Aligning an academic health system's IT and clinical leadership

Academic health systems are in an optimisation year, not an implementation year. Most already run their electronic health record platform. Third party research bears that out: KLAS Research found the number of hospitals affected by new electronic health record decisions fell roughly 40 percent in calendar 2025 against 2024, with capital moving instead toward artificial intelligence and operational efficiency.

So the competition inside IT has changed shape. It is no longer about which system to buy. It is about who gets the finite analyst and informatics capacity, with optimisation, integration, reporting and revenue cycle analytics in one queue.

Margin pressure is the force behind that shift. Kaufman Hall’s National Hospital Flash Report, drawn from more than 1,300 hospitals, put the calendar 2025 median hospital operating margin at 1.3 percent including allocations. The American Hospital Association records 2024 hospital expense growth of 5.1 percent against 2.9 percent general inflation, with labour at 56 percent of total expenses. Neither is academic-medical-centre specific; no academic-only margin benchmark is published.

Artificial intelligence arrives into the same queue, ahead of its own governance. A peer-reviewed study in the American Journal of Managed Care found 62.6 percent of Epic-using US hospitals had already adopted ambient documentation AI. The Joint Commission and the Coalition for Health AI published their responsible use guidance in September 2025, after most of that adoption had happened. Systems are now standing up oversight across informatics, quality, compliance and clinical operations for tools that are already live.

All of it lands on the same constraint: too much demand, one queue, and no agreed way to decide what goes first. That is a prioritisation problem before it is a technology problem, and prioritisation problems are decided by people who do not usually sit in the same meetings.

The figures above are third party research about the sector, from KLAS Research, Kaufman Hall, the American Hospital Association, the American Journal of Managed Care and the Joint Commission. None is an outcome of this engagement.

The programme manager running the electronic health record portfolio described the problem better than we could have.

Her teams were siloed. There were no limits on work in progress, so every new project entering the pipeline pulled people who were already committed elsewhere. The projects with the loudest stakeholders moved to the front and everything else slipped quietly. She had drawn the current state out herself, as an illustration, to show her leadership what was actually happening.

That is the ordinary shape of an unmanaged electronic health record portfolio. A service line wants a build, a regulatory change forces one, a payer requirement forces another. Each has a sponsor who considers it urgent, each is agreed in a separate conversation, and none is ranked against the others. Nothing is refused, so everything is late, and the delay lands hardest on the work with the quietest sponsors.

What made this solvable was that she had already done the diagnosis, and she wanted her IT and clinical leaders hearing it together, from someone who was not her.

The last part matters more than it sounds. An internal champion who has made the same argument for a year has usually exhausted their credibility on it. Not because the argument is wrong. Because it has become familiar, and familiar arguments get heard as position rather than analysis. The value of an outside voice is often that the argument arrives without its history.

Two days of preparation before anyone entered the room. The week before, we reviewed the client’s own value stream artefacts with her, and built the workshop on what her organisation had already documented about itself rather than a standard curriculum.

The order matters. Design the day first and gather context afterwards and you get a generic deck with the client’s logo on it. A senior clinical audience detects that inside twenty minutes, and once they decide the material is generic they stop bringing their own problems into the room. From then on you are presenting.

One day, both groups, together, on Zoom. The system’s directors and vice presidents, IT and clinical side by side, 8am to 4pm.

Running the groups separately is more convenient and wrong. Whoever goes second hears a summary of what the first group concluded, and a summary of someone else’s conclusion is something you negotiate with. Two sessions produce two accounts and a reconciliation meeting nobody schedules. One produces a single account, with the disagreements surfaced in front of each other.

The agenda came from her diagnosis: work in progress, the real cost of unlimited intake, and how a planning cadence changes who gets to decide what moves first. Certifying anyone was never the goal. Certification brings a fixed syllabus and an exam window, and would have spent the opening hours on framework vocabulary these people did not need.

Structured for the format rather than adapted to it. A full day virtually with senior clinical leaders is the hard part of this delivery. Executives leave for clinical obligations, cameras go off, attention fragments. Lecture format punishes that. Step out for forty minutes and you return to a slide with no idea what happened, so you stay quiet for the rest of the day.

We built it as short working blocks against a shared board, so people were producing something rather than receiving slides, and so anyone pulled away could rejoin and carry on. The board is the memory of the room, and the artefact they keep.

That board needed the client’s security team to approve a tool exception, which their programme manager secured two days before we ran. The facilitation technique is useless if it cannot clear information security. Leave it to the morning of and the fallback is slides, the exact failure the design was built to avoid.

We could have pushed for two days in person. Most firms would have, and it would have been an easier day to deliver and a more impressive looking engagement on paper.

A room of that seniority in a health system does not have two consecutive days. Their calendars are held by clinical obligation, board cycles and the same intake pressure the workshop was about. Asking for two days would have produced one of two outcomes: the session happens months later with half the attendance, or the scheduling conversation runs out of energy and it never happens.

A full day on Zoom with a tight agenda was the version that could take place while the intake problem was still urgent to her leadership. That window closes when the problem stops being the loudest one, and it closes on its own schedule.

Scoping to what the client can genuinely absorb, rather than to what we would prefer to sell, is usually the difference between a workshop and a diary invitation that gets moved twice and then cancelled.

Let the internal champion own the diagnosis and bring the outside voice to deliver it. Reversing that is the common pattern and the weaker one: an independent diagnosis takes weeks, arrives with less detail than the person who lives there already has, and hands the champion someone else’s version of their own argument.

Put both sides in the room at the same time. If IT and clinical leadership cannot be got into one session, that scheduling difficulty is not a logistics problem to work around. It is the first measurement of the alignment problem.

Build from the organisation’s own artefacts, even rough ones. A hand-drawn current state the room recognises does more work than a polished reference model belonging to somebody else.

Expect the conversation to move up. Work in progress is where these engagements start because it is visible and everyone feels it. Funding and prioritisation is where the problem actually lives, and it involves a different set of people.

The clearest signal came without being asked for. Over the five months after the workshop, ten of their people enrolled themselves on further training. Nobody was directed to do that and none of it was sold. In a health system where every hour of leadership time is contested, people spending their own budget and calendar on more of the same method is the most reliable evidence available that the first day worked.

The conversation itself has moved up a layer. It began as one workshop about work in progress. It is now about how the portfolio gets funded and prioritised, which is where the intake problem actually lives, and which is a conversation with a different set of people in the room. That shift happened without a follow-on engagement holding it open.