Delayed Discharge: An Interface Problem, Not a Capacity Problem

This is the third article in a series examining the boundary between local authority social care and the NHS. The first article established that this is a crossing between constitutional domains. The second examined what happens when both sides of that crossing are dissolved and recreated simultaneously. This article examines the most expensive symptom of the ungoverned boundary: delayed discharge.

In September 2025, NHS England published — for the first time — the financial cost of delayed hospital discharge. The figure was £220 million in a single month. An average of 12,459 beds were occupied by patients who were medically fit to leave. The Institute for Government calculated that if this is a typical month — and the data across the preceding year suggests it is — the annual cost exceeds £2.6 billion. That is not far short of what the government spent on the entire courts and tribunals system in the same year.

By November 2025, 291,000 bed days were lost to delayed discharge. 49,688 patients experienced some form of delay. 2,654 waited 21 days or more. The Health Foundation reported that approximately 11% of all hospital bed days in England were occupied by patients who no longer needed to be there — worse than the 10% recorded a year earlier. Throughout 2024/25, an average of 12,663 patients per day remained in hospital despite meeting the criteria for discharge.

The conventional narrative is that this is a capacity problem. Not enough social care places. Not enough care home beds. Not enough domiciliary carers. And capacity is genuinely constrained — the Skills for Care report for 2024/25 showed that while vacancy rates had fallen to pre-pandemic levels, the improvement was driven almost entirely by international recruitment now being curtailed by visa changes from July 2025.

The human cost is not captured in the financial figures. Ten days in a hospital bed produces muscle deterioration equivalent to ten years of ageing in people over 80. Patients who remain in hospital after they are clinically ready to leave lose independence, acquire infections, develop confusion, and decline — often to the point where they need a higher level of care on discharge than they would have needed had they left when first assessed as ready. The Nuffield Trust found that delays for patients who had been in hospital for 21 days or more were driven primarily by the need for more intensive post-discharge services — services they may not have needed had the discharge crossing worked sooner.

But the NHS’s own data tells a different story about where the delays actually occur. And the language it uses to describe them reveals more than the numbers alone.

The Five Delay Categories and the Boundary They Describe

Since June 2024, acute trusts have submitted daily data to NHS England on reasons for discharge delay, categorised into five groups. The January 2025 data is representative of the pattern across the year:

Service capacity outside the hospital: 32.8% (3,218 patients). Interface processes: 28.2% (2,609 patients). Hospital processes: 20.4% (1,884 patients). Care transfer hub processes: 13.4% (1,243 patients). Wellbeing concerns: 5.1% (474 patients).

The Lowdown’s analysis of twelve months of data from October 2024 to September 2025 found that these proportions were essentially stable throughout the year, despite delayed discharge being a stated priority for every Integrated Care Board. On average, only 40–45% of patients assessed as ready for discharge actually left hospital when ready each month.

The category that matters most for understanding the boundary problem is “interface processes” — 28.2% of all delays, costing the NHS over £62 million in September 2025 alone. NHS England defines this as delays caused by “NHS trusts and system partners discussing patients’ onward care.” In practice, this means the hospital and the local authority — or their commissioned services — negotiating across the boundary about who is responsible for what, under which pathway, funded by which mechanism.

“Interface processes” is the NHS’s own term for what happens at the constitutional boundary described in the first article in this series. It is the crossing itself causing the delay. Not the capacity on either side. Not the clinical processes within the hospital. The act of transferring responsibility from one constitutional domain to another.

But the true share of boundary-caused delay is larger than 28.2%. The capacity figure — 32.8% — includes delays waiting for services that span the boundary. The King’s Fund’s analysis found that of the 3,203 daily capacity delays in March 2025, the single largest category (966 patients) was waiting for “bed-based rehabilitation, reablement or recovery services” — services commissioned by NHS trusts, local authorities, or both jointly. Even care transfer hubs, at 13.4%, are themselves boundary mechanisms: they exist precisely to coordinate the crossing between hospital and post-acute care. When the NHS adds up hospital-based processes (20.4%) and care transfer hub processes (13.4%), those two NHS-side categories account for 33.8% of delays — almost as much as the capacity figure usually attributed to social care.

Every one of the five categories describes a different facet of the same structural problem: the absence of governed infrastructure at the crossing between two constitutional domains.

From DTOC to D2A: What We Chose to Stop Asking

There is a critical gap in the data that the King’s Fund identified in May 2025: in 2020, NHS England stopped separating delay reasons between health and social care. The previous metric — Delayed Transfers of Care (DTOC) — attributed each delay to the NHS, to social care, or to both jointly. That metric was discontinued. The current categories do not make the attribution.

This was deliberate. The old regime — Schedule 3 of the Care Act 2014 — set time limits for social care assessments and imposed financial penalties on local authorities for every day a discharge was delayed where the delay was attributable to the local authority. The Health and Care Act 2022 repealed Schedule 3 entirely. The reimbursement regime was abolished. The requirement to complete social care assessments before discharge was removed. The formal notifications between NHS bodies and local authorities that previously structured the handoff were revoked.

