A community nurse parks on a terraced street at half past eight in the morning. In the boot of her car sits a soft bag of testing kit: a small handheld analyser, a strip of cartridges, a lancet, a control vial that has spent the night at whatever temperature the car park offered. Inside the house is a man in his seventies with a chest infection, breathless, being managed as an acute patient without ever leaving his own front room. She will take a sample, run a test on his kitchen table, and a number will appear. On that number, a doctor sitting in a hub several miles away will decide whether he stays at home tonight or is brought into hospital. The ward, in every sense that matters clinically, is that living room. The laboratory is the boot of a car.
This is not a thought experiment. It is one of the largest shifts in how the NHS delivers care in a generation. Virtual wards, also called hospital at home, and urgent community response teams are expanding rapidly, and they are genuinely transformational: they let people receive acute-level care in their own homes, they ease pressure on hospital beds, and most patients would far rather be at home than on a trolley in a corridor. I am an enthusiast for the model. But there is a quiet dependency inside it that rarely gets the attention it deserves, and it is the whole reason the model can work at all.
Here is my thesis. You cannot safely run an acute patient at home without the ability to measure them at home, and the moment you can measure them, you have moved a laboratory into the community whether you planned to or not. Virtual wards run on point-of-care testing. Yet in far too many virtual ward designs, the diagnostics and the quality governance around them are an afterthought, bolted on once the pathway is already live. Get the testing right and virtual wards are a triumph. Get it wrong and you have not saved anyone; you have simply relocated clinical risk from a monitored ward into an unmonitored living room.
What a virtual ward actually is, and why it is spreading
The idea is deceptively simple. Take a patient who would otherwise occupy a hospital bed, someone with a chest infection, a flare of heart failure, a frailty crisis, an infection that needs intravenous antibiotics, and manage them at home instead, with a remote clinical team providing oversight and a mix of in-person visits and remote monitoring. Urgent community response teams sit alongside this, responding within hours to people at risk of a hospital admission and trying to prevent it. NHS England has backed both as core parts of delivery, not fringe experiments.
The drivers are not hard to see. Beds are scarce and expensive. Deconditioning and hospital-acquired infection are real harms of admission, especially for older people. And the technology to keep a distant eye on someone, wearables, remote observations, video review, has matured to the point where continuous supervision no longer requires a physical bay. The direction of travel is right, and I want to say that clearly before I complain about anything.
But observations are not diagnostics. A pulse oximeter and a blood pressure cuff tell you how a patient is trending. They do not tell you the patient's potassium, their kidney function, their inflammatory markers, their blood gas, their clotting. And for an acute patient, those numbers are frequently the ones that decide the next move. Which is exactly where point-of-care testing walks onto the stage.
The diagnostics that quietly power the ward
Think about what a hospital does to an acute inpatient that a set of vital signs alone cannot replace. It takes bloods. It runs a gas. It checks renal function before the next dose of a nephrotoxic drug. It measures an infection marker to decide whether this is settling or escalating. It checks an INR before adjusting anticoagulation. It measures glucose in someone whose diabetes is being buffeted by acute illness. Move that patient home and none of those questions go away. What changes is only where, and by whom, the answer is generated.
Point-of-care testing is what makes that possible. At the bedside-at-home, a virtual ward may need rapid biochemistry, blood gases, infection markers, kidney function, glucose and coagulation, generated in minutes, without a sample making a round trip to a central laboratory that would take too long to inform a decision that has to be made now. This is the invisible engine of the whole model. Without it, a virtual ward is a monitoring service that can watch a patient deteriorate but cannot say why or how fast, and cannot act with confidence.
And here is the thing that ought to focus every commissioner's mind: NHS England has itself published guidance on integrating in vitro point-of-care diagnostics into urgent community response and virtual ward services. That guidance is explicit that pathology teams have a role in oversight, including quality control and the governance of devices and consumables. This is not a niche opinion held by a few quality obsessives. It is national guidance saying, in effect, that the laboratory does not stop being responsible just because the sample was taken on a sofa.
A virtual ward without diagnostics is a monitoring service that can watch a patient deteriorate but cannot say why. The point-of-care test is the engine, not the accessory.
Who owns quality when the ward is a whole town?
This is the hard part, and it is the part that gets skipped. In a hospital laboratory, the accountability for a result is unambiguous. There is a named quality manager, a physical environment with controlled temperature, a competency file for every operator, a quality control regime run to a schedule, external quality assessment enrolment, and a documented chain from sample to signed-out result. Now scatter that ward across a town.
The devices no longer live in a controlled room. They live in cars, in cupboards, in kit bags, subject to heat, cold and vibration that a laboratory would never tolerate. The operators are no longer a small trained team you can see; they are a mobile workforce of community nurses spread across shifts and geographies, and sometimes, increasingly, the patients themselves or their relatives. The results no longer travel down a corridor to a hub; they have to be transmitted, reliably and quickly, to a remote clinical team that may never physically meet the patient. Every one of the quiet assumptions that make laboratory governance work has been broken, not by carelessness, but by the geography of the model itself.
