Week 8 of 10
Neighbourhood health centres
Estate is the most concrete part of the policy and the easiest to get wrong, because a building can lock in a service model for thirty years.
A guidance note with a date in it
The neighbourhood health centre guidance published in April 2026 is not a description of a service. It is a planning instruction with a deadline. NHS England tells regions and integrated care boards to "work in partnership to develop their full pipeline of schemes", and requires that by 28 May 2026 regions working with their ICBs set out, for each ICB, how centres will be organised across its footprint, where capital investment is likely to be prioritised, and the mix of upgrades and new builds proposed over time (Neighbourhood Health Centre Guidance, 2026).
The centre is defined before it is designed. Centres "will be expected to be open at least 12 hours a day and 6 days a week", and must "include on-site general practice as a core element of the model, not solely GP staff or sessional input" (Neighbourhood Health Centre Guidance, 2026). With general practice at the core, the guidance expects GP provision at the scale of a primary care network, serving a population footprint of around 50,000 people, though other services work at other scales.
Four estate archetypes are offered: hub and spoke, upgrading or extending existing NHS buildings, described as typically "the quickest and most affordable route"; repurposing community or civic space; cohort-specific hubs such as women's health hubs; and purpose-built centres. They are not "prescribed templates", and where suitable estate exists the guidance expects upgrading to be preferred to new build. The numbers sit in the Neighbourhood Health Framework: 250 centres by 2035, 120 of them by 2030, with 20 per cent of new builds funded from public capital and the rest through public private partnerships (Neighbourhood Health Framework, 2026). The first wave, for 2026/27, will largely repurpose existing NHS buildings in the most deprived areas.
Funding is released against seven criteria: strategic alignment with neighbourhood health objectives, coherence between the service model, general practice and the physical estate, intelligent use of existing estate before new build, deliverability, revenue affordability, governance and partnership maturity, and local alignment. The fifth decides whether a scheme survives contact with a finance director. Proposals must demonstrate "clear revenue (RDEL) affordability, including how running costs will be met within existing or agreed funding flows", covering workforce, unitary charges, occupancy and utilisation (Neighbourhood Health Centre Guidance, 2026). Approval panels were expected from early June 2026, with business cases then taken through a new Neighbourhood Estates Investment Committee.
The specification writes a service model into the walls
The design and performance specification published a day earlier is the more revealing document, because it does the opposite of what a technical rulebook does. It "sets out shared principles to guide design, rather than prescribing a single model" (Neighbourhood Health Centre Specification, 2026). Its own account of its status matters: it "does not introduce new burdens or compliance mechanisms", and "must" is "stipulated only where there is an existing legal or regulatory requirement". The opening-hours expectation is therefore a principle, not a statutory duty, and a scheme that treats it as a rule has read the document wrongly.
What the principles fix is the operating model. A centre is designed around "adaptable, shared spaces rather than fixed, service-specific rooms", with rooms "shared between services on a timetabled basis" and "all services operating at 80% utilisation of spaces" (Neighbourhood Health Centre Specification, 2026). The utilisation assumptions are explicit: a general practice population of 30,000 for a core centre, 50,000 for a core plus centre with community care, and 100,000 for a core plus plus centre with diagnostics. It has a single shared reception desk, a shared booking system, and a centre manager accountable for whether the promised benefits appear. The specification distinguishes an "unintegrated" model, where rooms are allocated to individual organisations and sized to each service's peak, from an "integrated" one, and states that centres "are designed to support the integrated model".
That is a service redesign specified ahead of the building, and several of its choices outlast the government that made them. Each centre is to be delivered as an all-electric, net zero energy building, is expected to be dementia-friendly from the outset, and is meant to be "long life, loose fit" so that it can change through its life. Read the two April documents together: the estate is meant to follow the service model, and the specification is that service model written down. The argument of this week is that this ordering is correct, and that it is the one most easily reversed, because a building is legible, fundable and visible while a service change is slow, contested and hard to photograph.
