Walk onto any large acute hospital site and ask a simple question: right now, this minute, exactly who is authorised and currently competent to run a blood gas on that analyser in the emergency department? Not who was trained at some point. Not who attended the roadshow last spring. Who, today, has a valid, in-date competency for that specific device.
In most trusts, nobody can answer that question quickly, and a surprising number cannot answer it at all. The information is scattered across a shared spreadsheet, a filing cabinet, a departmental inbox, and the memory of one or two people who are on annual leave. Meanwhile the analyser keeps producing results, and clinicians keep acting on them, because the machine has no idea who is standing in front of it.
This is the quiet scandal at the heart of point-of-care testing. We have poured effort into device connectivity, quality control rules and result thresholds, and we have left the single most common failure mode almost entirely to goodwill. My argument is blunt: POCT competency is a systems problem, not a training-day problem, and if you cannot show me today, on demand, exactly who is competent to test on each device, you do not have a competency system. You have a filing cabinet.
The arithmetic nobody wants to do
Start with the numbers, because the numbers are where the comfortable assumptions fall apart. POCT looks small when you picture it: a handful of dedicated operators, a friendly coordinator, a couple of devices. That picture was accurate twenty years ago. It is a fantasy now.
A single large acute trust can carry several thousand registered POCT operators across its device fleet. Glucose meters live on every ward. Blood gas analysers sit in emergency, theatres, critical care and maternity. Urinalysis, coagulation, HbA1c, infectious disease and pregnancy testing spread the operator base wider still. Every nurse, every healthcare assistant, every junior doctor who touches a device is an operator who must be trained, assessed and periodically reassessed. That is what structured POCT training actually has to cover.
Now do the multiplication, illustratively. Say you have three thousand operators. Say each one needs reassessment once a year to stay current. Say a proper reassessment, done well, takes forty minutes of a coordinator's time when you include the observation, the record, the follow-up and the inevitable chasing. Three thousand times forty minutes is two thousand hours of coordinator time, every year, on reassessment alone, before you have trained a single new starter.
Two thousand hours is roughly one and a quarter full-time staff doing nothing but reassessment, all year. Most POCT services run on one or two coordinators who are also managing connectivity, quality control, audits, procurement, incident investigation and the standing tide of email. The demand does not fit the capacity. It never fitted. We just did not measure it.
And that is the static picture. It ignores the thing that makes POCT competency genuinely unmanageable by hand: churn.
Churn is the real enemy
A hospital is not a stable population of operators. It is a river. Bank staff appear for a single shift. Agency nurses cover a gap and are gone by Friday. Junior doctors rotate on a national timetable that empties and refills whole departments twice a year. Return-to-practice nurses, international recruits, students on placement, redeployed staff during a winter surge: every one of them may need to run a test, and every one of them arrives outside your tidy annual cycle.
The spreadsheet model assumes a birthday-cake world where everyone is assessed on a neat anniversary. Reality is a firehose of arrivals and departures that never pauses. By the time you have updated the sheet for this week's starters, next week's cohort has already tested on devices they were never formally signed off to use, because the ward was busy and the machine let them.
The machine has no idea who is standing in front of it, and in most trusts neither does the record.
This is why "we did a training day" is one of the most dangerous phrases in the whole field. A training day is an event. Competency is a state. Confusing the two is like confusing a wedding with a marriage. The roadshow was well attended, the sign-in sheet is full, and six weeks later half the attendees have moved wards, a third have never touched the device since, and the sign-in sheet proves attendance, not ability. Attendance is not competency, and a signature on a register is not a demonstration of skill.
Why competency decays, and why nobody notices
Competency is not a permanent acquisition. It decays. An operator who was genuinely excellent in March, assessed properly and performing well, will have drifted by the following March if they have used the device rarely, if the workflow changed, if a new lot of strips behaves differently, or if they simply forgot the step that matters. This is not a moral failing. It is how human skill works, and it is precisely why international standards require periodic reassessment rather than one-off certification.
ISO 15189:2022, the standard against which medical laboratories and their point-of-care activities are assessed, requires that the people performing testing, including non-laboratory staff, are trained, have their competency assessed, and are reassessed at defined intervals. That is not a bureaucratic nicety. It exists because the evidence of decades is that skill fades and error creeps back in, and the only defence is to check again.
Picture competency over time as a sawtooth. Each reassessment resets an operator back up toward full currency, and then it declines until the next check. Managed well, the troughs never fall into dangerous territory. But when a reassessment slips, because the coordinator is stretched, because the operator was on leave, because nobody flagged it, the line keeps falling and crosses into a zone where the person is lapsed but still testing. And here is the point that should keep quality managers awake: nothing stops them.
The device does not lock. The result does not carry a warning. The clinician downstream sees a number, not a competency status. A lapsed operator produces results that look identical to a current operator's results, which is exactly why competency lapse is the most quietly dangerous POCT failure there is. It does not announce itself with an alarm. It surfaces, if it surfaces at all, as a subtle drift in quality, a cluster of odd results, an incident review months later that traces back to a hands-on step done wrong by someone whose sign-off expired last spring.
