Pharmacy First and the POCT Quality Gap

Walk into a community pharmacy on a Tuesday morning and you are standing in the most accessible clinical setting in the country. No appointment, no referral, no car park. A pharmacist who has trained for years, who knows the local families by name, who catches the things a rushed ten-minute GP slot misses. For a decade we have said we should use this setting better, and at last we are. Since the launch of NHS Pharmacy First in England on 31 January 2024, community pharmacists have been able to complete whole episodes of care for seven common conditions. In its first fifteen months, to April 2025, community pharmacists delivered 6.4 million consultations, close to five million of them inside the first year alone. The direction of travel is right, and I want to say that plainly before I say anything else.

But something else is happening in that same room, quietly, alongside the headline scheme. A finger-prick sample is going onto a strip. A cartridge is going into a small analyser on the dispensary bench. A screening service is booked in for the afternoon. Commissioned and private point-of-care testing in community pharmacy, for cholesterol, for HbA1c, for C-reactive protein, for a widening list of screens, is growing fast, and a result off one of those devices is a laboratory result whether or not anyone in the building thinks of it that way. It carries the same core obligations any laboratory result carries. And that is where the trouble starts.

Here is my thesis, and I hold it as a friend of community pharmacy rather than a critic. We have moved diagnostic testing into a setting that was never built around a laboratory quality system, and we have not yet moved the quality system with it. The clinical governance that protects a result in a hospital laboratory, quality control with acceptance rules, competency assessment, external quality assessment, disciplined adherence to the instructions for use, and a guaranteed route to escalate an abnormal result, has not scaled with the ambition. This gap is nobody's fault in particular, and it is fixable. Closing it is not a burden imposed on pharmacy. It is the price of doing testing properly.

What Pharmacy First actually changed, and what it did not

It is worth being precise, because the accuracy matters and because a sloppy version of this argument does pharmacy a disservice. Pharmacy First did not turn pharmacies into laboratories. The seven clinical pathways lean on structured history, examination and validated clinical scoring tools. The sore throat pathway, for instance, uses a validated clinical scoring tool, the FeverPAIN score, to decide who might benefit from antibiotics, not a mandated point-of-care blood test. There is no requirement in the scheme to run C-reactive protein or any other bench assay to complete those consultations.

What Pharmacy First did change is arguably more profound. It normalised the community pharmacy as a place where real clinical decisions are made and episodes of care are owned end to end. It moved the centre of gravity. Once the public and the profession accept the pharmacy as a place you go to be assessed and treated, the addition of a test on the bench feels like a small, natural next step. And commissioners and private providers have noticed. Point-of-care and screening services are being layered on precisely because the clinical credibility is now there to carry them.

Pharmacy First did not put a lab on the dispensary bench. It made the bench a place where putting a lab there feels natural. That is exactly why the quality question cannot wait.

So the testing growth is not the scheme. It runs alongside it, commissioned locally, sold privately, encouraged nationally. NHS England has committed to explore and support point-of-care testing in community pharmacy, and guidance for POCT in this setting exists. The ambition is explicit. What has not kept pace is the scaffolding underneath it.

The signs are already concrete, not hypothetical. A commissioned cholesterol-testing pilot in north east London, a seven-minute finger-prick test run with Barts Health and Heart UK, grew from a single pharmacy in November 2024 to seventy across east London inside a year. In the subset that was formally evaluated, more than 550 people were tested; around one in five carried a ten-year cardiovascular risk of 10 percent or more, fourteen had cholesterol high enough to act on there and then, and two dozen were started on a statin who might otherwise have waited years for the conversation. HbA1c, the analyte used to diagnose and monitor diabetes, is following the same route onto pharmacy benches, with point-of-care pathways now being built and validated specifically for this setting. Cholesterol, HbA1c and C-reactive protein are laboratory analytes with laboratory reference ranges and laboratory failure modes. They are now being measured on dispensary benches, at pace, and every published pilot points to more of it, not less.

What a laboratory quietly does that a pharmacy has never had to

To see the gap clearly, look at what happens to a result inside an accredited laboratory before anyone is allowed to trust it. None of it is glamorous, and every step earns its place.

Every analyte in scope runs internal quality control at a defined frequency, and those control results are checked against acceptance rules before patient results are released. A control that drifts out stops the run. Every operator is trained against the instructions for use, formally assessed as competent, and reassessed periodically, because a device is only as good as the hands that run it and the sample that goes into it. The service enrols in external quality assessment, sending blind samples to a scheme such as UK NEQAS or WEQAS, so that its performance is compared against peers and its blind spots are exposed by someone other than itself. Procedures are written, controlled and followed. And when a result is abnormal or simply does not fit the patient in front of you, there is a defined, non-optional route to escalate it.

A community pharmacy has extraordinary clinical strengths. It does not, historically, have any of that machinery, because it never needed it. Nobody built a pharmacy around a quality management system, because until recently a pharmacy did not generate its own diagnostic numbers. The competencies, the accountability, the counselling skill are all superb. The laboratory-style obligations that arrive the moment a device starts producing results are new, and they are not optional extras that a diligent pharmacy can add later at leisure. They are the definition of a trustworthy result.