Colocation is an enabler, and only under conditions
The guidance asserts that "providing a range of services in a single, trusted location improves access for residents and supports real-time collaboration between professionals". The literature is more conditional. Lalani and Marshall evaluated integrated community care teams in a borough of East London, using 80 hours of participant observation and individual and group interviews, and found that co-location "can be an effective enabler for service integration providing a basis for joint working, fostering improved communication and information sharing if conditions such as shared information systems and professional cultures (shared beliefs and values) are met" (Lalani and Marshall, 2022, Health and Social Care in the Community). Organisations should not assume that "decades of silo working in health and social care and strong professional cultures will be resolved by co-location".
The wider reviews reach the same place. Painter and colleagues ran a systematic umbrella review of co-located services on primary care estates supporting disadvantaged groups, screening 2,626 studies and including 13 reviews. They found that co-located services can improve access, engagement and quality of care, but that the evidence was "inconsistent when considering the impact of co-location on efficiency" (Painter et al., 2026, Primary Health Care Research and Development). Pourat and colleagues measured integration directly in health centres that already had dental services on the same site, producing scores from 31% to 73% with a mean of 64% (Pourat et al., 2020, Journal of Evidence Based Dental Practice). Their title carries the point: colocation does not equal integration. Baskin and colleagues, evaluating co-located public mental health services across six case study sites in England, found that co-location reduced stigma and lowered barriers to access largely because the setting was not a clinical one (Baskin et al., 2023, BMC Public Health).
The twelve-hour day rests on evidence the requirement does not mention
Opening hours are the most concrete commitment in the policy and the thinnest evidenced. Cowling, Harris and Majeed analysed a national patient survey covering 903,357 respondents across 8,005 practices. Participation in the extended hours access scheme was associated with a 1.25 point increase in satisfaction with opening hours on a 100 point scale, rising to 2.08 points for patients who could not take time off work, and the authors concluded that participation had "a limited association with three patient experience measures" (Cowling et al., 2017, BMJ Quality and Safety). Burch and Whittaker found no association at all between practices offering extended access and patient experience, in a study that put the cost of extended access at an estimated 250 million pounds a year (Burch and Whittaker, 2022, BJGP Open).
A later case study explains why a hub behaves differently. Studying two large-scale providers, Burch and colleagues found that extended access was used to bolster a struggling service, that it "provides a different service to in-hours general practice", and that continuity is hard to deliver when the patient is seen at another site by another team (Burch et al., 2024, Journal of Health Services Research and Policy). Access also has a geography. Murphy and colleagues linked 25,408 extended access appointments across 158 practices in four clinical commissioning groups and found that use of the service was associated with distance, with practices further from the hub using it less (Murphy et al., 2021, International Journal of Population Data Science). A centre open for 12 hours but sited where people must travel to reach it can miss the patients the hours were meant to help, which is why the guidance prefers existing community focal points and warns against siting a centre away from one.
Polyclinics, and the last time a building led the service
England has run this experiment before. Under proposals for London, polyclinics were to serve populations of around 50,000 and co-locate primary, community and secondary care professionals. Imison, Naylor and Maybin reviewed the evidence and the experience of the LIFT programme for the King's Fund and reached a conclusion the 2026 guidance has restated rather than answered: "In practice, co-location alone is often not sufficient to generate co-working or integration of care" (Imison, Naylor and Maybin, 2008). Their reading of LIFT was that "bringing staff together in one place does not necessarily change the way in which they work", with the absence of clinical and managerial leadership the key factor. They also flagged that the larger model would mean patients, most of them elderly, travelling further.
The 2026 guidance is aware of that history and says so in one line: "Neighbourhood health is not limited to a building; it is an operating model." It adds that "the development of NHCs is not an end in itself", and that estate planning is "both a prerequisite and an accelerant". That is a more careful framing than the polyclinic programme managed, and it is also the framing most easily lost once a capital programme acquires a delivery director, a wave number and a biddable specification. The guidance cites no published evaluation of a neighbourhood health centre model, so the evidence base is being written while the estate is built.