The spreadsheet was never going to hold
None of this is a criticism of the people running POCT services. They are, in my experience, some of the most conscientious professionals in the health service, holding a vast operator base together with spreadsheets, reminders and sheer determination. The criticism is of the method. A manual, document-based model cannot scale to thousands of operators churning constantly, and pretending otherwise is how organisations end up exposed.
Consider what the spreadsheet cannot do. It cannot stop a lapsed operator from testing, because it is a passive record, not an active control. It cannot tell you your live compliance position without someone manually counting rows. It cannot distinguish the operator who runs forty tests a day from the one who ran one test last year, so it treats both as needing the same reassessment on the same schedule, which is simultaneously wasteful and unsafe. It cannot survive the coordinator being off sick, because the knowledge lives partly in their head. And it cannot give you, at an audit or a CQC inspection or an incident review, the one thing that matters: proof, on the day, of exactly who was competent to do what.
A filing cabinet answers the question "can we find a record of training if we look". A competency system answers the question "who is competent, right now, to test on this device". Those are different questions, and only the second one keeps patients safe.
What a real competency system looks like
If the manual model is broken, what replaces it? Not another spreadsheet with more columns. A systems approach, where competency is designed into the way testing happens rather than bolted on afterward. Here is what that means in practice, and what I would push any POCT service to build toward.
- Role-based access and device lockout. The single highest-value change. If an operator is not current for a specific device, the device should decline to accept them, or the result should be flagged and quarantined. The control has to live at the point of testing, not in a report read weeks later. When lockout is real, lapse stops being a silent risk and becomes an immediate, visible prompt to reassess.
- Competency built into the workflow. Reassessment should be a small, frequent, low-friction event woven into normal work, not an annual ordeal that everyone dreads and defers. Short observed checks, quick knowledge confirmations at the device, and evidence captured as testing happens will always beat a once-a-year marathon that half the workforce misses.
- Train-the-trainer cascades. One or two coordinators cannot personally assess thousands of people. They should not try. Build a network of ward-based or department-based assessors, trained and quality-assured by the POCT team, so competency propagates through the organisation instead of bottlenecking on two individuals. This is the only way the arithmetic ever balances. Our consultancy work almost always starts here, because it is the change that unlocks all the others.
- Risk-based reassessment depth. Not every operator and not every device carries the same risk. A high-throughput blood gas in critical care is not a monthly pregnancy test on a quiet ward. Match the depth and frequency of reassessment to the clinical risk and the operator's actual usage, so effort goes where harm is most likely rather than being spread thinly and uselessly across everything.
- Live visibility of current competency. Leadership, quality managers and coordinators should see, at a glance, the real-time compliance position: who is current, who is lapsing this month, which devices have coverage gaps, which wards are drifting. Not a report you commission and wait a fortnight for. A dashboard that is true when you look at it.
Notice what these have in common. Every one of them moves competency from a passive record toward an active control, and from a coordinator's private burden toward an organisational system. That is the whole shift. You can support it with good foundational training and sensible documentation, but the templates and courses are the raw materials, not the machine. The machine is the system that makes competency visible, enforceable and shared.
What to do on Monday
You do not have to boil the ocean. You have to change the questions your service can answer. Here is where I would start.
- Count your true operator base. Not the ones on the training list. Every person who could physically run a test on a live device this month. The gap between that number and your reassessment capacity is the size of your problem, and you cannot fix a problem you have not measured.
- Find your lapsed-but-testing population. Cross-reference who is out of date against who has actually run tests recently. If you cannot do this in an afternoon, that itself is the finding, and it tells you your record is a filing cabinet, not a system.
- Pick one high-risk device and make lockout real on it. Prove the principle somewhere it matters most, then spread it. A single blood gas analyser that refuses lapsed operators teaches the whole organisation what a control feels like.
- Stand up a train-the-trainer network. Identify credible assessors in each area, train and quality-assure them, and stop trying to do everything centrally. This is the change that makes the arithmetic survivable.
- Build one honest dashboard. Even a simple, live view of who is current by device and ward changes the conversation with your board from anecdote to evidence, and it turns competency from a thing you hope about into a thing you know.
If you want a structured way through this, our consultancy and training programmes are built around exactly this shift, and the resources library has starting points you can use today.
The test that matters
Come back to the question I opened with. Right now, this minute, who is competent to test on that device? A mature POCT service answers instantly, with evidence, because competency is a live, enforced state and not a stack of paper. An exposed service goes quiet, opens a spreadsheet, and starts counting.
The time bomb is not that operators lapse. Operators will always lapse, because skill decays and people move on, and no amount of enthusiasm changes that. The bomb is that most services cannot see the lapse until it has already produced a result someone acted on. Make competency visible and enforceable, and you defuse it. Leave it in the filing cabinet, and you are simply waiting to find out, the hard way, who was never really competent at all.