Structural risk in healthcare sits at the boundary
When we talk about clinical risk, we typically mean risk inside an organisation — the kind addressed by clinical safety management systems under DCB 0129 and DCB 0160, and scrutinised by CQC under Regulation 12 (Safe Care and Treatment).
But the most consequential clinical risk in modern NHS delivery does not sit inside any single organisation. It sits in the joins between them. This is structural risk — and it takes three forms.
| Risk Type | What Happens | Provider Impact |
|---|---|---|
| Accountability Gap | No single organisation owns end-to-end clinical safety across a pathway. Responsibility is implied, not verified. | CQC non-compliance, litigation exposure, indefensible governance under investigation |
| Information Decay | Clinical intent, reasoning, and context degrade as they cross organisational boundaries. Data transfers; meaning does not. | Clinical error, misinterpreted referrals, patient harm at handoff points |
| Operational Friction | Governance overhead at boundaries creates delays, duplication, and workarounds that slow care delivery. | Cost inefficiency, workforce burden, pathway backlogs |
Every NHS provider model operating across organisational boundaries is exposed to all three. The question is where and how severely.
Four models, four architectures
The UK healthcare market is consolidating around a handful of provider archetypes. Each has its own logic, its own strengths, and its own theory of how scale and quality coexist.
The digital front door. Patients enter through an app or employer scheme, receive remote triage and consultation, then get routed onward — to NHS services, specialists, or physical providers. The model is efficient and scalable. It is also, by definition, a handoff machine. Every patient journey creates at least one organisational boundary crossing, often more. These platforms are increasingly common in corporate healthcare and insurance-backed provision, and they are beginning to appear in NHS pathways.
The corporate contract holder. Large organisations holding NHS primary care and urgent care contracts across multiple sites and populations. The complexity here is structural — multiple practices, subcontracted services, ICS pathway integration, statutory obligations distributed across a corporate delivery network. Every contract boundary creates a governance seam. Every subcontracted clinical service creates a responsibility surface that falls squarely within CQC's safe care mandate.
The clinician-led federation. GP practices collaborating through shared services and ICS alignment, maintaining local autonomy while achieving collective scale. The governance challenge is different again — responsibility distributed across a network of semi-independent partners, each with their own clinical and operational accountability. These federations are often deeply trusted locally, but the shared responsibility model introduces its own ambiguity, particularly when clinical safety obligations under DCB 0160 apply to systems deployed across multiple partner organisations.
The scaled regional provider. Organisations delivering primary care, urgent care, and community services across regional networks — often hundreds of thousands of registered patients, hundreds of thousands of appointments annually, organised into geographic clusters. Their strategic direction typically aligns with integrated neighbourhood health, prevention, and digital innovation, placing them squarely at the intersection of multiple care boundaries.
Four genuinely different approaches. Different ownership structures, different clinical models, different commercial strategies, different scales.
And yet.
Where they all break
Every one of these models operates across organisational boundaries. That is not incidental to their strategy — it is their strategy. Integration, collaboration, pathway coordination, network delivery. These are boundary-crossing activities by nature.
The problem is that NHS clinical governance infrastructure was not designed for this. It was designed for a world where organisations were largely self-contained and accountability sat within institutional walls. The frameworks we rely on — CQC registration, DCB 0129/0160 clinical safety management, even the statutory duty of quality — are organisational in scope. They assume a responsible entity. They do not address what happens in the space between responsible entities.
The Data (Use and Access) Act 2025 makes boundaries more porous. GP Connect, the Federated Data Platform, and expanded data sharing obligations mean that clinical information flows more freely between organisations than ever before. But making boundaries more porous does not make them more governed. Risks are surfacing in audit trails across these newer data pathways, but the structural accountability for those risks — who owns the clinical safety obligation when data and responsibility cross an organisational boundary — remains largely invisible.
When care crosses an organisational boundary today, several things happen simultaneously.
Responsibility becomes implicit. There is rarely a structured, verifiable moment where one organisation accepts clinical accountability for a patient and another relinquishes it. Referral letters imply transfer. Acceptance is assumed from silence. Rejection may not surface for days or weeks. In the interim, the patient exists in a governance gap — nominally someone's responsibility, practically no one's.
