A patient is being discharged from hospital next Tuesday. Her readmission risk is shaped less by her medication regime than by the temperature of her living room on Sunday night. The Building Research Establishment puts the first-year treatment cost to the NHS of poor housing in England at £1.4bn, with £857m attributable to excess cold alone (BRE, 2023). Two boards in her local system hold direct authority over those numbers. Neither one is on the other's agenda.
This is not an evidence problem. The NHS lost roughly 121,695 bed days in 2024 to 2025 to mental health discharge delays driven primarily by housing shortfalls, at an estimated cost of £102m (Social Care Today, 2026). Total delayed-discharge cost across the NHS is now running at approximately £2bn a year (The Lowdown, 2025). The unit cost of an English NHS bed day in 2025 to 2026 sits at £562 (Parallel Parliament, 2025). The diagnosis writes itself. Housing is the prevention infrastructure the NHS already relies on by accident, and the NHS is the customer the housing sector has never been allowed to name.
What the housing CEO is not being told
A social housing CEO sitting on a stock retrofit business case has been told it is a cost programme, funded by Warm Homes and Social Housing Decarbonisation Fund tranches. The CEO has not been told the same retrofit is a revenue-eligible intervention against a defined NHS prevention outcome.
The 2025 to 2026 Better Care Fund framework names local housing authorities as required signatories to joint plans, carrying £5.6bn of NHS minimum contribution and £2.6bn of Local Authority Better Care Grant (NHS England, 2025). The forthcoming 10-Year Health Plan goes further, opening the door to budget pooling and reprofiling toward prevention in named demonstrator areas, with Greater Manchester first in the queue (House of Commons Library, 2025). Retrofit stops being a one-way cost when its outcomes are bilaterally measured.
A registered provider with verifiable data on heating performance, tenant respiratory admissions, and the readmission profile of recently discharged residents owns the only artefact the ICB needs to commission against. Without the data, the conversation stays at sentiment. With it, the conversation moves to a Section 75 pooled budget against a measurable cohort. The line on the housing balance sheet shifts from grant-funded asset improvement to NHS-funded prevention income.
What the ICB Chair is not being asked
An ICB Chair under pressure on delayed-discharge figures has typically responded by buying more domiciliary care, more discharge-to-assess capacity, and more reablement weeks. None of those reduces the rate at which the same patient returns to A&E in February. Supported housing, by contrast, runs at roughly a third of the cost of an inpatient bed and is associated with NHS savings of between £53m and £65m a year for mental health populations alone (Social Care Today, 2026).
Virtual wards deliver clinical outcomes equivalent to or better than inpatient care, but only where the home environment supports the model (Parliamentary Office of Science and Technology, 2024). The home environment is the registered provider's product. The NHS spends £562 a bed day to substitute for a building it would rather not be using. The registered provider has the building and is being asked to bid against itself for capital grant.
An ICB writing a five-year prevention contract with three or four large registered providers in its footprint is buying virtual ward capacity at a price no acute trust is able to match. The unit economics are not subtle.
The plumbing already exists
Gentoo Group and Sunderland Clinical Commissioning Group ran a scheme called Boilers on Prescription a decade ago and recorded a 60 per cent reduction in GP appointments for participating tenants, with A&E attendances down 30 per cent and emergency admissions down 25 per cent (Gentoo Group, 2016). The evidence has been there since 2014. The legal vehicle exists in Section 75 of the NHS Act 2006, which has been under government review since 2023 with explicit proposals to widen the scope of pooled budgets to include housing providers as eligible parties (Mills and Reeve, 2023). The funding exists in the Better Care Fund. The political opening exists in the 10-Year Plan and the prevention demonstrators.
What does not exist, in most parts of England, is the meeting.
Two C-suites with overlapping populations, overlapping outcomes, and entirely separate board packs. The first registered provider and the first ICB to close that gap with data and a contract, rather than another memorandum of understanding, will rewrite their commissioning narrative for the rest of the decade. Everyone else will continue to discuss it at conference.
If you are a housing CEO or an ICB Chair reading this, the question worth taking into your next board is a simple one: do you know the name of your counterpart in the other system, and have you ever sat in a room with one shared cohort, one shared outcome metric, and one pooled budget on the table?