put in the board packet
First, what specific problem is this migration solving, and would optimizing the incumbent solve it more cheaply? Second, what is the total five-year cost including internal labor, dual-running, third-party integrations and the productivity reserve, not just the vendor contract? Third, how does the platform improve measured data exchange performance under TEFCA-era volumes? Fourth, which strategic initiatives are being deferred to free up clinical change capacity, and who signed off on that trade? Fifth, what AI governance requirements are non-negotiable in the build?
None of these questions requires a vendor scoreboard. They require a system to state its own theory of value and accept accountability for it. Given that a conversion is the largest single IT capital commitment most health systems will make this decade, that is a low bar to set.
