Under the EU In Vitro Diagnostic Medical Devices Regulation (IVDR, Regulation (EU) 2017/746), the performance evaluation is the basis for demonstrating that an IVD achieves its intended purpose safely and reliably. It rests on three pillars – scientific validity, analytical performance and clinical performance (Article 56 and Annex XIII) – and its results are documented in the performance evaluation report.
The framework below helps manufacturers structure this report, from the product overview and risk analysis to the individual elements of the evaluation. It draws on EN 13612:2002 (performance evaluation of IVDs) and the IMDRF/GHTF guidance documents GHTF/SG5/N6:2012, GHTF/SG5/N7:2012 and GHTF/SG5/N8:2012. For clinical performance studies, EN ISO 20916 should also be considered. The example used throughout is an HIV test.
The 12 sections of the report
- Abbreviations – e.g. “IVD MD: in vitro diagnostic medical device”.
- Product – name, version, Basic UDI-DI (where available), UMDNS code, GMDN code.
- Related documents – the performance evaluation SOP and the performance evaluation plan.
- Intended purpose – description of the device’s intended purpose.
- Risk analysis – summary of the risk analysis report.
- Medical background and state of the art
- 6.1 Medical background: in which clinical settings the IVD is used – an HIV test, for example, may be used for screening.
- 6.2 State of the art: how this is done today – e.g. what currently happens to patients undergoing HIV screening.
- Scientific validity – usually based on literature research: the association of the analyte with a clinical condition. Example: an HIV test is scientifically valid because HIV is known to lead to AIDS and early detection improves treatment outcomes.
- 7.1 Literature search method: keywords and sources, e.g. bibliographic databases (MEDLINE, EMBASE), specialised databases (MEDION), systematic review databases, clinical trial registers (CENTRAL, NIH) and adverse event databases (MAUDE, IRIS) – see the examples in GHTF/SG5/N7:2012.
- 7.2 Search results: database terms, relevant publications with title, authors and abstract.
- 7.3 Conclusion: e.g. “detecting HIV is useful because HIV subsequently leads to AIDS and early detection is beneficial”.
- Analytical performance – the metrics the test should achieve, such as sensitivity and specificity: what performance can be expected when testing, say, 100 blood samples from different patients? This requires testing on a defined test set and analysing the results.
- 8.1 Method: describe the approach; for a machine-learning model, also the test set, why it was chosen and the metrics used.
- 8.2 Results.
- Clinical performance – similar to scientific validity, but about the device’s performance in its target patient population. Analytical data only reflect the assay level; in real patients, sensitivity and specificity may differ, for example because comorbidities such as other viral diseases cause false positives. Clinical performance can be demonstrated by literature research or clinical performance studies.
- 9.1 Method.
- 9.2 Results.
- Conclusion – summarise why the IVD is safe and performs as intended, referring to the intended purpose, the risks in the risk analysis, and the scientific validity, analytical and clinical performance.
- Date and signature – electronic signature if your document management system supports it, otherwise a handwritten signature.
- Qualification of the evaluators – attach the CVs of the people who prepared the performance evaluation; they must have adequate skills and training.
Mapping EN 13612 to the documentation
| EN 13612 clause | Documented in |
|---|---|
| 3.1 Responsibilities and resources | Performance evaluation plan |
| 3.2 Documentation | Performance evaluation plan |
| 3.3 Final assessment and review | Performance evaluation report |
| 4.1 Prerequisites | Performance evaluation report |
| 4.2 Evaluation protocol | Performance evaluation plan |
| 4.3 Site and resources | Performance evaluation plan |
| 4.4 Basic design information | Performance evaluation plan |
| 4.5 Experimental design | Performance evaluation plan |
| 4.6 Records of the performance study | Performance evaluation plan |
| 4.7 Observations and unexpected results | Performance evaluation plan |
| 4.8 Evaluation report | Performance evaluation report |
| 5 Modifications during the performance study | Performance evaluation plan |
How EVERBIZ can help
EVERBIZ supports IVD manufacturers in preparing performance evaluation plans and reports, literature searches and performance studies under the IVDR. Contact us.
This article is for general information only and does not constitute legal advice.