How to Manage a Laboratory Proficiency Testing Programme
Use external comparison to find bias your own controls cannot see.
What a how-to is not
A how-to is SPEQ’s practitioner method, not a procedure. It does not replace your own SOP, it is not a validated approach, and the judgement calls in it belong to your quality unit.
Internal quality control detects drift against your own values. It cannot detect bias, because a laboratory that is consistently wrong is consistently wrong against its own controls too. Proficiency testing and other interlaboratory comparison is the only routine mechanism that measures a laboratory against the outside world, which is why it is required rather than recommended and why the results deserve more attention than they usually get.
- 1
Enrol for every test and specialty that requires it
Coverage is per specialty and per test, and gaps are found by mapping your test menu against your enrolments rather than by assuming the provider covers what you do. Where no scheme exists for an analyte, arrange an alternative comparison and document the rationale.
- 2
Handle the samples exactly as patient or study samples
Same staff, same methods, same instruments, same number of replicates, integrated into the routine run. A sample given special handling measures your best case, which is not the thing being assessed and not what your patients receive.
- 3
Never refer or discuss before the deadline
Referring a sample to another laboratory or comparing results with one before results are due defeats the measurement and carries severe consequences in some regimes. The rule is absolute and the intent behind breaking it does not mitigate.
- 4
Investigate every unacceptable result to a cause
Treat an outlier as a nonconformity: cause, impact on patient or study results already reported, correction and prevention. The impact assessment on results already issued is the step most often skipped and the one that matters most.
- 5
Look at the trend, not only the pass
A run of results consistently at one end of the acceptable range is bias that has not yet failed. Waiting for a failure means acting after the point at which the data were already skewed, which is precisely what the programme exists to prevent.
- 6
Feed the outcomes into review and competence
Performance belongs in management review and in individual competence assessment. A programme whose results are filed and never used is a cost with no control attached, which is the most common way this requirement is satisfied on paper.
- !Enrolment gaps found only at assessment, because the test menu was never mapped against the schemes.
- !Samples given special handling, so the programme measures the best case rather than routine performance.
- !An unacceptable result closed without assessing the patient or study results already reported.
- !Consistent bias inside the acceptable range treated as a pass rather than as a trend to act on.
How to Manage a Laboratory Proficiency Testing Programme: frequently asked questions
Common questions on manage a laboratory proficiency testing programme.
Why is internal quality control not enough?
Because it compares a laboratory against itself. A systematic bias — a miscalibrated instrument, a shifted calibrator lot — moves the controls and the patient results together, so the control passes. External comparison is the only routine mechanism that measures you against everyone else, which is where bias becomes visible.
What if no scheme exists for an analyte?
Arrange an alternative comparison: exchange samples with another laboratory, analyse a reference material, or split samples with a laboratory using a different method. Document the approach and the rationale. Absence of a commercial scheme is not an exemption from the underlying obligation.
What should happen after an unacceptable result?
A full nonconformity investigation, including whether results already reported to clinicians or sponsors were affected and whether any need correcting. The retrospective impact assessment is the step most often skipped, and it is the one that determines whether anyone was actually harmed by the error.