How to test a scheme without turning the specification into a tick list
The specification makes the test easy to run badly, because it looks like a checklist. It is not one. It introduces no compliance mechanism, it uses "must" only where law already requires it, and it applies its principles proportionately to refurbishment and repurposing as well as to new build. The revenue question is what decides whether this is a service change or a building programme, and the capital record is unambiguous. Naylor's 2017 review of NHS property and estates set out its ordering in a single sentence: "The form of the estate must follow the service strategies evolving through local Sustainability and Transformation Plans" (Naylor, 2017). The same review recorded that capital investment was then "insufficient to fund transformation and maintain the current estate", with backlog maintenance at around 5 billion pounds, and that both PFI and LIFT had delivered modern buildings "but at a considerable cost in on-going financing and facilities payments".
The National Audit Office put figures on that cost. In 2016/17 there were 716 operational PFI and PF2 deals with a capital value of around 60 billion pounds; annual charges came to 10.3 billion pounds, and future charges running to the 2040s totalled 199 billion pounds. Providers holding PFI deals carried unitary charges ranging from 5.6% to 20.1% of turnover (National Audit Office, 2018). The 2026 guidance proposes a new public private partnership model for new build schemes, expects 80 per cent of new builds to be delivered that way, and states that the unitary charge paid to the private partner "typically covers the capital cost of the facility, hard facilities management, lifecycle maintenance and associated funding costs" (Neighbourhood Health Centre Guidance, 2026). A specification that fixes a floor plan to a service model for three decades is not a metaphor; it is a contract term.
The defensible way to use the specification is as a set of questions about ordering. Does the clinical brief exist before the site is chosen, or was the site chosen and the brief written to fit it? Which service change is the building enabling, and who is accountable for it? What did the seven funding criteria produce when the scheme was tested against them? If the scheme is a hub, what does the distance evidence say about who will use it, and what is being done for those who will not travel? And what will be measured, by whom, in the third year, when the rooms are either full or empty?
The contested question the week leaves open is whether capital investment in buildings should lead service redesign or follow it. The evidence gathered here points one way. Co-location enables integration only where shared records and professional cultures already exist, and it does not create them. Extended hours have not, on the English evidence, transformed patient experience. Polyclinics were judged on their buildings, and the buildings did not carry the case. Capital that arrives without revenue is a commitment of 25 to 30 years, and a building that opens before the service model is settled will decide the service model by default. The ordering that survives scrutiny is the one Naylor set out in 2017 and the guidance repeats in 2026: agree the service, then design the estate. The risk is that a programme with a wave number, a specification and a delivery director will proceed in the other order, and that the specification's flexibility will be used to justify whatever has already been drawn.
The contested question
Whether capital investment in buildings should lead service redesign or follow it.
Working across professions
One building, several doors
A shared building forces decisions that a shared plan avoids: reception, confidentiality, queuing, noise and who owns the front door. Run a design session in which each profession names one thing about the building that would break their work. The output is the negotiation, not the floor plan.
For your profession
Written for the groups City St George's trains. Read your own first, then the two nearest to you: the disagreements between them are the working material.
General practice and primary care (including physician associates)
Physician Associate Studies (Pre-Registration) MPAS
What changes for you. The practice becomes an anchor tenant rather than the owner of a building. The specification expects on-site general practice at the core of every centre, operating at the scale of a primary care network, and assumes a core centre sized around a general practice population of 30,000 with rooms shared between services on a timetabled basis. The practice list, its premises and who books its consulting rooms all move into the scheme business case.
The decision on your desk. Whether to give up the practice's own premises and occupy the centre on a short or medium term licence, or stay out and remain a spoke served by the same extended hours rota.
What goes wrong if this is done badly. The practice gives up its premises without securing the booking priority it assumed. Its consulting rooms are let to another service, and the practice meant to anchor the centre becomes a tenant competing for space.
Community nursing and district nursing
Public Health and Specialist Community Nursing MSc/PGDip
What changes for you. District nursing gains a team base in a building open at least 12 hours a day and 6 days a week, shared with adult social care and mental health colleagues. The specification sizes a core plus centre around a community care population of 50,000 and expects rooms to be booked rather than held, so the caseload is planned around a shared diary instead of a base the team controls.
