The Questions That Rewrote Clinical Evaluation Forever

Written by HATEM RABEH, MD, MSc Ing

Your Clinical Evaluation Expert And Partner

in
S

This is what happens when regulators realize the guidance wasn’t clear enough. MDCG 2021-6 didn’t introduce new requirements. It revealed how many of us had been misreading the existing ones.

I’ve seen files built on interpretations that seemed reasonable at the time. Solid work. Defensible positions. Then this Q&A document clarified what was actually expected. And those positions collapsed.

Why This Document Exists

MDCG 2020-13 laid out the framework for clinical evaluation. It was comprehensive. But implementation exposed hundreds of questions.

What level of technical equivalence is sufficient? How much clinical data is enough for an established technology? When does the State of the Art actually require new studies?

Manufacturers asked these questions. Notified Bodies answered differently. The resulting inconsistency forced regulators to clarify. MDCG 2021-6 is that clarification.

It’s structured as questions and answers. But don’t mistake the format for simplicity. Each answer redefines boundaries that were previously unclear.

The Equivalence Earthquake

If there’s one section that forced rewrites across the industry, it’s the clarification on equivalence.

The MDR allows you to rely on clinical data from an equivalent device. Most manufacturers understood this as: if your device is similar enough, you can reference someone else’s data.

MDCG 2021-6 sharpened that interpretation dramatically.

This shift matters because it changes what you need to demonstrate. Previously, manufacturers focused on technical comparison tables. Similar materials. Similar design. Similar intended use.

The Q&A clarifies that technical similarity is necessary but not sufficient. You must show that the clinical and biological interaction with the body will be the same.

This is why surface coatings matter. Why contact duration matters. Why subtle design differences that seem trivial from an engineering perspective become critical from a clinical evaluation perspective.

What This Means for Your Equivalence File

First, the technical comparison must go deeper. It’s not enough to list specifications side by side. You must explain why each difference is clinically irrelevant.

Second, the biological and clinical sections must explicitly address transferability. Why will the data from the equivalent device predict the performance of yours?

Third, any difference that could affect safety or performance requires additional data from your device. This is where most equivalence claims break down under review.

I’ve seen claims rejected because a minor difference in sterilization method wasn’t adequately justified. The manufacturer argued it was a validated process. The reviewer asked: does this process affect the biological response? The file had no answer.

The State of the Art Reckoning

Article 61(1) requires that clinical evaluation considers the current knowledge, the State of the Art. Everyone knew this. Few knew what it actually demanded.

MDCG 2021-6 clarifies that SOTA isn’t just a literature review section. It’s the benchmark against which your device is measured.

The questions in the document make this explicit. What does SOTA include? Clinical outcomes. Safety profiles. Alternative treatments. Benefit-risk considerations from existing options.

Here’s the consequence: your clinical evaluation must demonstrate that your device meets or exceeds what’s currently available. Not in marketing terms. In clinical evidence terms.

If the SOTA for a surgical mesh shows 5% recurrence at five years, your data must address this. If you have shorter follow-up, you must justify why your data is still sufficient. If your outcomes appear worse, you must explain why the benefit-risk remains favorable.

What Changes in Practice

Your literature review must now explicitly identify the SOTA benchmarks. This means going beyond device-specific studies. You need clinical guidelines. You need systematic reviews. You need registry data.

Then your appraisal must compare your device against these benchmarks. Not implicitly. Explicitly.

I’ve reviewed files where the SOTA section was thorough but disconnected. Beautiful literature tables. But no analysis of what the literature establishes as current performance expectations. No comparison of the device data against these expectations.

The Q&A makes clear this isn’t acceptable. SOTA must inform your evaluation. It must shape your conclusions.

How Much Data Is Enough?

This question appears repeatedly in MDCG 2021-6. For good reason. It’s the question every manufacturer asks. It’s the question where Notified Bodies varied most widely.

The answers don’t give numbers. They give logic.

The sufficiency of clinical data depends on the device risk class, the novelty of the technology, the available evidence base, and the specific claims you’re making.

For well-established technologies where extensive literature exists, you can rely more heavily on that literature. But you still need device-specific data for verification.

For novel devices or novel applications, you need substantial clinical data from your device. How substantial? Enough to demonstrate the safety and performance claims you’re making.

This is circular reasoning until you understand the underlying principle. The data must be proportionate to the uncertainty.

Proportionality in Practice

If you’re launching a new orthopedic implant with a novel surface coating, the uncertainty is high. You need clinical data showing the coating performs as expected in humans. Literature on similar coatings helps but doesn’t replace your data.

If you’re modifying an established implant by changing the manufacturer of a well-characterized material, the uncertainty is lower. You might demonstrate equivalence with strong technical and biological characterization plus limited clinical verification.