The rationale was reasonable: the old regime had created perverse incentives. Hospitals would issue assessment notifications to start the penalty clock. Local authorities would rush assessments to avoid fines. Patients were being assessed in hospital environments that distorted the picture of their needs. The government’s justification, as Community Care reported, was that “the requirement to carry out assessments before discharge has resulted in some individuals experiencing delayed hospital discharge as they wait for their assessment to be carried out.”

In its place came the Discharge to Assess model — D2A — with four pathways: Pathway 0 (home, no new needs), Pathway 1 (home with support), Pathway 2 (community bed-based recovery), and Pathway 3 (new residential or nursing placement). The “home first” principle. Assess in the community, not the ward. The Hospital Discharge and Community Support Guidance replaced the legislative framework with guidance — setting out how NHS bodies and local authorities “should” plan and deliver discharge services, “affordable within existing budgets.”

D2A solved a real problem. Assessing people in their own homes produces better outcomes. The NHS Confederation reported that D2A contributed to a 28% reduction in patients staying over 21 days during winter 2020/21. It freed up 30,000 acute beds during the pandemic. The Age UK analysis found that the average excess bed day costs £346, making D2A cost-effective by reducing unnecessary hospital stays.

But D2A solved the assessment-location problem. It did not solve the boundary-governance problem. It moved where the assessment happens — from the hospital ward to the patient’s home or a community setting. It did not change the fact that the crossing between hospital care and social care requires a transfer of responsibility between two constitutional domains with different legal bases, different funding models, different accountability structures, and different information systems. It did not create infrastructure for that transfer. It removed the old infrastructure — the notifications, the timelines, the penalty regime — and replaced it with guidance that local areas “should” develop discharge processes “affordable within existing budgets.”

The result is visible in the data. Despite D2A becoming the default model from April 2022, despite care transfer hubs being established at every acute site, despite the discharge agenda being a stated priority for every ICB — the number of delayed patients has remained stubbornly high. The Health Foundation noted that before the pandemic, the daily count of patients ready for discharge but still in hospital averaged 4,000 to 5,000. By 2024/25, it averaged over 12,600. D2A changed the process. It did not change the structure.

As the King’s Fund observed: we stopped asking which side causes the delays — but not because we found the answer. We stopped asking because the question itself revealed an uncomfortable truth. The boundary is the cause. Attributing delay to one side or the other is a category error when the problem is the crossing between them.

What “Interface Processes” Actually Looks Like: The Data Disconnect

Behind the 28.2% statistic is a process that repeats thousands of times every day across England. A patient is assessed as having no criteria to reside. The clinical team determines the appropriate discharge pathway. The care transfer hub is notified. And then the crossing begins.

For a Pathway 1 discharge — home with new or additional support — the hospital needs to communicate the patient’s functional needs to the local authority or its commissioned services. This requires: confirmation of the patient’s identity across systems that do not share identifiers natively (the NHS number is not the local authority client ID). An understanding of the patient’s consent position — specifically, whether the patient has consented to their clinical information being shared with social care, under which legal basis, and for what purpose (the lawful bases differ between NHS and Care Act processing). The clinical reasoning that led to the discharge decision — not just the current status, but why the clinician believes this patient is safe to leave, under what conditions, with what risks. An explicit statement of who holds responsibility during the transition — between the point when the hospital’s duty of care ends and the local authority’s duty to assess and meet eligible needs begins.

For a Pathway 2 or 3 discharge — to a community bed or care home — the crossing is more complex. The question of who pays must be determined or deferred. If the patient may have a primary health need, then Continuing Healthcare eligibility must be considered — but under D2A, the full CHC assessment happens post-discharge, so the patient is placed in a setting funded by one mechanism (short-term, jointly, or one-sided) while the determination of long-term funding responsibility is deferred. The Hospital Discharge and Community Support Guidance acknowledges the complexity of “cross-border” situations where a patient may fall under one ICB and a neighbouring local authority — and advises that they “should agree local arrangements to ensure that any decisions about the joint funding of care can be made swiftly.”

“Should agree local arrangements” is the recurring motif of boundary governance at the LA–NHS crossing. It acknowledges the problem. It does not provide the infrastructure to solve it.

The 28.2% — the “interface processes” — is the time consumed by organisations negotiating these questions case by case, patient by patient, crossing by crossing, without infrastructure to resolve them systematically.

Seven Flows at the Discharge Crossing: Where Each Governance Flow Fails

The Seven Flows framework asks, of any organisational boundary crossing: are the necessary conditions for safe handover met? Applied to the hospital-to-social-care discharge crossing, each flow exposes a specific structural gap.

Identity. The patient has an NHS number in the hospital’s PAS system and an entirely different identifier — or no identifier at all — in the local authority’s social care system. The NHS number is not universally used in social care. Reconciliation depends on manual matching or local integration. When the patient is unknown to social care — a new referral — there is no pre-existing record to match to.