The decision on your desk. Whether the team base moves into the centre or the service stays in existing community premises as a spoke, and who then holds the equipment, the keys and the records.
What goes wrong if this is done badly. A shared booking system replaces a base the team controlled. Visits are planned around building hours while the caseload is housebound and dispersed, and travel time grows in the hours meant for care.
Mental health nursing and mental health practice
Mental Health Nursing BSc (Hons)
What changes for you. The centre hosts primary care led mental health support, which the specification distinguishes from community based mental health centres that keep complex and acute work. Evidence from six co-located public mental health sites in England found that a non-clinical setting reduced stigma and lowered barriers to access. The practical change is a practitioner working alongside general practice rather than referring a patient out of the building.
The decision on your desk. Whether the mental health practitioner is based in the centre or keeps a community mental health centre base, and which patients are seen in primary care led provision.
What goes wrong if this is done badly. Primary care led provision absorbs the straightforward work while the specialist continuity the framework says community mental health centres retain is thinned. Crisis then arrives without the relationship that had kept it stable.
Speech and language therapy
Speech and Language Therapy BSc (Hons)
What changes for you. The specification offers standard treatment and group rooms, a bookable timetable and a shared reception desk, and says nothing about the acoustics, storage and equipment that swallowing and communication work requires. Rooms are shared across services and sized to a common model. A therapist holding dysphagia clinics and group sessions across children's and adult caseloads must negotiate for space rather than assume it.
The decision on your desk. Whether to accept the standard bookable room or decline to sign up to the centre until protected space with suitable acoustic treatment and equipment storage is designed in.
What goes wrong if this is done badly. A shared multi-purpose room cannot support safe swallowing assessment or confidential therapy. The caseload moves to a spoke or out of the building, and the profession disappears from the centre while remaining on the plan.
Adult social care
Partner profession: local authority adult social care
What changes for you. The specification has a section on social care integration, delivered through flexible non-clinical space, interview rooms and shared workspaces rather than fixed accommodation, and the guidance lists adult social care among the services to co-locate. The framework also states that co-location does not change local authority accountability or funding. The council is a partner in the building, not a service the building selects.
The decision on your desk. Whether the council takes flexible bookable space in a health led building on a licence, and what safeguarding and case records it keeps out of the shared system.
What goes wrong if this is done badly. Social care's presence is defined as flexible non-clinical space and shrinks as clinical demand for rooms grows. The council loses its own building without gaining secure space, and the warm handover has nowhere to happen.
Physiotherapy
Physiotherapy BSc (Hons)
What changes for you. The specification expects services such as physiotherapy to run from adaptable therapy and group spaces in the centre, as part of the shift of outpatient care out of hospitals. Rooms are booked and shared, not held, and the centre is designed for several sessions in one room across a longer day. Rehabilitation work moves from a hospital site into a neighbourhood building where the case for the space rests on utilisation.
The decision on your desk. Whether to move outpatient physiotherapy into a bookable therapy room in the centre or keep it on the hospital site and use the centre only for first contact work.
What goes wrong if this is done badly. Generic therapy rooms are designed without the equipment, plinths and exercise space a rehabilitation caseload needs. The rooms fill with consultation work, the muscle and joint caseload is displaced, and waits grow in a building that looks full.
Radiography (diagnostic and therapeutic)
Diagnostic Radiography BSc (Hons)
What changes for you. A core plus plus centre adds diagnostic space, including simple imaging such as ultrasound and mobile x-ray, alongside point of care testing, and the specification sizes diagnostics around a population of 100,000. It expects mobile diagnostic units to be accommodated rather than fixed suites. Radiographers therefore plan into a shared building whose imaging footprint is set by local activity modelling rather than hospital capacity.
The decision on your desk. Whether to base radiographers in the centre or keep imaging at the hospital and serve the centre with a mobile unit and visiting sessions.
What goes wrong if this is done badly. The centre is sized for mobile equipment and visiting sessions while the diagnostic waiting list needs fixed capacity. Walk-in diagnostics appear in the business case and the modalities that would clear the backlog are never installed.