The Q&A emphasizes that post-market data can contribute to demonstrating sufficient clinical evidence. But only if collected systematically through PMCF. And only if the pre-market data was already sufficient for the initial risk-benefit determination.

The PMCF Plan Requirements

Several questions in MDCG 2021-6 address PMCF planning. The clarifications here are specific and demanding.

Your PMCF plan must define methods. Not just state that you’ll conduct PMCF. What will you collect? How? From whom? When?

It must identify gaps and uncertainties from the pre-market evaluation. These become your PMCF objectives. If you claim there are no gaps, reviewers will question whether your pre-market evaluation was thorough enough.

The plan must also explain how PMCF data will be analyzed and fed back into the clinical evaluation. This closes the loop. Data collection without planned analysis is meaningless.

I see plans that treat PMCF as a compliance checkbox. A paragraph stating that PMCF will be conducted according to MDCG 2020-7. This isn’t a plan. It’s an acknowledgment that a plan is required.

What a Real PMCF Plan Contains

Specific data points you’ll collect. Why these data points address your gaps. What thresholds or trends would trigger action.

Defined follow-up intervals. Realistic patient numbers based on your market projections. Methods for capturing data from clinical practice, not just controlled studies.

A timeline for analysis and CER updates. PMCF isn’t a perpetual data collection exercise. It’s evidence generation that feeds regulatory decision-making.

The Q&A makes clear that PMCF plans will be assessed as part of your clinical evaluation. Weak plans signal weak clinical strategy. Reviewers notice.

Literature Search Depth

MDCG 2021-6 addresses how comprehensive literature searches must be. The answer: comprehensive enough to capture the relevant evidence.

This sounds obvious until you realize how much variation exists in practice. Some manufacturers search only device-specific terms. Others cast wide nets that generate thousands of irrelevant results.

The Q&A clarifies that your search strategy must cover your device, equivalent devices, similar technologies, the clinical condition, and alternative treatments.

You must search multiple databases. You must update regularly. You must document your strategy and justify any limitations.

Here’s what matters: the search must be reproducible. A reviewer should be able to follow your methodology and reach the same literature set.

When Updates Are Required

Several questions address CER update frequency. The regulation says at least annually for implantables and class III devices. But MDCG 2021-6 clarifies this isn’t mechanical.

Updates must occur when new information becomes available that could affect your benefit-risk determination. A major safety signal from a similar device. New clinical guidelines. Significant PMCF findings.

Annual updates for high-risk devices ensure regular review. But they don’t excuse you from updating when circumstances demand it.

I’ve seen manufacturers wait for their scheduled update while sitting on PMCF data showing unexpected performance patterns. This isn’t compliance. It’s risk.

The Q&A makes explicit that updates must be substantive. A paragraph stating “no new information” isn’t an update. It’s evidence you didn’t look hard enough.

The Qualification of Evaluators

One question addresses who can write clinical evaluations. The answer reinforces what the MDR already states: qualified clinical evaluators with appropriate expertise.

But the Q&A adds nuance. The evaluator must have clinical expertise relevant to the device and its clinical application. A general practitioner can’t evaluate complex cardiovascular implants. A surgeon experienced in one anatomical area can’t automatically evaluate devices for another.

This matters because I see CVs attached to CERs where the clinical connection is tenuous. The evaluator has medical credentials but no apparent expertise in the specific clinical domain.

Notified Bodies now routinely challenge evaluator qualifications. If you can’t demonstrate that your evaluator understands the clinical context deeply, your entire evaluation is questioned.

What This Means for Your Current Files

If your clinical evaluation was completed before MDCG 2021-6, the clarifications don’t invalidate your work. But they do set the current interpretation standard.

When you update your CER, apply these clarifications. When Notified Bodies review your file, they’ll apply these clarifications.

This means equivalence demonstrations need deeper justification. SOTA analysis needs explicit benchmarking. PMCF plans need concrete methodology. Literature searches need documented comprehensiveness.

The questions in MDCG 2021-6 reflect what reviewers were already asking. Now those questions have official answers. The informal variation in how different Notified Bodies interpreted the requirements is narrowing.

This is good for consistency. It’s challenging for manufacturers who built files under previous interpretations.

Looking Forward

MDCG 2021-6 won’t be the last clarification document. As implementation continues, more questions will emerge. More guidance will follow.

But this document represents a inflection point. It shows how regulators are tightening interpretation. How the initial flexibility is giving way to more defined expectations.

The manufacturers adapting fastest are those treating these clarifications as the new baseline. Not as additional burdens to minimize, but as the standard to meet.

Your clinical evaluation strategy should incorporate these Q&As directly. When planning literature searches, reference the specific expectations. When building equivalence demonstrations, follow the clarified logic. When justifying data sufficiency, use the proportionality framework.

The gap between compliant and non-compliant clinical evaluations is widening. MDCG 2021-6 defined that gap more clearly. Now we know which side we need to be on.