A US software campus in Verona, Wisconsin, is an unlikely place to feel a tremor in NHS trust boardrooms, but that is roughly what happened on Friday. Sumit Rana, president of Epic Systems and the man widely tipped to succeed 82 year old founder Judy Faulkner, told staff his last day would be August 14. He wrote of his father's death last November, of a mother in India who needs him present, of wanting to give more of himself to his wife and children. It is a personal letter, not a corporate memo, and there is nothing in it that should trouble anyone outside Epic's Wisconsin campus. Except that Epic is no longer just an American company's internal affair. It is now woven into the operational core of the NHS, and its succession question is a question the health service has no seat at the table to answer.

The scale of that dependence has crept up quietly. Epic's NHS market share rose by 3.5 percentage points to 9.7% in the last year alone, the fastest growth of any supplier in the electronic patient record market. It has moved beyond its early foothold in flagship trusts such as Cambridge and Guy's and St Thomas' into more rural and mixed geographies, most recently a £222 million federated contract across Somerset and Dorset that will retire legacy systems from System C, TPP and several patient administration vendors in one sweep. Trusts are choosing Epic partly because their neighbours already have a pattern that compounds rather than diversifies risk. The direction of travel, consistent with the wider consolidation of integrated care boards and the run-up to NHS England's abolition, is toward fewer, larger, more entangled digital estates built on a single vendor's architecture.

That entanglement was supposed to be the point. Fewer logins, shared data standards, less duplicated entry, easier flow between sites, the arguments for scale are genuine and NHS Digital Maturity data on reduced length of stay and lower per-spell costs bears some of them out. But scale of this kind also means that Epic's internal governance, ownership structure and leadership stability stop being a Wisconsin story and become an NHS resilience question. Faulkner has run Epic since 1979 and has resisted acquisition, flotation and outside investment throughout. Rana was the person most consistently named as her chosen successor and one of the original architects of MyChart, the portal now embedded in ten major English trusts and drawing roughly £3 million a year in NHS fees. His departure removes not just a leader but the clearest signal of continuity that NHS procurement teams had been quietly relying on when they signed decade long contracts.

None of this suggests Epic is about to falter. Its statement about a culture of growing internal leadership is plausible, and Faulkner shows no sign of stepping back herself. But the episode exposes something Sir Jim Mackey's accountability agenda has largely left untouched, which is that the NHS has built serious dependency on a handful of overseas technology firms without building equivalent mechanisms to interrogate their internal succession, ownership risk or long term strategic direction. Trusts scrutinise capital business cases in exhaustive detail. Almost none of that scrutiny extends to asking what happens to a supplier if its founder retires, sells, or simply changes course, even where that supplier now sits underneath clinical records for millions of patients.

For NHS leaders currently negotiating or renewing EPR contracts, the practical lesson is to treat vendor governance as a live commercial risk rather than a footnote, with the same seriousness applied to financial covenants or data security. For policymakers, it sharpens an argument that predates this episode, over whether the NHS App and the federated data platform can ever be more than a thin layer sitting atop foreign owned clinical infrastructure. For health tech firms watching Epic's UK expansion, a leadership change at the top of the market leader is an opening worth testing. And for patients, the honest answer is that a personal decision made in Wisconsin will not touch their care this year, or probably next. The unease lies further out, in a health system that has outsourced a great deal of its digital future to institutions it does not govern and increasingly cannot easily replace.