Why Is NHS Health Data So Fragmented?

Your GP holds one record. Your hospital holds another. Your specialist holds a third. Nobody has the full picture — including you. Here's why, and why it keeps affecting the patients but is impossible to change soon.

TL;DR

NHS health data is fragmented because of a combination of decentralised architecture (no single national record), competing IT vendors with incompatible systems, a £12 billion national IT programme that collapsed in 2011, inconsistent information governance across 6,500+ GP practices and 215 hospital trusts, and a fundamental design choice: provider records were never meant to capture the full patient.

The Darzi Report (September 2024) and the Sudlow Review (November 2024) both confirmed it. The Data Use and Access Act received royal assent in June 2025. And patients — especially those who move, travel, or manage chronic conditions — are still paying the price.

Start with what sounds like it should not be true: there is no central record of patient data in the NHS. Your GP surgery holds your primary care record. Each hospital you attend holds its own record of your treatment there. Your mental health provider holds theirs. Your community physiotherapist holds theirs. And when you leave one setting and enter another, the data rarely follows you.

This isn't a secret. The House of Commons Library confirmed it directly in its 2025 research briefing: data sharing practices vary across the country, there is no consistency in what data is held where and why.

What's less obvious is why this happened — and why it's so stubbornly persistent. The answer isn't laziness or one bad policy. It's a layered accumulation of structural decisions, failed programmes, commercial competition, and a fundamental misunderstanding of who the patient actually is.

The six reasons

1. The NHS was built decentralised — and that was a political decision, not a design choice

The NHS wasn't designed around data. It was designed around a compromise.

When Aneurin Bevan created the NHS in 1948, he faced resistance from two directions simultaneously: hospital consultants who feared losing their independence, and local authorities unwilling to surrender the municipal hospitals they'd spent decades building. The solution was a federated structure. Hospitals were nationalised but administered through regional Hospital Boards, not a single central body. GP practices were kept entirely independent — GPs became contractors to the NHS, not employees of it, a concession Bevan made to get the British Medical Association's acceptance. He famously said he "stuffed their mouths with gold."

The result was three separate administrative streams from day one: hospitals, family practitioners, and local health authorities — each with their own governance, their own budgets, and their own records. That three-way split was never fully resolved. Every major reorganisation since — 1974, 1990, 2012, 2022 — preserved some version of the primary/secondary/community divide. The data fragmentation we have now is, in part, a direct inheritance of the political architecture of 1948.

In practice, this means over 6,500 GP practices across England, each one legally responsible for the data it holds, and over 215 hospital trusts, each procuring their own IT. The Health Foundation puts the scale of it plainly: more than 300 million appointments conducted in primary care each year, across systems that were never required to talk to each other.

2. A £12 billion IT programme failed — and left a mess behind

In 2005, the NHS Connecting for Health Agency launched the National Programme for IT (NPfIT). The goal was a single nationwide electronic health record for every patient in England. By 2010. It was, by most assessments, the most expensive healthcare IT failure in history. Individual trusts were left to develop their own data systems. The result is the fragmented digital landscape that persists today: a patchwork of locally-procured, incompatible systems, some still running on software no longer supported anywhere else.

3. Two vendors dominate — and they don't speak the same language

In primary care, two systems hold most of the market: EMIS Web (originally Egton Medical Information Systems, now owned by UnitedHealth's Optum) and SystmOne (built by TPP). They were paid £77 million for primary care software in 2018 alone. These two systems operated in near-complete incompatibility for years. The first pilot of data exchange between them happened in May 2013. A formal cross-sharing agreement wasn't signed until 2015. Meanwhile, secondary care uses an entirely different zoo of systems — Epic, Cerner, Lorenzo, and more — each deployed at trust level, each with their own data models, each requiring custom integration work to share even basic patient information across a GP-hospital boundary. On over 11 million occasions a year, patients present to a hospital using a different EHR or paper record system than their previous attendance. That's 9.1% of all acute presentations where the clinician is flying without a chart.

4. Data was collected for billing, not for care

This one doesn't get enough attention. A significant portion of NHS data collection was designed for operational and financial purposes — ensuring hospitals get paid, managing resource allocation, commissioning. Not for clinical insight. Outpatient administrative data frequently lacks diagnostic information because the primary purpose is operational. The Commissioning Data Set (now in version 6.4) was designed to reflect activity, costing and commissioning — not the true complexities of clinical practice. When data is collected to justify a payment rather than to understand a patient, the resulting record reflects what was billed, not what was experienced. That gap is structural. Over 100 separate clinical registries have been created since 1974 precisely because the mandated NHS data flows couldn't capture what clinical teams actually needed to know.

5. Information governance created silos, not safety

The legal framework around data sharing is simultaneously well-intentioned and actively counterproductive. The Data Protection Act created genuine obligations for GPs and hospitals to protect patient data — but the interpretation of those obligations varied wildly across organisations and created a powerful incentive to not share rather than to share thoughtfully. EMIS's own Chief Clinical Officer acknowledged it directly: the legal framework is the main barrier, because the law creates a duty of confidentiality and a duty to share simultaneously, with no clear mechanism for resolving the tension. The result: a 2016 Digital Maturity assessment found that of 249 NHS Trusts surveyed, only 56 shared care records digitally with non-primary care providers.

