I spent the better part of two years navigating complex medical care across both private and NHS pathways. Multiple procedures. Multiple consultants. Multiple organisations. And at every boundary between them, the same thing happened. Information got lost, responsibility went undefined, and I ended up being the only person holding the thread of my own care together.
So I started writing my own clinical notes. Maintaining them on an iPad. Carrying them between appointments.
Not because I'm a difficult patient. Because neither system could be trusted to transfer them.
The View From the Consultant’s Desk: Fragmented NHS IT Systems
I wasn't the only one who could see the problem. One of my consultants told me that blood test results routinely took two weeks to travel from the service centre to the private hospital. Two weeks. Not because anyone was doing anything wrong, but because there was no infrastructure for the result to flow from where it was generated to where it was needed.
The same consultant told me he had to open eight separate applications to assemble a coherent picture of a single patient. Eight apps. Not because he'd chosen them. They'd been bought because they were cheap. Each one solved a narrow function at the lowest cost, and nobody had asked whether they'd work together. Nobody had asked what happens to the patient when eight isolated systems are expected to behave as one.
That consultant wasn't describing a technology problem. He was describing an architecture that had never been designed. A set of procurement decisions made in isolation that produced fragmentation by default.
Another surgeon described a different kind of failure. Referrals arriving on his desk that should never have left primary care. Pre-referral tasks that hadn't been completed. Protocols that existed on paper but had no mechanism for enforcement. The referral moved from GP to surgeon before the GP's work was done, wasting the surgeon's time, delaying the patient, and clogging secondary care with work that shouldn't have been there yet. The protocol existed. The infrastructure to enforce it didn't.
Then there was what I noticed at the GP end. The cognitive load was visible. At almost every appointment, the first three to four minutes of a ten-minute consultation were spent piecing together what had happened previously. The GP scrolling, clicking, reconstructing a timeline from scattered data points. Who did what, when, where, what was the result, what's still outstanding. All of it buried across screens and entries that were never designed to tell a coherent story about a patient's journey.
That's not a minor inefficiency. That's a third of the clinical encounter lost to data archaeology. And the cost isn't just time. It's cognitive. The GP is burning mental energy on assembly when they should be spending it on clinical reasoning. By the time they've pieced the picture together, they're already depleted for the actual decision-making.
And then there's the detail that makes people outside healthcare stop and stare. In 2025, letters between healthcare professionals are still sent by post. Clinical correspondence containing patient data and clinical decisions, put in an envelope and trusted to Royal Mail. Days in transit. No tracking. No confirmation of receipt. No audit trail. No way to know if it arrived, if it was read, if it was acted on. Not as an exception. As standard practice across huge parts of the system.
Meanwhile, phlebotomy departments are being run with a mobile phone and an A4 notebook. A clinical department handling blood samples that feed directly into diagnostic decisions, managed with handwritten notes and personal devices. Not because the people running them are incompetent. Because nobody gave them anything better. The procurement decisions that produced the consultant's eight apps didn't even reach them. They were left to improvise their own infrastructure with whatever was to hand.
Solving Healthcare Interoperability: Lessons from Enterprise ERP
What struck me as an engineer was how familiar this looked. Not from healthcare, but from enterprise.
In the early 1990s, large businesses had exactly the same problem. Finance ran on one system. Procurement on another. Logistics on a third. HR on a fourth. Each department had bought whatever was cheapest for their specific function. Data didn't flow between them. Nobody had a coherent picture of the business.
SAP and the ERP wave solved this. Not by replacing every system, but by creating a shared data infrastructure that the business could operate on as a whole. Common data models. Shared identity. Transactions that flowed across departmental boundaries without manual intervention.
Healthcare in 2025 is where manufacturing and finance were in 1990. But there is a crucial difference. SAP solved the problem within the enterprise. One company, multiple departments, shared infrastructure under a single governance structure. Healthcare's problem is inter-organisational. Different legal entities, different governance, different systems, different accountability structures. Nobody has authority over the space between them. And that space is exactly where patients get lost.
This problem is about to get significantly worse. The shift to neighbourhood health depends on more organisations collaborating more closely at local level. More boundaries. More handoffs. More transfers of responsibility between providers who have no shared infrastructure to support those transfers. The policy is pushing towards integration while the architecture remains fragmented. It is accelerating the problem, not solving it.