Pharmacy (pharmacists and pharmacy technicians)
Pharmacy MPharm (Hons) (awarded by Kingston University, delivered jointly with St George's, University of London)
What changes for you. The specification prefers co-location with a nearby community pharmacy and states that fully embedded pharmacies are not expected as standard, because of their space and operational requirements. The framework wants pharmacies to be a first port of call, with prescribing trained pharmacists growing in number. Pharmacy therefore gains a formal place in the neighbourhood offer without a guaranteed place in the building.
The decision on your desk. Whether the community pharmacy takes space in the centre or stays on the high street and works through a co-ordinated referral route with the centre.
What goes wrong if this is done badly. The centre opens as a clinical building without pharmacy access on site. Patients collect medicines elsewhere, the one stop offer the guidance describes does not materialise, and prescribing support is delivered by remote contact.
Critical commentary
Where this week is argued with
Full-text commentaries and critical analyses from peer-reviewed journals. The policy column tests the claim against what government actually commissioned. The practice column tests it against what staff can deliver. Read one from each before the contested question.
Policy-facing
Perspective from a former health minister arguing that the top-down, uniform GP-led health centre programme damaged the case for co-location, while locally led Integrated Care Pilots fared better. It supports the week's claim that a building cannot carry integration, yet disputes that a better national framework is the answer. Perspective.
Editorial from the Health Foundation arguing the plan needs upfront cash as well as reform, noting that capital investment in buildings and equipment is essential when the maintenance backlog stands near fourteen billion pounds. It tests the week's ordering thesis by treating a funded capital settlement, not a better-designed estate, as the precondition for the community shift. Editorial.
Policy evidence brief setting four conditions for integrated neighbourhood teams to reduce inequalities, including co-located multi-sector services, shared IT and community trust. It argues that professional hierarchies and differing terms of employment fracture co-located teams, so it tests the week's thesis by locating the decisive barrier in culture and funding rather than the building. Policy evidence brief.
Practice-facing
Analysis arguing that the physical space of general practice actively shapes patient and staff outcomes, with practical recommendations on design, privacy and shared staff areas. It tests the week's ordering by claiming that space does not merely follow the service model but helps constitute care, so estate decisions belong in clinical planning. Analysis.
Khan N (2025). Access to general practice. British Journal of General Practice, 75(754), 220-221.
FULL TEXTEditorial arguing that access is not simply a supply of appointments, since triage, digital-first systems and multidisciplinary working can worsen or improve it unequally. It tests the week's twelve-hour centre by showing that extended hours and more slots have not raised patient experience, so practices should target the barriers that matter. Editorial.
Guest editorial reporting early results from a 24/7 neighbourhood mental health centre, where self-referral rose and home treatment referrals fell. It argues that a neighbourhood base can reshape access and crisis care, yet warns that partial implementation risks producing a rebranded community team, so the building helps only when fidelity and legacy disinvestment follow. Guest editorial.
Key sources
Qualitative evaluation of integrated community care teams in a borough of East London, drawing on 80 hours of participant observation and both individual and group interviews. It found that co-location enabled integration only where conditions such as shared information systems and shared professional values were met, and warned against assuming that decades of silo working would be resolved by sharing a building. It is the week's central evidence that co-location is a condition for integration rather than a cause of it.
Systematic umbrella review that screened 2,626 studies and included 13 reviews of co-located services on primary care estates supporting disadvantaged groups. It found that co-located services can improve access to care, engagement in treatment and quality of care, with gains reported for social determinants and mental health, but that the evidence was inconsistent on efficiency. It sets out the limits of what estate-led policy can claim for access.
Study that built a framework and scoring tool for integrating medical and dental services and applied it to 12 health centres in Los Angeles County that already provided dental care on site. Integration scores ranged from 31% to 73% with a mean of 64%, and depended on warm hand-offs, leadership support and linked records rather than on sharing premises. It supplies the week's sharpest single statement that co-location does not equal integration.
