4.3 Technical Performance / Analytical Performance (Pillar 2)
Source: MDCG 2020-1 (March 2020), pages 12-13
4.3 Technical Performance / Analytical Performance
The manufacturer should verify that the MDSW reliably, accurately and consistently meets the intended purpose in real-world usage.
The relevant performance characteristics, as part of the GSPRs and linked to the analytical and / or clinical features, should be supported by evidence generated during verification and validation activities as part of good manufacturing practices for software, or by generating new evidence through the use of curated databases, curated registries, reference databases or use of previously collected patient data.
TECHNICAL PERFORMANCE / ANALYTICAL PERFORMANCE is confirmed by the examination and provision of objective evidence that the MDSW specifications conform to user needs and intended uses, and that the requirements implemented can be consistently fulfilled.1
For example, performance verification and validation in the intended computing2 and use environments34 can be characterised by the demonstration of
- availability,
- confidentiality,
- integrity,
- reliability,
- accuracy (resulting from trueness and precision),
- analytical sensitivity,
- limit of detection,
- limit of quantitation,
- analytical specificity,
- linearity,
- cut-off value(s),
- measuring interval (range),
- GENERALISABILITY,
- expected data rate or quality,
- absence of inacceptable cybersecurity vulnerabilities,
- HUMAN FACTORS ENGINEERING.
Identification of gaps during the validation of the TECHNICAL PERFORMANCE / ANALYTICAL PERFORMANCE could require generation of new evidence, for example, to demonstrate generalisability with real-life datasets or to extend the usability evaluation to omitted user groups.