Clinical intent gets reinterpreted. A referring clinician has a specific reason for the referral — a differential diagnosis, a concern, a required investigation. By the time that intent reaches the receiving organisation, it has been summarised, reformatted, or simply lost. The receiving clinician works from their own assessment, not from the structured intent of the referrer. If the outcome is poor, reconstructing what was intended and by whom becomes an exercise in retrospective interpretation rather than evidence.
Provenance chains fracture. The clinical context that informed a decision at one organisation does not travel intact to the next. Shared care records help, but they record data, not reasoning. The why behind a clinical decision — the differential that was considered, the risk that was weighed, the option that was ruled out — is rarely machine-readable or transferable. When something goes wrong across a boundary, the evidential chain has gaps that no amount of retrospective record review can fill.
Outcome feedback loops fail to close. The organisation that referred a patient rarely receives structured confirmation of what happened next. Did the patient attend? Was the referral appropriate? What was the outcome? This information exists somewhere in the system, but it does not flow back to the point of origin in any consistent or timely way. Without closed feedback loops, referring organisations cannot learn, cannot adjust, and cannot demonstrate the quality of their onward routing decisions.
And critically, no one can see the gap. There is no system-level visibility of where clinical responsibility sits at any given moment in a cross-organisational pathway. Not for the organisations involved, not for commissioners, not for regulators, and not for patients. The governance gap between organisations is not just unmanaged — it is unobservable.
Why the model doesn't matter
This is the structural insight that matters for anyone running a provider organisation or commissioning services across boundaries.
Digital front doors create responsibility discontinuity at the point of handoff to physical care. Corporate contract holders create accountability opacity across statutory obligations and subcontracted services. Clinician-led federations create shared responsibility ambiguity across partner organisations. Regional providers create governance complexity across geographic and service boundaries.
The surface presentation differs. The underlying failure is identical.
Clinical responsibility crosses organisational boundaries without structured acceptance, without time-bounded accountability, without machine-verifiable state, and without system-level observability.
This is not a technology gap. Interoperability standards, shared care records, and digital referral systems all exist. They solve the data problem reasonably well. They do not solve the responsibility problem at all.
A shared care record tells you what happened. It does not tell you who was accountable when it happened, whether accountability was explicitly accepted, or what the conditions of that acceptance were. An e-referral system routes a request. It does not verify that responsibility has transferred, or define the terms under which that transfer is valid.
What neighbourhood health makes harder
The NHS is moving toward neighbourhood health models, integrated care, prevention, and shared accountability. This is the right direction clinically and operationally.
But it makes the cross-organisational governance problem worse, not better.
Every additional partner in a neighbourhood team adds handoff surfaces. Every shared care pathway adds responsibility transfer points. Every integration initiative adds governance complexity. The more organisations collaborate — and they should — the more boundaries clinical responsibility must cross.
More collaboration means more boundaries. More boundaries mean more places where responsibility can become ambiguous, delayed, or invisible.
The current approach to managing this complexity relies on goodwill, local relationships, and informal professional agreements. These work well in stable conditions with known partners and established trust. They are brittle under pressure, opaque under scrutiny, and indefensible under investigation.
What a coroner would ask
The real test of cross-organisational clinical governance is not whether it works when everything goes well. It is whether it can withstand the questions asked when something goes wrong.
A coroner investigating a death that occurred between organisational boundaries will want to know: Who held clinical responsibility at the point of deterioration? When was responsibility accepted by the receiving organisation? What were the conditions of that acceptance? Was the referring organisation informed that their referral had been actioned? If not, what was the expected response time and what should have happened when it was exceeded?
These are not unreasonable questions. They are precisely the questions that current cross-organisational governance cannot answer with structural evidence. The answers exist, if they exist at all, in email threads, phone call recollections, and retrospective clinical record review.
That is not clinical governance. It is reconstruction.
The question for provider leadership
If you are running an organisation that delivers care across boundaries — and in 2026, that is effectively every provider — there is a question worth sitting with.
Can you show, at any given moment, where clinical responsibility sits in the pathways you operate across?
Not approximately. Not by inference from records after the fact. But structurally, in real time, with the kind of auditability that would satisfy a coroner's inquiry or a CQC investigation into cross-boundary care failure.
If the answer is no — and for most organisations it will be — that is not a criticism. It is a description of the current state of NHS clinical governance infrastructure. The frameworks, standards, and systems we have were built for a world of self-contained organisations. We no longer live in that world.
The question is whether the infrastructure catches up with the reality of how care is now delivered.
