Every executive in this survey got four picks. There were seventeen places to spend them. Watch where they went — and where they didn’t.
Each dot below is one of the 112 senior executives Redesign Health surveyed — CEOs, CIOs, CMIOs, all at health systems running Epic, two-thirds of them the final signature on clinical and administrative systems. They were asked to name the four functional areas where a brand-new startup has the best chance to compete.
Press play. The areas arrive in order of how many picks they drew.
Interoperability finished last. Four people out of a hundred and twelve.
That is the number the whole report will be quoted for, and it deserves to be. It is the most beloved problem in this field — the FHIR work, the data-liberation work, the reason a lot of us are here — and it drew four picks out of four hundred and forty-eight.
112 people × 4 slots = 448 picks, spread across 17 doors. If everyone had picked at random, each area would collect about 26 picks — 23.5%. That is the line to measure against, not zero.
Before the forced choice, the survey asked a rating question about the same list: on a scale of 1 to 5, how likely is it that an external vendor of any kind can compete against new and existing Epic solutions? No picks, no scarcity, no ranking — just a number for each area.
Here are both answers on their true scales. Top: the ratings, on the full 1–5 axis. Bottom: the picks, on a 0–100% axis.
Fifteen areas. The ratings occupy a 0.36-point band from 3.16 to 3.52, every one of them above the neutral midpoint, at a sample size where the margin of error is roughly ±9 points. Interoperability sits at 3.21 — a third of a scale point below imaging AI, and above bed and OR throughput, which drew six times as many picks.
The cliff in the bottom chart does not exist in the top one. It was manufactured by asking people to rank scarce attention instead of rate a market.
Not: 96% of buyers have decided a startup cannot win interoperability.
But: when 112 Epic-based executives had four slots of attention to spend across seventeen categories in early 2026, interoperability came seventeenth — while they rated it a 3.21, essentially tied with everything else.
Those are different claims, and only the second one is in the data. The second one may be worse news, not better: an area everyone considers moderately winnable and nobody thinks about is a harder place to raise money than one people argue about.
The report does not publish the forced-choice question wording — it publishes the rating question’s wording verbatim but not this one. It also never names two of the seventeen areas, and withholds pick shares for two more that it does name. The published shares sum to 331% where a top-four question should produce about 400%, so roughly 69 points of response are unaccounted for.
The publisher builds companies in these categories. Redesign Health is a venture studio, and the report says the survey was fielded “to inform both our investment thesis and the broader health tech market’s understanding.” The results are presented as High / Moderate / Low conviction cards.
The fielding window is four and a half months — October 29, 2025 to March 12, 2026 — and results are not broken out by date. Respondents early and late in that window were answering under materially different information about what Epic had shipped.
The sample is 52% for-profit community hospital systems and 8% academic-affiliated, which does not resemble the Epic install base. No weighting, recruitment method, response rate, incentive, or limitations section is published.
±9 points. At n = 112 the margin of error is roughly nine percentage points at 95% confidence. Every adjacent pair in the middle of the ranking is a statistical tie. The 4% and the 48% are the only two that clearly aren’t.
The companion piece to this one takes the same question to the federal certification record — 4,593 hospitals, and what actually gets installed alongside Epic. Four Percent →