Point-of-Care Testing Governance After ISO 22870

Somewhere in most hospitals there is a blood gas analyser nobody in the laboratory chose, a glucose meter bought from a ward budget, and a box of test strips in a drawer that expired in the spring. All three are point-of-care testing, all three produce results that go into patients’ records, and all three are the laboratory’s responsibility.

Since ISO 15189:2022 absorbed the withdrawn ISO 22870, point-of-care testing performed under a laboratory’s oversight is squarely inside that laboratory’s scope — including the devices it did not buy and the operators it does not employ. This guide covers what oversight actually has to include, and the four findings assessors reliably bring back from a ward.

What changed, and why it matters now

ISO 22870:2016 was a separate standard for point-of-care testing, applied alongside ISO 15189. When the fourth edition of ISO 15189 was published in December 2022, ISO 22870 was withdrawn and its requirements were incorporated. ISO’s own abstract for ISO 15189:2022 states that the document is applicable to point-of-care testing.

The practical consequence is that POCT is no longer a separate conversation with a separate standard. It is assessed as part of your medical laboratory accreditation, against the same clauses on competence, equipment, reagents, quality control and reporting that apply inside the laboratory. Any toolkit or management system written before 2022 treats it as an optional extra; it is not one.

The test of whether point-of-care testing oversight is real

Most laboratories monitor point-of-care testing to some degree. Far fewer can stop it, and stopping it is what oversight means.

The question worth asking of your own arrangement is simple: can the laboratory withdraw a device from a ward, or withdraw an operator’s access, and make it stick? If the answer is no, what exists is a reporting function — data flowing to a coordinator who can describe a problem but not end it.

Making that authority real needs three things written down and agreed: who holds it (normally the laboratory director), what triggers its use, and what the clinical service does instead when a device is withdrawn. That last one matters — removing a device without an alternative produces either an unsafe workaround or a different clinical harm.

Deploying a point-of-care testing device properly

Every device should arrive through the same route, including the ones a clinical service has already bought. A workable sequence:

1. State the clinical need — what decision the result supports, and why laboratory testing cannot serve it. Turnaround is a legitimate answer; convenience alone usually is not.
2. Select against defined requirements, including two that get forgotten at purchase and cost dearly afterwards: connectivity to the laboratory system, and the ability to lock out unauthorised or lapsed operators.
3. Verify the method and compare it against the laboratory method on paired patient samples.
4. Assess the location — bench space, temperature and humidity within the device’s specification, power, network, consumable storage, waste disposal, and whether the area’s cleaning regime is compatible with the device.
5. Train and authorise the operators, individually.
6. Define quality control: material, frequency, acceptance criteria, and what happens on failure.
7. Arrange external quality assessment for the POCT method in its own right.
8. Write a procedure for the operator, not for a biomedical scientist.
9. Define the critical value route from that device.
10. Sign an agreement with the clinical service and register the device.

Step 2 is worth insisting on at procurement. A device that cannot enforce operator authorisation transfers that control to human vigilance on a night shift, and it will be the reason for a finding two years later.

Operators: competence at ward scale

A single ward may have a hundred nurses, rotating doctors, night staff, and bank and agency workers, any of whom may need the analyser at three in the morning. None of them work for the laboratory.

A scheme designed like laboratory competence assessment — individual, lengthy, laboratory-delivered — will not scale and will quietly be bypassed. One designed to be undemanding will not produce competence. What works is a short, specific scheme with device-level enforcement behind it:

• The laboratory trains and authorises trainers or link practitioners at each site.
• They train and assess operators against the laboratory’s criteria.
• The laboratory audits a sample of those assessments each year and keeps the right to assess anyone directly.
• Authorisation is recorded per operator per device type and enforced by device lockout.

Two failure modes to design against. The cascade fails silently when a link practitioner leaves, so record who they are per site and check coverage at each review. And never solve rota coverage with a shared login — it destroys operator traceability, which is the only way to identify whose results to review when something goes wrong.

One piece of administration deserves care: tell a site before its operators’ authorisations lapse. A nurse discovering at 3 a.m. that the analyser will not accept them is a patient-safety event created by the way the scheme is run.