So ask the awkward questions early. Who is accountable for the quality control on a device that lives in a nurse's car? Who confirms that a community nurse is still competent to run this test six months after their initial sign-off, when nobody is watching them do it? Who notices when a result never reaches the hub because a phone had no signal in a rural dead spot? Who owns the consumable that was stored above its rated temperature over a hot bank holiday weekend? If the honest answer to any of these is "nobody in particular", you do not have a virtual ward. You have an untested pathway that has not failed yet.
The standards did not move; only the address did
It is tempting to argue that a test done in a patient's home is a different, lighter category than a test done in a hospital. It is not, and the standards are unusually clear on this. ISO 15189:2022 sets out the principles that make a result trustworthy, competency, quality control, external quality assessment and comparability of results across sites and devices, and those principles apply wherever the testing happens. A potassium measured on a kitchen table and a potassium measured in a core laboratory are the same clinical object, used to make the same kind of decision, and if we would not release the hospital one without governance behind it, we cannot rationally release the home one without it either.
Comparability deserves special emphasis in the virtual ward context. A patient may be tested on one handheld device during a home visit on Monday, a different device carried by a different nurse on Wednesday, and a hospital analyser if they are eventually admitted on Friday. If those devices do not agree, a genuinely stable patient can look like a deteriorating one, or worse, a deteriorating one can look stable. Harmonisation across the fleet is not a technical nicety. It is the difference between a trend you can trust and a trend that is an artefact of which box happened to be in the boot that day.
Ward pathway versus virtual ward pathway
The clearest way to see the challenge is to lay the two pathways side by side. In hospital, the diagnostic decision points, take the sample, run the test, review the result, decide, escalate, all happen inside one building with one accountable structure and a named owner at each step. In a virtual ward, the very same decision points still exist, but they are now distributed across a car, a living room, a mobile network and a remote hub, and the accountability at each step has to be deliberately assigned, because it no longer comes for free with the walls.
Nothing in that comparison is an argument against virtual wards. It is an argument for designing the diagnostics in from the first whiteboard sketch, rather than discovering, once volumes rise, that the pathway was built around a device nobody was formally governing. Retrofitting quality onto a live acute-at-home service is far harder, and far more dangerous, than building it in at the start.
Design the diagnostics in from day one
If you are commissioning, building or running a virtual ward or an urgent community response service, treat the testing and its governance as a founding component, not a later procurement line. Practical steps, in roughly the order I would take them:
- Choose devices for the environment, not the brochure. A device that performs beautifully on a laboratory bench may not tolerate a car boot in July. Select for ruggedness, temperature tolerance, ease of operation by a non-laboratory workforce, and the specific analytes your patient cohort actually needs, blood gases, renal function, infection markers, INR, glucose and the rest.
- Build connectivity in, so results are visible centrally. A result that stays on the device, or on a scrap of paper in a nurse's pocket, is not a virtual ward result. Design from the outset so that every test is transmitted to the hub, timestamped, attributed and visible to the clinical team that has to act on it, with a plan for what happens when a signal drops.
- Put a real competency framework behind a mobile workforce. Initial training against the instructions for use, an assessed sign-off, and, crucially, a way to reassess and re-evidence competency for people you cannot stand next to. This is one of the hardest problems in the model, and it will not solve itself.
- Run quality control in the field, and make it enforceable. Define the frequency, decide how controls are stored and transported, record every result, and set a rule that stops testing on a device when a control fails, wherever that device happens to be. A QC regime that is optional in the field is decorative.
- Write the escalation route down and rehearse it. Name who an abnormal result reaches, how fast, and who owns the decision to admit. In an acute-at-home service the escalation route is the safety net, and a net that exists only in principle is no net at all.
- Give pathology a seat at the design table. National guidance is right that pathology teams should oversee quality control and device governance. Bring them in on day one, not after the first near miss.
None of this requires turning a patient's living room into a hospital laboratory. It requires accepting that the ordinary, unglamorous discipline of laboratory medicine, competency, quality control, external quality assessment, comparability and escalation, has to travel with the ward when the ward leaves the building. If you want to understand which analytes carry which risks in a community setting, our analyte guides are a good starting point. Our training, including the free POCT Fundamentals course, is built for exactly the mobile, non-laboratory workforce these services depend on. And when you are standing a service up, our consultancy helps commissioners design the diagnostics and their governance in from the beginning, rather than bolting them on once the pathway is already carrying real patients.
The virtual ward is one of the best ideas in UK healthcare. Whether it saves people or quietly endangers them turns on a question nobody wants to own: who governs the test?
Move the ward home. Take the laboratory's discipline with it.
Virtual wards are not a fad, and they are not a compromise. Done well, they are better care in a better place, and I would not slow their spread for anything. But better care depends on trustworthy numbers, and a number you cannot trust is worse than no number, because it is acted upon by a clinician who cannot see the patient. We have moved the ward into the living room. Now we have to move the laboratory's discipline in with it, deliberately, on day one, with a named owner for every result. That is not bureaucracy getting in the way of innovation. That is the thing that lets the innovation be safe. It always was.