Realist evaluation of co-located public mental health services across six case study sites in England, using 65 interviews, four focus groups and two workshops. It set out five theories of how co-location works, including reduced stigma in non-clinical settings, lower cost and time barriers to access, and better sustainability through pooled resources. It is the week's evidence on what a community setting contributes that a clinical one does not.
Analysis of the General Practice Patient Survey covering 903,357 respondents and 8,005 practices, comparing patients at practices that took part in the extended hours access scheme with patients at practices that did not. Participation was associated with an increase of 1.25 points in satisfaction with opening hours on a 100 point scale, rising to 2.08 points for patients who could not take time off work, which the authors judged a limited association. It is the week's caution against treating extended opening as a proven route to better access.
Observational study using the English GP Patient Survey in 2018 and 2019 to test whether practices offering extended access had better patient experience at practice population level. It found no association with satisfaction with access, continuity of care or overall satisfaction, and noted that extended access costs an estimated 250 million pounds a year. It is the most direct available test of the assumption behind a twelve-hour centre.
Qualitative comparative case study of two large providers of supra-practice extended access, using clinician and patient observation, interviews with staff, managers, commissioners and patients, and routinely collected appointment data. It found that extended access was used to bolster a struggling primary care system, provided a different service from in-hours general practice, and struggled to deliver continuity for patients with complex problems. It explains why a hub can be open for twelve hours and still not be the same service.
Study linking 25,408 extended access appointments across 158 general practices in four clinical commissioning groups to the General Practice Patient Survey and area deprivation data. It found that use of the extended hours service was associated with distance, with practices further from the hub using it less, and that deprivation was not associated with rates of use. It is the week's evidence that a centre open for twelve hours can still miss the patients it was meant to reach.
Audit office briefing on the Private Finance Initiative and its successor, covering 716 operational deals with a capital value of around 60 billion pounds, annual charges of 10.3 billion pounds in 2016/17 and future charges of 199 billion pounds running to the 2040s. It records that health providers holding PFI deals carried unitary charges ranging from 5.6% to 20.1% of turnover. It supplies the week's figures for capital that arrives with a revenue commitment attached.
Independent review of NHS property and estates that states the ordering this week argues for, that the form of the estate must follow the service strategies evolving through local plans. It found the capital budget insufficient to fund transformation and maintain the estate, put backlog maintenance at around 5 billion pounds, and recorded that PFI and LIFT had delivered modern buildings at a considerable cost in ongoing financing and facilities payments. It is the week's clearest statement of the estate-follows-service principle.
King's Fund review of whether polyclinics would deliver integrated care, drawing on published literature and original research into facilities similar to the polyclinic model, including the LIFT programme for new general practice premises. It concluded that co-location alone is often not sufficient to generate co-working or integration, and that bringing staff together in one place does not necessarily change how they work and can make joint working harder. It is the week's account of the last English attempt to let a building lead service change.
National guidance published in April 2026 that instructs regions and integrated care boards on the planning work required to develop a pipeline of neighbourhood health centre schemes, including the 28 May 2026 submission, the four estate archetypes, the seven funding criteria and the metrics against which schemes are assessed. It expects centres to open at least 12 hours a day and 6 days a week, with on-site general practice at the core, and requires revenue affordability within existing or agreed funding flows. It is the week's primary planning document.
Design and performance specification for new-build neighbourhood health centres, published in April 2026, setting out the integrated operating model, the core, core plus and core plus plus centre types with their utilisation populations of 30,000, 50,000 and 100,000, the 80% utilisation assumption, shared bookable rooms, the single reception and the centre manager role. It states that it introduces no new burdens or compliance mechanisms and uses the word must only where law already requires it. It is the document that writes the service model into the building.
Framework published in March 2026 that sets out the delivery expectations for neighbourhood health, including the commitment to 250 neighbourhood health centres by 2035 and 120 by 2030, with 20 per cent of new builds funded from public capital and the rest through public private partnerships. It records that the first wave for 2026/27 will largely repurpose existing NHS buildings across NHS Property Services and LIFT estates in the most deprived areas, and that the form of the estate and the quality of existing buildings are central constraints. It sets the pipeline numbers this week tests.