What training has to cover

Keep it to what changes the result, and include what the device cannot do:

Comparability: the difference nobody warned the ward about

Point-of-care devices commonly use a different measurement principle from the main laboratory method, and their calibrators may trace differently — or not at all.

So compare them on paired patient samples collected at the same time, not on control material and not on a sample transported to the laboratory hours later for a measurand that moves with time. Repeat on device replacement, on lot change where the manufacturer indicates it, and at a defined interval.

Then tell the clinical service the answer in plain terms: how much difference to expect, in which direction, and at what values it matters. A ward measuring a glucose of 3.9 on a device and receiving 4.6 from the laboratory an hour later will report it as a laboratory error unless somebody has quantified and communicated the expected difference in advance.

Results have to reach the record

A point-of-care result is an examination result. It belongs in the patient’s permanent record, identifiable as such, with the device, the operator, the time and the patient identity attached.

Where a device is not connected, the manual route needs defining and verifying — and it should be recorded as a risk with a plan to close it, not accepted as normal. Hand-transcribed results are the weakest link in the whole pathway, and results that never make it into the record at all are common enough that a monthly comparison of tests performed against results transmitted is worth running.

Reports should identify point-of-care results as such, so that a later reader knows which method produced the number they are comparing.

The four findings assessors bring back from a ward

These are consistent across services, and none of them are visible from inside the laboratory:

1. Expired consumables in use. Strips and cartridges in a drawer past their date. The check has to be scheduled, assigned and recorded — not left to a busy clinical team’s goodwill.
2. Quality control performed but never reviewed. The site runs it faithfully and nobody looks at the results, so a drifting device runs on. The laboratory reviews POCT quality control, not just the site.
3. Lapsed operators still testing, because the device does not enforce authorisation.
4. Results that never reached the record, because the device is not connected and transcription did not happen.

The common thread is that all four are found by going to the site. Monitoring conducted entirely from data in the office misses the device on a windowsill in the sun and the laminated instruction sheet somebody wrote themselves.

Reviewing the programme, including what to remove

Review annually: the sites, the devices, the volumes, the performance, the incidents, and whether each deployment is still clinically justified. Services change, and a device deployed for a pathway that no longer exists is usually still there, still consuming quality control effort and still creating risk.

Removing it is as much a part of oversight as deploying it.

Our guide to ISO 15189 covers how point-of-care testing fits alongside the rest of the standard, and our ISO 15189 Toolkit gives it a section of its own: a governance policy with the deployment route and the withdrawal authority, an operator training and authorisation procedure built for ward-scale cascades, and a device and operator register that tracks quality control compliance, expired consumables, unauthorised use and results reaching the record, site by site.

Frequently asked questions

Is ISO 22870 still current?

No. ISO 22870:2016 was withdrawn when ISO 15189:2022 was published, and its point-of-care requirements were incorporated into that standard. There is no longer a separate POCT standard to be accredited against.

Does ISO 15189 apply to point-of-care testing?

Yes. ISO states that the document is applicable to point-of-care testing, and testing performed under a laboratory’s oversight falls within its management system — whether or not those examinations appear on the accreditation schedule.

Who is responsible for point-of-care testing?

The laboratory director is accountable for testing under the laboratory’s oversight, usually through a POCT coordinator, with the clinical service lead at each site responsible for their own operators and local compliance. The arrangement should be written down and signed by both.

How often should point-of-care devices be compared with the laboratory method?

On paired patient samples at a defined interval — commonly at least annually — and additionally after device replacement and on consumable lot changes where the manufacturer indicates it. The result should be communicated to the clinical service in plain terms.

Do point-of-care examinations need external quality assessment?

Yes, in their own right. The main laboratory’s participation does not cover a ward device using a different method. Where a device is deployed at many sites, a documented rotation across sites is acceptable; leaving sites permanently unassessed is not.

What is the most common point-of-care testing finding?

Expired consumables in use at the site. Close behind it: quality control performed but never reviewed, operators whose authorisation lapsed while the device kept accepting them, and results that never reached the patient’s record.