ISO 13485 · Medium · 12 min lesson
Performance qualification in depth: sample size, acceptance criteria and the validated state
ISO 13485:2016 clause 7.5.6 and GHTF SG3 N99-10 sections 5.5, 5.6 and 6
What you will learn
- Design a performance qualification under routine conditions and justify the number of consecutive runs.
- Calculate sample sizes from confidence and reliability, for attribute and variable data, and read a Cpk.
- Review a validation report and keep the validated state through change control, monitoring and revalidation.
Key takeaways
- PQ shows that routine production, with real operators, lots and shifts, consistently meets requirements (GHTF 2.3 and 5.5).
- GHTF sets no number of runs; three consecutive runs is a convention the protocol must justify, and every run counts.
- Sample size follows from reliability and confidence: 29 zero-failure pouches for 90 and 95 percent, 59 for 95 and 95.
- Acceptance criteria must be able to fail, and a failed run is investigated and reported, never deleted.
- The validated state is maintained by change control, monitoring and revalidation criteria (clause 7.5.6 f and g; GHTF section 6).
Free practice question
Lumora protocol PQ-022 rev A, section 6 (fictional): 'Three runs will be performed on 12, 13 and 14 March by senior technician A, film lot F-2201, pouch size S, directly after the annual service of the sealer. Dye test n=14 per run, zero failures allowed.'
What is the main weakness of this PQ design?
The runs cover only a best-case day: one lot, one operator, one size and a freshly serviced machine.
Correct. GHTF section 5.5 asks for challenges that simulate real manufacturing and the range of action levels, so clause 7.5.6 consistency is unproven here.
Three runs are too few because GHTF N99-10 requires at least ten.
Not quite. N99-10 section 5.5 sets no run count; it asks for enough repeats to be meaningful, so quoting a fixed minimum of ten is wrong.
Zero failures allowed is forbidden in a performance qualification.
Not quite. GHTF Annex A describes acceptance sampling with confidence statements, so a zero-failure plan is acceptable if the sample size supports the claim.
Dye penetration cannot be used to test seals.
Not quite. GHTF section 5.2 encourages standard test methods that replicate use, and dye penetration is a recognised seal integrity method, so the method is not the flaw.
Key terms
- performance qualification: Objective evidence that the process, under anticipated conditions, consistently produces product meeting all predetermined requirements (GHTF 2.3).
- operational qualification: Objective evidence of the process control limits and action levels that give conforming product (GHTF 2.2).
- reliability: The proportion of future product that conforms to requirements, in a confidence statement.
- confidence: How sure we are that the stated reliability is true, for example 95 percent.
- attribute data: Pass or fail results such as dye penetration; they need larger samples than measured data.
- variable data: Measured numerical results such as peel strength in g/in, allowing tolerance bounds and Cpk.
- Cpk: Capability index that compares the nearer specification limit with three standard deviations; it shows spread and centering together.
- action level: A value inside the specification at which the process is adjusted before it reaches worst-case conditions (GHTF 5.4).
Official sources
- GHTF/SG3/N99-10:2004 (Edition 2) Quality management systems - Process validation guidance (sections 2, 5.2, 5.5, 5.6, 6, Annex A and B)
- ISO 13485:2016 Medical devices - Quality management systems (clauses 7.5.6, 7.5.7, 7.6, 8.4, 8.5.2)
Go further
The full lesson has narrated slides, more scenario questions with explained answers and a simulation.
Open the AcademyWatch on YouTubeIllustrative example. Educational only, not legal advice.