MDCG 2020-6: Sufficient clinical evidence for legacy devices, explained

Who this applies to

You are affected if you are bringing a legacy device — one CE marked under the old directives (MDD or AIMDD) — through its first MDR conformity assessment. The guidance addresses both manufacturers and notified bodies, and it interprets MDR Article 61(1): confirmation of conformity must be based on clinical data providing sufficient clinical evidence, and you must specify and justify the level of evidence appropriate for your device and its intended purpose.

What "sufficient clinical evidence" means

The MDR never defines the term. The guidance fills the gap: sufficient clinical evidence is "the present result of the qualified assessment which has reached the conclusion that the device is safe and achieves the intended benefits" — a conclusion that has to hold against the current state of the art, and has to be re-confirmed continuously as part of the device lifecycle. Data generated for MDD/AIMDD certification still counts, but it may no longer be enough on its own: the MDR adds consideration of alternative treatment options, a defined equivalence regime (MDCG 2020-5), and mandatory incorporation of post-market surveillance and PMCF data.

The five stages (MDR Annex XIV Part A)

The guidance structures the whole exercise along the clinical evaluation stages of Annex XIV:

1. Establish or update the clinical evaluation plan — including a gap analysis of your directive-era evidence against the MDR's GSPRs.
2. Identify available clinical data — all of it, pre-market and post-market: investigations, literature on your device or a demonstrably equivalent one, PMS and complaint data, PMCF studies, registries.
3. Appraise the data — using validated assessment tools (the guidance names examples such as the Cochrane tool, MINORS, and the Newcastle-Ottawa Scale) to weigh methodological quality, bias, and relevance.
4. Generate new data — where gaps remain. Critically: evidence must be sufficient before CE marking under the MDR. New PMCF studies started under the MDR must not be used to bridge gaps such as unsupported indications.
5. Analyse — determine whether the totality of appraised evidence demonstrates conformity, benefit by benefit and risk by risk, against up-to-date alternatives.

The evidence hierarchy

Appendix III ranks twelve types of evidence from strongest to weakest — roughly:

Two hard rules frame the table: class III and implantable legacy devices that are not well-established technologies need evidence at level 4 or better, and reliance solely on complaints and vigilance is never sufficient. The guidance also warns against complaint-rate arithmetic — incidents divided by sales is not generally acceptable proof of safety.

Well-established technologies

The WET concept (Articles 52(5) and 61(8)) is not limited to the devices listed in Article 61(6b) — it extends to similar devices meeting all four criteria: relatively simple, common, stable designs with little evolution; a generic device group with well-known safety and no past safety issues; well-known clinical performance in a standard-of-care role with little evolution in indications; and a long history on the market. Devices that genuinely qualify may confirm conformity through cumulative evidence from further down the hierarchy, including data from similar devices. Devices that don't must not borrow the label.

What this means for you, practically

1. Start with the gap analysis. Map your directive-era evidence against the MDR GSPRs in the clinical evaluation plan (Appendix II of the guidance suggests a minimum structure for legacy devices).
2. Appraise before you argue. Grade every data source with a recognised tool and document why it does or doesn't count — this is exactly what the notified body's CEAR reviewer (MDCG 2020-13) will probe.
3. Close gaps before certification, not after. If the evidence doesn't support a claim, the guidance says to narrow the intended purpose until it does. PMCF confirms conclusions; it doesn't rescue unsupported indications.
4. Use Article 61(10) carefully. The non-clinical route exists for cases where clinical data is "not deemed appropriate" — but it is never available for class III or implantable devices, and it demands justification from your risk management.