6. Provider records were never designed to capture the full patient story

This is the deepest reason, and the one most rarely stated. NHS records capture what happened in an NHS setting. They do not capture what happens at 2am when a patient is managing symptoms alone. They do not capture medication adherence, or the pattern of what a patient describes to one GP but not another, or the health history accumulated across four countries before someone arrived in the UK. The richest data about most patients' health — daily life context, symptom patterns, cross-border history — was never meant to live in an EHR. It was always meant to live with the person. The patient is the missing interoperability layer.

What it costs

Fragmentation isn't just an inconvenience. It produces measurable harm.

An estimated 1.8 million undetected transition medication errors occur in NHS England every year, resulting in harm across approximately 31,604 patient episodes. 52% of those harmful errors happen during hospital admission — the moment of highest transition between data systems.

33% of hospital trusts still cannot electronically access outside patient data, despite clear evidence that data-informed decisions improve treatment outcomes by 30%.

The Darzi Report was blunt: the NHS is "in the foothills of digital transformation." The richness of NHS data is "largely untapped" for clinical care, service planning, and research. Darzi identified chronic capital underinvestment — a £37 billion shortfall over the past decade and a half — as a key driver.

Professor Sir Martin Landray of Oxford put it plainly in his response to the report: "Data are in silos — hospital by hospital, GP surgery by GP surgery — with little if any data available centrally on mental health or community care."

The Sudlow Review, published two months later in November 2024, confirmed the same picture. The Data (Use and Access) Act received royal assent in June 2025, aiming to address this through information standards that now apply to IT providers across health and social care. It's a step. It is not yet a solution.

Who it hurts most

Fragmentation hurts everyone. But it hurts some people more than others.

People with complex or chronic conditions — diabetes, Parkinson's, HIV — often receive care across multiple settings simultaneously. Research on NHS App rollout found that these patients were actively less supported by digital access, because the app primarily served as a gateway to GP systems. Their hospital records, their specialist notes, their community health data: largely inaccessible through any patient-facing tool.

People who move. If you move to a new GP, your records can be transferred via the GP2GP system — but data is notoriously lost in translation between different EHR systems during the process. If you've attended hospitals in different regions, those records stay where they were created. If you came from another country, your pre-UK medical history doesn't exist in any NHS system at all. The 9.8 million migrants living in England are not systematically recorded in primary care electronic health records.

People who are older and have a long history. The longer someone has been a patient, the more providers have touched their care. Each transition is a potential information gap. Each gap is a potential harm.

What's coming

The government has committed to a Single Patient Record via the NHS App, aiming for full access by 2028. As of early 2025, 91% of secondary care trusts have an electronic patient record system. Progress. But of 215 NHS trusts, only 84 have implemented patient engagement portals — and only 26 have integrated those portals with the NHS App.

The structural challenge remains: even a perfect Single Patient Record would still only contain what NHS systems recorded. It would still not capture what happened before someone became an NHS patient. It would still not include health history from other countries, or the daily lived experience of managing a condition that never prompted a GP appointment.

Provider records are incomplete by design. Not as a failure of ambition, but because they were built to document clinical encounters — not to represent a person's entire health life.

The patient is the missing interoperability layer.

Frequently asked questions

• Can I access my full NHS health record? You can access your GP record through the NHS App, and under the Data Protection Act 2018 you can request data held about you by any NHS provider, and Health Data Avatar helps you optimise this. However, hospital records, specialist notes, and mental health records are largely not accessible through a single interface. Records from different providers remain in separate systems and portals.

• Why do NHS hospitals use different IT systems?

After the collapse of the NHS National Programme for IT in 2011 — a £12 billion programme that failed to deliver a unified national EHR — individual NHS trusts were left to procure their own systems. This produced a patchwork of different vendors (Epic, Cerner, Lorenzo, and others) at trust level, with incompatible data formats and no mandated interoperability standards until very recently.

• What is FHIR and why does it matter for NHS data?

FHIR (Fast Healthcare Interoperability Resources) is an international standard for exchanging healthcare data. It provides a common format so different EHR systems can share patient data without custom integrations. NHS England has adopted FHIR as a key part of its interoperability strategy. Health Data Avatar supports FHIR export, meaning patients can take their organised health records and share them in a format that clinical systems can read.

• What is the NHS Single Patient Record?

A government commitment to create a centralised, accessible patient record available via the NHS App by 2028, so that records no longer sit only with individual GP practices or hospitals. Implementation is underway but complex, dependent on legacy IT upgrades across hundreds of NHS trusts.

• What can patients do now about their fragmented health records?

You have the legal right to request your health records from any NHS provider. The practical problem is doing anything useful with what arrives: PDFs, scanned letters, discharge summaries, coded GP exports in formats nobody explained to you. Health Data Avatar (HDA) was built for exactly this — upload everything from any provider, any country, any format, and get it structured, translated where needed, and searchable. So you have the complete picture even when the systems don't.