Author: Stephanie Knighten, Senior Director of Medical Device Writing Services
Last week I attended RAPS Convergence in Charlotte, and one of the sessions I found especially useful was “MDR/IVDR State of Play.”
There has certainly been no shortage of discussion about MDR and IVDR over the past several years. Most of us working in this space have lived through the transition periods, capacity concerns, changing guidance, certification timelines, EUDAMED delays, and the growing list of regulatory requirements that interact with the regulations.
What made this session useful was that it moved beyond simply describing those problems. The discussion was really about what the next version of the system could look like.
The European Commission’s December 2025 proposal to revise the MDR and IVDR is intended to simplify the framework and reduce administrative burden while continuing to protect patient safety and device availability. The proposal touches a broad range of areas, including certification, coordination, digitalization, international considerations, and the interaction of MDR/IVDR with requirements for areas such as cybersecurity and artificial intelligence.
But I think the more interesting theme underneath all of those individual changes is proportionality.
From fixed requirements to risk-based oversight
One of the concepts presented during the session was a move away from uniformly fixed procedures and toward a more risk-based approach.
That makes sense to me.
A regulatory framework should absolutely establish a high bar for safety and performance. But applying the same intensity of oversight regardless of what we know about a device is not necessarily the same thing as improving safety.
A new device with limited clinical experience should not necessarily be treated in the same way as a mature device with years of clinical use, stable performance, a well-characterized risk profile, and substantial post-market evidence.
That idea came through particularly clearly in the Notified Body portion of the session.
Sabina Hoekstra-van den Bosch discussed the concept of risk-adaptive surveillance, where surveillance intensity could decrease as confidence in a device increases. Importantly, that would not be a one-way decision. The model presented was evidence-driven: if new information changed the risk picture, surveillance could increase again.
That feels fundamentally different from simply reducing oversight.
It is about directing oversight where the evidence says it is needed.
And I think that distinction matters.
What does that mean for clinical evidence?
For those of us working in clinical evaluation and post-market evidence, there is another side to this.
A more proportionate regulatory system only works if the evidence supporting those decisions is reliable.
If surveillance intensity is going to reflect confidence in a device, then manufacturers need to be able to demonstrate why that confidence is justified.
Clinical evaluation, PMS, PMCF, vigilance, complaints, trend data, real-world experience, and risk management cannot operate as separate exercises that happen to generate different regulatory documents. Together, they should tell a consistent story about how the device is actually performing.
The discussion around EUDAMED fit into this as well. The Notified Body perspective emphasized the value of having access to meaningful registration, certificate, and vigilance information so that emerging signals can be identified and oversight can be better targeted.
In theory, that means less duplication where the evidence is reassuring and more attention where the evidence indicates a problem.
That is a much better use of regulatory resources than applying the same activity at the same frequency simply because the calendar says it is due.
PMCF is a good example
Gert Bos’s industry perspective included several changes that manufacturers are likely to care about: improved timelines, removal of the five-year recertification requirement, and treating updated PMCF information as part of ongoing surveillance rather than automatically as a reason to re-review the device.
The PMCF point is particularly relevant to clinical evaluation.
PMCF is supposed to be part of a continuous clinical evaluation process. New information should feed back into the CER, risk management, PMS, and the overall benefit-risk assessment.
But there is a difference between evaluating new information and reopening an assessment simply because new information exists.
If the incoming data remain consistent with the established safety and performance profile, an ongoing surveillance model seems much more aligned with the lifecycle approach that MDR itself is intended to create.
If those data identify a new risk, worsening trend, change in performance, or other meaningful signal, then that should drive additional scrutiny.
Again, the regulatory activity follows the evidence.
The proposal also acknowledges some very practical problems
There was also a practical side to the discussion that I appreciated.
Notified Bodies have to maintain qualified staff and specialist expertise, and that capacity is not unlimited. Sabina’s presentation addressed Notified Body sustainability directly, including the challenge of retaining the people needed to conduct assessments.
That matters because a regulatory system cannot function well if highly specialized resources are consistently being used for activities that add limited regulatory value.
The same applies on the manufacturer side.
Every hour spent responding to unnecessary duplication, administrative reassessment, or overlapping requirements is an hour that cannot be spent investigating a signal, improving a PMCF strategy, strengthening clinical evidence, or addressing an actual device risk.
Reducing unnecessary burden is therefore not inherently at odds with patient safety.
Done correctly, it may help regulatory systems focus more effectively on it.
There are still some important details to work through
The session was not simply an endorsement of the proposal.
There are still open questions, and implementation will matter enormously.
One of the items raised from the industry perspective was the continued need for a clearer definition of significant change. Other considerations included EUDAMED functionality, post-market requirements, Notified Body processes and timelines, and how the revised framework will ultimately operate in practice.
The relationship between MDR/IVDR and the EU AI Act is another area worth watching.
One of the directions discussed during the session was greater integration of AI-related requirements into the medical device framework rather than forcing manufacturers to navigate two parallel regulatory systems for the same product.
For companies developing software and AI-enabled medical devices, how that interplay is ultimately resolved could have very practical consequences.
Where I landed after the session
Erik Vollebregt described the Commission proposal as a competent and well-targeted attempt to address genuine implementation problems.
I think that captures the tone of the discussion well.
The proposal does not appear to be about walking back the underlying objectives of MDR and IVDR. It is about asking whether the regulatory mechanisms being used to achieve those objectives are always proportionate to the risk and the evidence available.
That is an important distinction.
After several years of MDR and IVDR implementation, we have much more information about where the framework works, where it creates unintended burden, and where limited regulatory resources are being consumed without a corresponding improvement in patient safety.
The next phase should be able to learn from that experience.
For me, the most promising direction discussed at RAPS was not simply simplification.
It was the possibility of a system in which the level of regulatory scrutiny changes as the evidence changes.
That still requires strong clinical evidence, meaningful post-market surveillance, capable Notified Bodies, and robust regulatory oversight.
It just uses those tools more deliberately.
And that feels like a worthwhile direction for the next phase of MDR and IVDR.
