Why Your Clinical Data May Not Count as Clinical Data Under MDR

What a Notified Body reviewer accepts, with Dr. Engin Çamer (SZUTEST NB2975) and Dr. Hatem Rabeh (Clinical Evaluation Navigator)

Manufacturers lose months when a Notified Body decides their evidence does not meet the MDR definition of clinical data. A retrospective image dataset, or clinical use logged outside Europe: both can look like clinical data and fail at review. In this session, Dr. Engin Çamer (internal clinician at SZUTEST NB2975) and Dr. Hatem Rabeh take the Article 2(48) definition apart word by word, then map the four sources a reviewer will accept, including where software and AI evidence holds or breaks.

Key Takeaways from This Webinar

Clinical data is defined by Article 2(48). Clinical data is information concerning safety or performance that is generated from the use of the device. The key phrase is generated from the use of the device, in or on living human subjects. The definition itself contains no timing word such as prospective or retrospective.

There are four sources of clinical data. Article 2(48) lists clinical investigation of the device, clinical investigation or other studies of an equivalent device, peer-reviewed reports on other clinical experience of the device or an equivalent device, and clinically relevant information from PMCF.

Retrospective analysis is clinical data, not a clinical investigation. MDCG confirms that data analysed retrospectively, separately from the decision to use the device, does not meet the definition of a clinical investigation. It can still be used as clinical data, but it is appraised differently.

Equivalence is not similarity. A similar device is not a source of clinical data: it informs the state of the art and benchmarking. To rely on an equivalent device you must demonstrate technical, biological, and clinical equivalence under Annex XIV, and the equivalent device must be CE-marked (MDCG 2020-5).

Implantable and class III devices default to a clinical investigation. Article 61(4) requires a clinical investigation, with four exceptions: equivalence to your own CE-marked device, contract access to an MDR-certified device, legacy devices with valid CE marking (MDCG 2020-6), and devices on the Article 61(6)(b) well-established technology list.

Article 61(10) is a real pathway when justified. For some devices, clinical evaluation does not have to be based on clinical data, but only when the state of the art supports it and the manufacturer justifies it in the clinical documentation. MDCG 2020-6 notes there is no formal definition of sufficient clinical evidence under MDR.

Software clinical benefit usually sits in patient management. MDCG 2020-1 notes the clinical benefit of medical device software often differs from physical devices. In most cases it is expressed as driving or informing patient management under Article 2(53), rather than a direct patient outcome.

AI and software evidence rests on three pillars. Notified Bodies evaluate software through valid clinical association, technical (analytical) validation, and clinical validation. Retrospective image datasets can support clinical validation only when ground truth, labelling, data quality, and representativeness cover the full intended population and all disease stages.

What We Cover in This Session

● The Article 2(48) definition of clinical data, and why every word matters

● The four sources of clinical data and how a Notified Body reads each one

● Clinical investigation under Article 2(45), and what involvement of human subjects means

● Where retrospective studies fit: clinical data compared with clinical investigation

● Prospective compared with retrospective, and the exposure to the device concept (MDCG 2021-6)

● Equivalence compared with similarity, and when equivalent device data can be used (Annex XIV, MDCG 2020-5)

● The Article 61 routes: 61(1), 61(3), the 61(4) exceptions, and the 61(10) pathway

● Sufficient clinical evidence, and how to justify the level of clinical evidence (MDCG 2020-6, MEDDEV 2.7/1 Rev. 4)

● Clinical data and clinical benefit for medical device software (MDCG 2020-1, IMDRF)

● Valid clinical association, technical validation, and clinical validation for AI and software

● Using retrospective image datasets: ground truth, data quality, and representativeness

Meet the Speakers

Dr. Engin Çamer

Clinical Unit Manager, SZUTEST GmbH (Notified Body 2975)

Dr. Engin Çamer is a medical doctor working as an internal clinician for SZUTEST, a Notified Body designated in Germany under the EU MDR. With approximately 15 years of experience in the medical device field, he is directly involved in evaluating clinical data for medical devices, including clinical evaluation reports, benefit-risk documentation, and PMCF plans submitted by manufacturers seeking CE marking.

Dr. Hatem Rabeh

Medical Doctor | Biomedical Engineer | Clinical Evaluation Expert

Dr. Hatem Rabeh is a Medical Doctor with a Master of Science in Biomedical Engineering and over 8 years of experience as a Clinical Evaluation Expert for medical devices. He has supported manufacturers across all device classes, from Class I to Class III, including AI software, robotics, navigation systems, and implantable devices.

As the founder of Clinical Evaluation Navigator, Dr. Rabeh has published over 100 articles on clinical evaluation best practices and has supported multiple medical device startups and established companies with clinical evaluation strategies, SOTA reports, and regulatory documentation required for MDR compliance.

Join Dr. Rabeh’s community →

Who Should Watch This Webinar?

This session is for anyone deciding what evidence supports a clinical evaluation under the EU MDR:

✓ Clinical Evaluation Specialists deciding what qualifies as clinical data in a CER
✓ Regulatory Affairs Managers choosing the clinical evidence route under Article 61
✓ Medical Device Software and AI developers preparing clinical association and validation evidence
✓ Manufacturers of implantable and class III devices weighing a clinical investigation against the Article 61(4) exceptions
✓ Quality and PMCF Managers linking post-market data back to the clinical evaluation
✓ MedTech Startups and SMEs building their first clinical evaluation strategy under MDR

Also Watch: More Notified Body Webinars

Two earlier sessions with SZUTEST NB2975 cover related ground: how Notified Bodies interpret the State of the Art during review, and how to CE mark AI software under the MDR and the AI Act.

Watch the AI Software Webinar →
Watch the SOTA Webinar →

Continue the Conversation

Join the Clinical Evaluation Navigator community to access weekly expert calls, discuss real-world challenges with regulatory professionals, and get direct guidance from Dr. Hatem Rabeh on clinical data, clinical evaluation, and MDR compliance.

Join the Community

Free to join • Weekly Friday calls • 100+ published articles

Related Articles

I create your clinical strategy and documents, including CEP, CER, PMCF, IP, and IB, in line with MDR and MDCG guidelines. My focus on efficiency and precision ensures swift CE approval for your medical device.

Contact Info

Get In Touch

Overcome MDR and MDCG hurdles with my proven method. I offer personalized guidance to accelerate your CE marking through expertly crafted clinical documentation and strategy. Ready to boost your medical device’s journey? Reach out to me today.