In short: NHS WiFi has quietly become clinical infrastructure — electronic patient records, e-observations and bedside devices all depend on it — but most hospital wireless estates were designed as an office convenience, running over an ageing and radio-hostile building stock. The result is dead zones and dropouts exactly where care happens, and the fix is to treat clinical connectivity as critical infrastructure: engineered coverage, separated traffic, and licensed-spectrum private networks where WiFi cannot reach.
Key Takeaways
- The EPR rollout turned WiFi into clinical infrastructure — with around nine in ten NHS trusts now running electronic patient records, drug charts and observations live on wireless devices at the bedside, so a WiFi dropout is no longer an IT ticket but a clinical safety issue.
- Hospital buildings are engineered against radio — Victorian wings, 1960s concrete, lead-lined imaging suites and fire doors create dead zones that no amount of access points fully cure, on an estate carrying a maintenance backlog of well over £13 billion.
- Trusts are separating clinical traffic onto engineered networks — dedicated coverage for EPR carts, e-obs and patient flow, with private 5G filling the spaces WiFi cannot serve, leaves patient and visitor WiFi as an amenity rather than a competing load.
In a nutshell
Why is NHS WiFi so unreliable?
Ask a ward nurse about NHS WiFi and you will get a specific, practical answer: the computer-on-wheels drops its connection outside bay four, the e-observations tablet has to be carried to the corridor to sync, and the barcode scanner for drug administration works everywhere except the side rooms where it is needed most. Ask a patient and you get a different answer about the login page and the buffering. Both are describing the same underlying problem from different ends of it.
That problem is worth stating plainly. Over the past decade, and at accelerating pace since NHS England's frontline digitisation push, the hospital ward round has become a wireless activity. Electronic patient records from the likes of Epic, Oracle Health and System C now run live at the bedside in the large majority of trusts — the national ambition was 90% EPR coverage by 2025, and the big go-lives at Manchester, Frimley, and Guy's and St Thomas' made the headlines on the way. Every one of those systems assumes a device in a clinician's hand with a working connection. Electronic prescribing, e-observations feeding early-warning scores, closed-loop medicines administration, VoIP handsets replacing bleeps, porter and patient-flow tracking: all wireless, all clinical, all riding on infrastructure that in many hospitals was specified when WiFi meant email in the office.
So when NHS WiFi fails, it no longer fails as a convenience. It fails as clinical infrastructure. That reframing — from amenity to utility — is the single most useful shift a trust's digital team can make, because it changes what "good enough" means.
What does the ward actually ask of the network?
It helps to be concrete about the load. A typical acute ward now carries: EPR sessions on carts and handhelds at every bed space, observation devices syncing every few minutes, barcode medication scanning that must work inside every side room, staff VoIP and task-management apps that follow clinicians as they walk, infusion pumps and monitors with wireless telemetry, and location tags on equipment that goes missing on foot. Layer onto that the patient and visitor WiFi — a service the NHS committed to providing free across its estate back in the 2016-2019 NHS WiFi programme — and the media-heavy expectations of a ward full of smartphones.
Two features of this load are unforgiving. The first is mobility: clinical devices move constantly, between bays, through fire doors, into lifts, and WiFi's handover between access points is precisely its weakest behaviour — the walking clinician's dropped session is a protocol characteristic, not a configuration error. The second is criticality mixing: a visitor's video call and a nurse's drug round currently contend for the same unlicensed spectrum, and on a contended channel WiFi offers no real way to guarantee the drug round wins.
Why do hospital buildings defeat wireless?
Even a well-designed WiFi estate then meets the NHS's real enemy: the buildings. The hospital estate spans Victorian pavilion wards with metre-thick masonry, 1960s and 70s slab-concrete blocks, endless retrofitted partitions, lead-lined X-ray and CT suites, Faraday-caged MRI rooms, basements full of pathology and pharmacy, and fire doors everywhere. Each is a radio obstacle; together they produce the dead-zone map every hospital IT team knows by heart. And this is an estate with a maintenance backlog that NHS Digital's own estates returns put well north of £13 billion — wireless refresh competes for capital with leaking roofs and failing lifts, and understandably loses.
The honest engineering conclusion is that in parts of many hospitals, WiFi alone will never deliver clinical-grade coverage at a defensible cost. The question is what to do about that, and the answer emerging across the sector has three layers.
What does fixing NHS WiFi actually involve?
Firstly, engineer for the clinical load, not the floor plan. That means radio surveys done on working wards rather than empty ones, coverage specified room-by-room including side rooms, lifts and stairwells, and the clinical device estate — carts, scanners, pumps — treated as the design driver. Several trusts have found that a survey-led remediation of a genuinely bad ward block costs less than the incident reviews the dead zones were generating.
Secondly, separate the traffic. Patient and visitor WiFi is a legitimate service — for many patients it is their only link to family, and for virtual-ward step-downs it matters clinically too — but it should not contend with the drug round. Clinical systems belong on their own engineered layer with reserved capacity, so that a full waiting room streaming football cannot slow an EPR cart upstairs.
Thirdly, use licensed spectrum where WiFi cannot win. This is where private 5G has moved from pilot to practical. Ofcom's shared access licences give a trust interference-free spectrum within its own estate; cellular's handover and building penetration are designed for exactly the moving-device, hostile-building problem wards present; and a private network extends naturally beyond the walls — to the ambulance deck, the multi-storey car park where community teams sync their visits, the portacabin clinics, and the community diagnostic centres and satellite sites that hospital-grade imaging is migrating to. We have made the case elsewhere that virtual wards and lone-working community staff hinge on connectivity nobody currently owns; inside the acute estate, the same discipline applies — someone has to own the wireless layer as infrastructure, with an SLA a clinician would recognise.
None of this is exotic. It is the same layered pattern that factories and airports have adopted: WiFi for what WiFi does well, licensed spectrum for what it does not, and a deliberate decision about which traffic rides where.
Can stretched NHS budgets afford this?
The candid answer is that the money is tight and the business case must be honest. A full private 5G overlay of a large teaching hospital is a significant project, and for some trusts a disciplined WiFi remediation will capture most of the benefit. But three things shift the arithmetic. Downtime already has a cost: reverting a ward to paper because the network dropped consumes nursing hours and creates transcription risk that safety teams can price. The capital comparison is not 5G versus nothing, it is 5G versus recabling and re-surveying a listed Victorian block, and licensed wireless frequently wins that comparison on the hardest 20% of the estate. And a managed-service model — the network delivered as a monthly service with uptime commitments, rather than a capital project the trust must staff — matches how stretched digital teams can actually consume infrastructure.
It is also worth saying what will not work: another round of access points sprinkled onto the existing design, procured as commodity IT. That is the cycle many trusts have been through twice already, and the dead zones survive it every time.
Where should a trust start?
Start where the risk is: ask your clinical safety officer and your ward managers where connectivity already features in incident reports and workarounds, and put a surveyor on those wards during a working shift. That gives you a map of the estate where WiFi can be remediated and the harder spaces where it cannot, and the second list is usually shorter and more affordable than feared.
We build and manage private networks for health and care settings — community sites, care homes and the awkward corners of acute estates — at a price point set for stretched public-sector budgets rather than for hyperscale hospital programmes. If your EPR is only as good as your worst ward's signal, talk to us about surveying it.