Consent. Hospital care is delivered under NHS Act authority with implied consent for treatment. Social care assessment and provision operates under Care Act 2014 authority with separate consent requirements. Sharing a patient’s clinical information with the local authority for care and support assessment requires a lawful basis that is different from the basis on which the information was originally collected. The patient may not understand that their information is about to cross a constitutional boundary — from a universal, free service to one that is means-tested and may require financial contribution.

Provenance. The clinical observations, assessments, and decisions that led to the discharge are held in the hospital’s electronic patient record. What crosses the boundary is typically a discharge summary — a compressed narrative that strips the clinical reasoning of its context. The social care assessor receives the summary but not the audit trail. If the patient’s condition changes after discharge, the community team cannot trace the clinical reasoning that underpinned the discharge decision.

Clinical Intent. The clinician’s judgment that the patient no longer meets criteria to reside is a clinical decision. But the post-discharge pathway is a cross-domain decision — it requires both clinical judgment about what the patient needs and social care judgment about what can be provided, under what legal authority, at what cost. The discharge pathway selection is the point where clinical intent meets constitutional boundary, and neither system has infrastructure to govern the junction.

Alert and Responsibility. This is where the boundary causes the most direct harm. At the moment of discharge, responsibility transfers — but the transfer is implicit, not explicit. The hospital’s duty of care ends when the patient leaves. The local authority’s duty to assess eligible needs begins when it becomes aware of the person’s appearance of need. In between — in the hours between discharge and first contact from community services — responsibility is held by no one, or by everyone, or by the patient themselves. The guidance advises “a safety and welfare check on the day of discharge for people on pathways 1, 2 and 3.” An advisory check is not an explicit responsibility transfer.

Service Routing. The patient must be routed to a service with both the capacity and the capability to meet their needs. Pathway selection (0, 1, 2, 3) is the coarsest possible routing decision — four categories for the full spectrum of post-acute need. Within each pathway, the actual service depends on local availability, commissioning arrangements, and the real-time capacity of providers that may be commissioned by the NHS, the local authority, or both. The care transfer hub exists to manage this routing — but as the data shows, care transfer hub processes account for 13.4% of delays. The routing infrastructure is itself a bottleneck.

Outcome. Does the hospital know what happened to the patient after discharge? In most cases, no. The discharge is recorded as complete when the patient leaves. Whether the social care package was actually delivered, whether the patient was readmitted within 48 hours, whether the placement met the patient’s actual needs — this information rarely flows back across the boundary to the discharging team. The loop is not closed. The sending organisation does not know whether its discharge decision resulted in a safe outcome.

Every one of these gaps exists because the crossing between hospital and social care is ungoverned as infrastructure. Each gap is managed — to the extent it is managed — through individual effort, local workaround, and professional relationship. The 28.2% “interface processes” delay is the measurable cost of managing structurally ungoverned crossings through negotiation rather than infrastructure.

The Discharge Crossing as Test Case for Boundary Infrastructure

Delayed discharge is not the only consequence of the ungoverned LA–NHS boundary. But it is the most measured, the most costed, and the most politically visible. It is the place where the boundary failure converts directly into numbers that NHS leaders, local authority directors, and politicians cannot ignore: beds occupied, money spent, A&E waits lengthened, elective procedures delayed.

It is also the highest-volume crossing. Every acute trust in England manages it every day. The NHS’s acute discharge situation report collects daily data from every trust. The data infrastructure exists to measure delay at this crossing with more granularity than at any other point on the LA–NHS boundary.

This makes the discharge crossing the natural test case for boundary governance infrastructure. If the Seven Flows can be implemented at the discharge crossing — if identity can be reconciled, consent can be governed, provenance can be preserved, clinical intent can be carried across the boundary, responsibility can be transferred explicitly, routing can be informed by real capacity, and outcomes can close the loop — then the same infrastructure can extend to every other LA–NHS crossing: CHC assessment, safeguarding referral, neighbourhood health integration.

The discharge crossing is also where the cost of inaction is most precisely quantifiable. At £62 million per month for interface processes alone — the crossing itself, not the capacity on either side — the business case for governed boundary infrastructure writes itself.

At Inference Clinical, we have built the Seven Flows audit methodology specifically for this analysis: mapping the governance maturity at each organisational boundary crossing, identifying where MVRT requirements are not met, and quantifying the risk at each gap. Applied to the discharge crossing, a Seven Flows assessment produces a structured understanding of which flows are failing, why, and what infrastructure would close them — documented in a form that any organisation on either side of the boundary can act on.

The organisations that treat delayed discharge as a capacity problem will continue to pursue capacity solutions — more beds, more carers, more funding — and those solutions will help at the margins. The organisations that recognise it as a boundary governance problem will build the infrastructure that makes the crossing itself faster, safer, and more reliable. Inference Clinical’s Boundary Risk Assessment applies the Seven Flows methodology to the discharge crossing, producing a structured Boundary Risk Score that quantifies where governance gaps cost the most. The £62 million monthly cost of “interface processes” is the price of the second approach not yet existing.

Julian Bradder

Founder & CEO, Inference Clinical

Building infrastructure for inter-organisation responsibility transfer in healthcare. 30 years’ experience in digital transformation across government, FTSE 100, and healthcare.