Clinical performance is based on the intended use defined by the manufacturer and needs to be done at the user’s end, in a clinical laboratory or in a hospital environment with use conditions, on the target population, and by the intended user. The clinical performance plan defines the methods involved.
For some devices the following methods may be sufficient:
- Published literature
- Experience gained from routine testing
But for novel devices/markers:
- Comparison with a clinical standard practice or method comparator is necessary by conducting a clinical performance study to establish the trueness of the value or the diagnostic accuracy
Clinical Performance studies define the parameters such as diagnostic sensitivity and specificity, positive and negative predictive values and likelihood ratios, expected values in normal and affected population, and also the sample collection and handling requirements. For new devices, such studies can confirm performance that cannot be tested by analytical performance studies. They are similar to clinical trials and therefore require ethics committee approval. The manufacturer needs approvals from the Competent Authority of the member states before conducting such studies. The studies are planned in such a manner as to protect the patients in every possible way. Depending on the type of product, different types of studies may be required, such as non-interventional, interventional, or those involving special cases like minors, pregnant women etc. (refer to Articles 57-58 and Annex XIII). If a manufacturer thinks that clinical studies are not required, proper justification needs to be provided.
The Scientific Validity Report, Analytical Performance Report, and Clinical Performance Report are collated in the Performance Evaluation Report. They are a very critical part of the technical documentation. The PER is updated throughout the life cycle of the device, with new and factual data gathered from the market with the help of a Post Market Performance Follow-up plan, or PMPF plan. It is the clinical evidence data that actually helps define how PMPF should be done. The results are documented in a PMPF report and updated in the PER.
An important point to note is that the evaluation of the device may help understand new and emerging risks. Therefore, a system is required to link the performance evaluation and PMPF to the Risk Management process.
For Classes A and B, the low-risk devices, the PER is updated as necessary, but at least once every three years is recommended. It also must be made available to the notified body as and when required. For Class C and Class D, the PER should be updated when necessary, but at least annually.