Why the standards now say this out loud

This is not merely my opinion; the standards world has already moved. For years, point-of-care testing sat awkwardly under a separate standard, ISO 22870, which is now withdrawn. The current international standard for medical laboratories, ISO 15189:2022, brings point-of-care testing inside the laboratory quality standard itself. The message is unambiguous. Testing is testing. The place it happens does not change the principles a good result requires, even if the depth of the system is scaled to the risk and the setting. Competency, quality control and external quality assessment apply wherever the test is done, whether that is a pathology laboratory, an emergency department, a care home or a pharmacy dispensary.

What that does, in effect, is remove the excuse that a pharmacy is somehow a different category where the rules are lighter. It is not. A CRP result generated on a bench in a high street pharmacy and a CRP result generated in a hospital core laboratory are used to make the same kind of decision, even where the method, the sample matrix and the pathway around them differ. If we would not accept the hospital one without QC, competency and EQA behind it, we cannot rationally accept the pharmacy one without them either.

The same clinical object, though, is not always the same number. When the NHS Diabetes Prevention Programme compared more than seventy thousand point-of-care HbA1c results against the laboratory, the bench devices read on average about 2.5 mmol/mol lower, and one model closer to 3. That is not a broken analyser. It is a known, correctable difference between methods. The trouble is that you cannot see it from inside the pharmacy. A systematic offset like that only reveals itself when your results are held up against a reference, or against your peers, which is precisely what external quality assessment is for and precisely what a lone device on a dispensary bench will never do for you.

Where things actually go wrong, and it is rarely the machine

When point-of-care testing produces a wrong or misleading answer, the analyser is seldom the culprit. Modern devices are good. The failure modes cluster upstream and downstream of the box, in exactly the places a laboratory quality system is designed to police.

The evidence is unusually clear on this. When researchers audited more than a hundred thousand point-of-care tests across three devices in UK acute settings, the chemistry was rarely the issue, yet the recorded error rates were far from trivial, and most of what went wrong was traced to the operator and the sample rather than the box. On the blood-gas style analysers the total device-recorded error rate sat between roughly 7 and 9 percent. On the C-reactive protein analyser, the single device most relevant to a Pharmacy First style conversation about whether to reach for antibiotics, almost one run in three tripped an error, and the single largest cause, thirty-five of the fifty probable operator errors, was simply too little sample going in. None of that is the assay failing. All of it is the pre-analytical discipline a laboratory takes for granted and a pharmacy has never been handed.

• Pre-analytical error. A finger squeezed too hard, a sample taken too soon after a meal, a specimen at the wrong temperature, an expired strip left out of the fridge. Sample technique is where a startling share of point-of-care error lives, and it is invisible in the number the device reports.
• Skipped or unread quality control. A control that is run but not checked against a rule, or not run at all because the clinic was busy, means the device could have been drifting all week and nobody would know.
• Instructions for use treated as a suggestion. Wrong sample type, wrong volume, storage outside the stated range, a lot not verified before use. The instructions are the validated conditions under which the result is meaningful. Outside them, the number is decorative.
• No escalation route. A frankly abnormal result with no defined, guaranteed path to a clinician who can act on it is not a safety net with a hole in it. It is no net at all.

Notice that not one of these is a criticism of pharmacists' clinical ability. They are systems failures, and systems are exactly what a quality management system supplies. The absence is structural, not personal, which is precisely why the answer is structural too.

What "good" looks like in a pharmacy doing POCT

So what would it take to do this properly? Less than people fear, and more than a vendor demonstration. The honest way to picture it is as a ladder. At the bottom sits testing that happens on the strength of a vendor demo and a confident operator. At the top sits a full quality management system of the kind an accredited hospital service runs. On the evidence of the pilots and evaluations published so far, much community testing today sits nearer the bottom than the top, and the job is to climb.

The rungs are not mysterious, and none of them require a pharmacy to become a hospital. In order, "good" means:

1. Trained, competency-assessed operators. Every person who runs the test is trained against the instructions for use and formally signed off as competent, with periodic reassessment. Sample technique is taught, watched and confirmed, not assumed.
2. Internal quality control with acceptance rules. Controls run at a defined frequency, results recorded, and clear rules for what happens when a control is out. Crucially, a rule that stops patient testing until the problem is resolved.
3. EQA enrolment. The service participates in an external quality assessment scheme for the analytes it offers, so its performance is checked against peers by an independent party, and poor performance triggers investigation.
4. Controlled procedures. Written standard operating procedures for each test, version-controlled, accessible at the bench, and actually followed. Lot verification and storage discipline built in.
5. A guaranteed escalation route. A defined, documented path for an abnormal or unexpected result to reach a clinician who can act, with clear responsibility for who owns that decision. Not a hope. A pathway.