Clinical negligence lawyers already know this. Cases are emerging where the transfer of responsibility between organisations is the exact reason the case exists. Not a clinical error made within a team. The handoff itself. The moment where one organisation believed responsibility had passed and another didn't know it had arrived. That gap between organisations isn't a quality improvement issue. It's a structural liability that will only grow as neighbourhood health multiplies the number of organisational boundaries that patients cross.
The Clinical Cost of Missing Longitudinal Health Records
But the real cost of this failure isn't measured in consultant frustration or wasted time, as significant as those are.
Towards the end of my care journey, something emerged that changed my understanding of the whole experience. A specific post-operative medical issue turned out not to be a new problem at all. It was an exacerbation of a condition that had been afflicting me for six years.
Six years. Multiple presentations across different providers, in different contexts, at different times. Each clinician saw the episode in front of them and treated it in isolation. Each one was doing their job competently within the information they had available. But none of them could see the pattern, because the pattern was spread across years of fragmented records in disconnected systems.
That pattern, had it been visible, would have changed everything. Earlier diagnosis. Different treatment. Fewer procedures. Months of time off work that could have been avoided. Years of reduced quality of life that didn't need to happen.
This is what the absence of longitudinal health records actually costs. Not in the abstract language of data governance strategies and interoperability roadmaps. In months of someone's life. In suffering that was avoidable. In a condition that was active and visible in the data for years, if only someone had been able to see it all in one place.
From Observation to Engineering
It was clear to me that the challenges didn't lie within any single organisation. The problems lay in the movement of information between NHS organisations themselves, and between the NHS and the private sector actors in my care pathways. The gaps existed in the spaces between providers, in the transitions that nobody owned.
After much experimentation, this led me to look beyond healthcare altogether. I started studying safety-critical systems. Aviation. Nuclear. Defence. Industries where responsibility transfer isn't optional and failure isn't tolerated. I wanted to understand what safe responsibility transfer would actually look like in a clinical environment if you engineered it from first principles rather than patching it with workarounds.
What emerged was that clinical handoffs don't fail in one way. They fail in seven.
The Seven Flows of Care Framework
The answer isn't one thing. It's seven.
Identity needs to follow the patient across every boundary. Not just demographic data, but verified, trusted identity that every provider can rely on without re-entering it from scratch.
Consent needs to travel with the data. Not as a one-time checkbox, but as a dynamic, auditable record of what the patient has agreed to, with whom, and under what conditions.
Provenance needs to be traceable. When a blood test result arrives, who ordered it, where was it processed, what chain of custody did it follow? Without provenance, data is just noise.
Clinical intent needs to be explicit. When a consultant orders an investigation, the intent behind that decision needs to follow the request through every system it touches. Otherwise the receiving end has a task without context.
Alert and responsibility needs clear ownership. At every point in a patient's journey, someone needs to be explicitly holding clinical responsibility. When that responsibility transfers between organisations, the transfer itself needs to be a first-class transaction, not an assumption.
Service routing needs to be intelligent. Referrals, requests, and clinical communications need to find their way to the right service, in the right organisation, with the right context attached.
Outcomes need to flow back. The result of every intervention, referral, and handoff needs to be visible to everyone involved in the patient's care. Not locked inside the system that happened to capture it.
These seven flows aren't a theoretical framework. They're what I watched fail, repeatedly, as a patient. And they're what every consultant, nurse, care coordinator, and social prescriber watches fail every day.
Building the Missing Infrastructure Layer for Healthcare
At Inference Clinical, we're building the infrastructure layer that healthcare skipped. Not another app to add to the pile. Not a ninth application for that consultant to open. The connective tissue between organisations that makes clinical data flow safely, with consent, with provenance, with clear responsibility at every step.
This isn't a new idea. It's an overdue idea. Every other major industry made this architectural decision a generation ago. Healthcare didn't, and patients have been paying the price ever since.
I know because I was one of them. Carrying my notes in a folder, being my own integration layer, watching a six-year pattern go unrecognised because the systems that held the data were never designed to talk to each other.
Nobody should have to be their own health record. The infrastructure should exist. Now we're building it.
