When an electronic patient record programme fails, the post‑mortem almost always lands in the same place: adoption. Clinicians didn’t engage. Training didn’t stick. The change team was under‑resourced. All of which may be true — and none of which is where the failure started.
By the time a programme is fighting for adoption, the decisive mistakes are usually two or three years old. They were made below the waterline, in the foundations, long before a single clinician saw a login screen. Adoption is where the damage surfaces. It is rarely where it began.
I’ve spent a long time on the delivery side of these programmes, and the pattern is consistent enough to be diagnostic. So this is a piece about looking in the right place — and doing it early enough to matter.
What “below the waterline” actually means
Every EPR programme sits on a stack of foundations: the network and infrastructure it runs on, the state of the data it will inherit, the interoperability it can rely on, the governance that will steer it, the clinical‑safety discipline that will protect it, and the benefits logic that justifies it. None of these is glamorous. All of them are load‑bearing.
The trouble is that foundations are invisible when they’re sound and catastrophic when they’re not — and they’re decided early, when the programme is still a business case and an optimistic Gantt chart. That’s exactly when the pressure to look decisive is highest and the appetite for uncomfortable truths is lowest.
Adoption problems are foundation problems wearing a clinical face.
Consider a few of the ways this plays out.
The infrastructure everyone assumed was fine
A trust signs for a modern, cloud‑adjacent EPR on the quiet assumption that the estate can carry it. Nobody audited the wireless coverage on the older wards, the age of the devices at the bedside, or the single points of failure in the network. Two years later, clinicians are blaming “the system” for freezes and dropouts that are, in fact, a 2014 access layer buckling under 2026 load. The EPR gets the reputation. The infrastructure keeps the secret.
The data nobody looked at until migration
Legacy data is where optimism goes to die. Duplicate patient records, inconsistent coding, free‑text where structure should be, allergy information scattered across three systems and two decades. If the first honest look at data quality happens during migration planning, it’s already late — and the choice becomes migrate the mess or delay the programme. Neither is a good look in front of a board that has already announced the date.
The governance that couldn’t say no
Weak governance doesn’t announce itself. It shows up as scope creep that no one owns, as a risk register that’s a list rather than a live instrument, and as decisions taken in corridors because the forums to take them properly were never stood up. By go‑live, the programme is carrying a backlog of unmade decisions, and every one of them lands on the clinical floor as friction.
Why the diagnosis lands late
If foundation problems are so predictable, why do they keep surfacing at adoption? Three reasons, and they’re human before they’re technical.
First, foundations are boring to fund. A business case that says “we need eighteen months and real money to fix data, network and governance before we buy clinical software” is a harder sell than one that promises transformation on a poster. So the foundational spend gets trimmed, and the programme starts with a limp it will never quite lose.
Second, the people who could see the problem early are often the people with the least incentive to raise it. Flagging a foundation risk in month three makes you the person who slowed things down. Flagging it in month thirty makes you a realist. The incentives are backwards.
Third, and most awkwardly, the vendor selection often happens before the foundational assessment. The platform is chosen, the date is set, and only then does anyone ask whether the ground will take the weight. By then the assessment can only confirm bad news, not prevent it.
Diagnosing it while you can still fix it
The good news is that a foundation diagnosis is quick and cheap relative to what it saves. It isn’t a two‑year study; it’s a structured look at the load‑bearing layers, done honestly, with the findings written down in language a board can act on. A few questions do most of the work.
Can the estate carry the load you’re about to put on it? Not in theory — measured. Coverage, device age, network resilience, and the failure modes that will matter at 3am on a busy ward.
Do you know the true state of your data? Duplicate rates, coding consistency, the systems your record is actually spread across, and an honest view of what “clean enough to migrate” will cost.
Is your interoperability real or aspirational? The national and regional connections you’re relying on — do they exist and work, or are they on a roadmap you don’t control?
Is your governance an instrument or a diagram? Are there forums that can actually take decisions, a risk register that changes week to week, and a clinical‑safety function with the authority to hold the line?
Does the benefits logic survive contact with reality? If the business case promises benefits that depend on foundations you haven’t built, the case is fiction — and everyone will discover that at the worst possible moment.
The alignment argument
This is why we map programmes as a single connected system rather than a stack of separate workstreams. The layer where value is felt — the clinician’s screen, the patient’s access — is entirely dependent on layers they never see. Optimise the experience while the foundations sag, and you’re polishing the deck of a ship that’s taking on water.
None of this makes adoption unimportant. Workforce enablement is real work and it deserves real investment. But if your programme is struggling for adoption, resist the reflex to pour more change resource onto the surface. Go and look underneath first. The odds are good that what looks like a people problem is a foundation problem that finally reached the light.
Find it early, write it down honestly, and fund the fix before you fall in love with a platform. That single discipline separates the programmes that deliver from the ones that spend the next five years explaining why they didn’t.
Written by Jonah Aburrow‑Jones, Founder of Undine Digital. If this raised a question about your own programme, start a conversation — or read more of our insights.