That is the whole of it. It is neither exotic nor expensive relative to the harm a single confidently wrong result can do. It is the ordinary, unglamorous discipline that has protected laboratory medicine for decades, applied to a new and welcome setting.

We owe community pharmacy the same quality scaffolding the laboratory has always had. Not because pharmacists cannot be trusted, but because results cannot be trusted without it.

What this means for your service

If you commission, run or advise on testing in community pharmacy, this is not a reason to stop. It is a reason to build the scaffolding deliberately, before volume makes retrofitting it painful. Practical steps, in the order I would take them:

• List every test you actually offer. Commissioned and private, on every device on the premises. You cannot govern what you have not counted. Include the screens people forget are diagnostic.
• Put a competency framework behind every operator. Initial training against the instructions for use, an assessed sign-off, and a reassessment date. Make sample technique an explicit, observed part of it.
• Set up internal QC with written acceptance rules. Decide the frequency, record the results, and above all define what happens when a control fails, including the rule that testing stops until it is fixed.
• Enrol in an appropriate EQA scheme for each analyte. Treat a poor EQA return as a signal to investigate, not a form to file. It is one of the cheapest ways to find out you have a problem before a patient does.
• Write and control your procedures. One SOP per test, kept current, kept at the bench. Cover lot verification, storage, and the exact steps of the assay.
• Guarantee the escalation route in writing. Name who an abnormal result goes to, how, and within what timeframe. Rehearse it. A route that exists only in principle does not exist.

If you are starting from a standing start, begin with the free POCT Fundamentals course, which covers exactly this ground: quality control, competency, sample technique and escalation, in plain language for a non-laboratory setting. Our wider training goes deeper for teams taking on more analytes. When you are ready to formalise it, our templates in the resource library give you SOP, competency and QC log starting points you can adapt, and our consultancy helps commissioners and pharmacy groups stand up a proportionate quality system sized to the service rather than to a hospital.

The setting was always right. Now make the result trustworthy.

Moving testing closer to the patient is one of the better ideas in UK healthcare, and community pharmacy is the right place to do it. I would not slow that down for anything. But a result you cannot trust can be worse than no result, because it is acted upon. The pharmacy earned the clinical role. Now we owe the pharmacy the quality system that makes its numbers safe. That is not extra bureaucracy. It is the other half of the same job.

Sources and notes

This article draws on published NHS statistics and on peer-reviewed and professional sources. Consultation totals are NHS Business Services Authority data, reported through NHS England and The Pharmaceutical Journal. The device error rates in Figure 3 are from a UK acute-setting evaluation and are the best published proxy for the pre-analytical failure modes discussed here; the HbA1c comparison in Figure 2 is drawn from more than seventy thousand paired point-of-care and laboratory results in the NHS Diabetes Prevention Programme. The activity curve in Figure 1 plots published NHSBSA consultation milestones with the connecting line interpolated, and its governance-maturity line is qualitative, illustrating an argument rather than a measured dataset.

1. NHS Business Services Authority. NHS Pharmacy First Service clinical pathways activity data. 9,976 pharmacies delivered over 423,000 clinical pathway consultations in the first quarter alone.
2. The Pharmaceutical Journal. Pharmacy First passes 5 million consultations landmark. 5.4 million consultations to the end of February 2025, of which 2.4 million followed the seven clinical pathways.
3. NHS England. Point of care testing in community pharmacy. Reports the 6.4 million Pharmacy First consultations delivered between January 2024 and April 2025.
4. Barts Health NHS Trust. Free cholesterol tests expanded to 70 east London pharmacies. Grew from one pharmacy in November 2024, with around one in five people at raised cardiovascular risk.
5. Community Pharmacy England. Service case study: cholesterol point-of-care testing in community pharmacy. Formal evaluation subset: 556 people, 111 with a QRISK score of 10 percent or more, 24 started on a statin.
6. The Pharmacist. Cholesterol testing in North East London pharmacies.
7. Barron and colleagues. Experience of point-of-care HbA1c testing in the English NHS Diabetes Prevention Programme. BMJ Open Diabetes Research and Care, 2020 (n=73,703 paired results; point-of-care read 2.48 mmol/mol lower than the laboratory on average).
8. Radley and colleagues. Implementation of point-of-care HbA1c instruments into community pharmacies. Annals of Clinical Biochemistry, 2024, on the quality control, EQA and competency needed for robust community testing.
9. Fanshawe and colleagues. Pre-analytical error for three point-of-care venous blood testing platforms in acute ambulatory settings. PLOS One, 2020 (i-STAT n=113,266, 6.8 percent error; Alinity n=1,829, 9.2 percent; Afinion CRP n=556, 32 percent, of which insufficient sample was the largest cause).
10. UK NEQAS. Point of care testing (POCT) external quality assessment.
11. International Organization for Standardization. ISO 15189:2022, Medical laboratories, requirements for quality and competence, which now brings point-of-care testing inside the laboratory quality standard, superseding the withdrawn ISO 22870.