Clinical Investigations Under EU MDR: When You Need One and How to Plan It

Running a clinical investigation is the most direct way to generate clinical evidence for your device — but it is also the most resource-intensive, and most manufacturers would rather avoid it if alternatives are available. Understanding when you genuinely need one, and what is involved if you do, is one of the more important planning questions in the clinical evaluation process.

When a clinical investigation is actually required

MDR does not require a clinical investigation for every device. Article 61 gives you three routes for generating clinical evidence: clinical investigations, clinical literature on your own device, and data from equivalent devices. Whether you need to run an investigation comes down to whether the other routes can give you sufficient evidence.

For Class III devices and implantable devices, Article 61(4) adds a presumption: you generally need clinical investigations unless your device is based on a predicate that already has sufficient clinical data, or you can demonstrate equivalence with access to the predicate's technical documentation. For many teams working on novel Class III devices — or devices where equivalence can't be established — a clinical investigation isn't optional.

The other situation that drives investigations is data gaps. If your existing literature doesn't cover your specific intended purpose, target population, or use environment adequately, and you can't close those gaps through post-market data, a pre-market investigation may be the only viable path.

What Article 62 and Annex XV require

Article 62 sets the ethical and procedural requirements for clinical investigations. The key requirements:

• Investigations must be designed to protect the rights and safety of subjects. This means proper informed consent procedures, independent ethics committee approval, and a Data Safety Monitoring Board (DSMB) for higher-risk studies.
• A clinical investigation plan (CIP) — defined in Annex XV Part A — must be produced before the investigation begins. The CIP defines the objectives, design, methodology, statistical approach, monitoring plan, and stopping rules.
• Investigations must be registered in EUDAMED before enrolment begins. The registration is part of the transparency requirements under MDR.
• Sponsor obligations include reporting serious adverse events, device deficiencies, and urgent safety measures to competent authorities within defined timelines.

The competent authority in each member state where the investigation will run must be notified (or approve, for certain device classes). The process is parallel to the ethics committee review, and the timelines from both must be satisfied before enrolment can start.

The approval process in practice

One thing that catches teams off guard is how long the regulatory and ethics pathway takes. A multi-country investigation can take 6–12 months from submission to first patient enrolled, depending on the member states involved, the device risk class, and how complete your submission package is.

The submission to each competent authority requires: the CIP, the Investigator's Brochure (IB), the risk management file for the investigation, the informed consent documents, the protocol risk analysis, and the sponsor's quality management information. Gaps in any of these generate queries and delay the clock.

MDCG 2022-2 (on clinical investigations under MDR) is the practical guide for this pathway. It covers what must go into the CIP, how the competent authority assessment is structured, and what the EUDAMED registration requirements look like in practice. Read it before you start drafting your CIP, not after.

Designing the investigation to answer the right questions

The most common design mistake is treating the investigation as a documentation exercise rather than a scientific one. Investigations that produce data the NB can't use — because the endpoints don't map to clinical claims, the statistical analysis doesn't address the primary objective, or the patient population doesn't match the intended use — generate major findings at technical review.

A well-designed investigation answers specific, pre-defined questions about safety and performance that your CER needs answered. The CIP objectives should connect directly to the open questions in your clinical evaluation. If you can't articulate what gap in the CER this investigation is designed to close, the design needs more work before you submit.

Post-investigation: feeding data into the CER

Once the investigation is complete, the clinical investigation report (CIR) becomes a primary evidence source for your CER. The CIR needs to be appraised like any other study — with critical evaluation of what was found, not just a summary of the results. Notified Bodies will compare the CIR endpoints with the pre-specified objectives in the CIP to check for post-hoc outcome switching, which is a known manipulation to manage.

The investigation data also feeds forward: if your investigation identified any residual risks or open clinical questions, those become inputs to your PMCF plan. The loop from CER to investigation to PMCF is the core of a defensible clinical evidence strategy for novel high-risk devices.