Premarket clinical evidence for medical devices: what MDR and ISO 14155 expect

Before a medical device reaches patients in Europe, its manufacturer has to show, with clinical evidence, that it performs as intended and that its benefits outweigh its risks. The Medical Device Regulation (EU) 2017/745 (MDR) provides the regulatory framework for the requirements regarding clinical evidence. When a clinical investigation is required, manufacturers need to plan sufficient time and budget to collect the premarket clinical evidence properly.

This article explains what the MDR expects from premarket clinical evidence and how the international standard for clinical investigations, ISO 14155 (fourth edition, March 2026), fits in.

What the MDR requires

Under the MDR, clinical evaluation is a mandatory part of proving that a device is safe and meets its intended purpose. Demonstrating conformity with the general safety and performance requirements “shall include a clinical evaluation in accordance with Article 61” (Article 5(3)). Manufacturers carry it out in line with Article 61 and Annex XIV, including post-market clinical follow-up (PMCF) (Article 10(3)).

The definitions in Article 2 show how the regulation thinks about evidence:

• Clinical evaluation is “a systematic and planned process to continuously generate, collect, analyse and assess clinical data” to verify safety and performance (Article 2(44)). It is not finished at certification.
• Clinical evidence is clinical data and evaluation results of sufficient amount and quality to assess whether the device is safe and achieves its intended clinical benefit (Article 2(51)).
• Clinical data can come from clinical investigations, studies of an equivalent device, peer-reviewed literature, and post-market surveillance, in particular PMCF (Article 2(48)).
  
  

  When is a clinical investigation required?

  Under the MDR, a clinical investigation is mandatory for implantable devices and class III devices unless one of four narrow exemptions applies, as explained in MDCG 2023-7. Each exemption carries its own conditions, which the manufacturer has to document.

  A clinical investigation is “any systematic investigation involving one or more human subjects, undertaken to assess the safety or performance of a device” (Article 2(45)). According to guidance document MDCG 2023-7, Article 61(4) requires clinical investigations for implantable devices and class III devices, with a limited set of exemptions:

  • a modification of the manufacturer’s own marketed device, where equivalence is shown and endorsed by the notified body;
  • a device equivalent to another manufacturer’s device, where a contract gives ongoing full access to its technical documentation;
  • legacy devices with sufficient clinical data and compliance with relevant common specifications; and
  • well-established technologies such as sutures, staples, screws, plates and clips, under the same conditions.

  Where no exemption applies, the guidance states that a clinical investigation is mandatory for these classes of devices, and at minimum a pivotal one. For non-implantable devices and class II devices, the need for an investigation follows from the gaps the clinical evaluation identifies.

  Investigational devices must meet Articles 62 to 80 and Article 82 of the MDR and Annex XV (Article 21(1)(a)), which cover how investigations are applied for, conducted and reported.

  ISO 14155:2026: good clinical practice for device investigations

  ISO 14155 specifies good clinical practice (GCP) for the design, conduct, recording and reporting of clinical investigations in human subjects. The fourth edition, published in March 2026, replaces the 2020 edition. It aims to protect subjects’ rights, safety and well-being, ensure credible results, define sponsor and investigator responsibilities, and support conformity assessment. For background on the standard, see our ISO 14155 overview.

  Its principles include ethical conduct in line with the Declaration of Helsinki, a favorable benefit-risk balance, a scientifically sound clinical investigation plan (CIP), prior ethics committee approval (according to country specific guidelines), informed consent, qualified personnel and quality systems covering every aspect of the study.

  Planning. The design must be justified by the clinical evaluation and risk assessment, including endpoints, comparator, randomization and blinding. The CIP, investigator’s brochure and case report forms are prepared up front, monitoring is risk-based, and sites are selected and qualified.

  Conduct. The study cannot start before ethics committee approval and, where required, regulatory authority approval. During the study, the sponsor manages adverse events and device deficiencies, controls documents, validates data systems and accounts for every device and subject.

  Close-out and reporting. The investigation ends with a clinical investigation report. Registration in a public database is expected before recruitment, with results entered at completion.

  Several changes in the 2026 edition matter for study planning. It provides:

  • a clearer distinction between risks of using the device and risks of procedures required by the CIP outside routine care, plus a required residual risk assessment;
  • a new section on clinical events committees; and
  • the non-mandatory principles of estimands in study design and analysis.
  
  

  How the MDR and ISO 14155 fit together

  The MDR decides whether clinical evidence is needed and what must be demonstrated; ISO 14155 describes how a clinical investigation is run so that its data are credible and subjects are protected.

  The two meet at several points along the route from evidence gap to market:

  1. Clinical evaluation identifies the gap. It shows which claims lack sufficient data.
  2. The design is justified from the evaluation. ISO 14155 requires the design to be based on the evaluation of pre-clinical and clinical data and the risk assessment (clause 6.3).
  3. Approvals are obtained. The study does not start before ethics committee and, where required, authority approval (clause 7.1). Clinical investigations are also recorded in Eudamed (MDR Article 33).
  4. The study is conducted to GCP. Risk-based monitoring, safety reporting and validated data systems provide the audit trail.
  5. Results feed back. The report updates the clinical evaluation, which is kept current through PMCF.

  A sponsor therefore works with two sets of expectations at once: the EU regulation with its national implementation, and a GCP standard. A good plan reconciles both from the start.

  Common pitfalls in premarket evidence generation

  Most delays trace back to a handful of recurring issues, many of them visible long before a study starts.

  • Study design disconnected from the clinical evaluation. ISO 14155 expects the design to follow from the Clinical Evaluation Plan and risk assessment; a protocol written in isolation struggles to justify its endpoint, comparator and sample size.
  • Assuming an exemption applies. Equivalence, legacy status and well-established technologies each carry conditions, such as ongoing access to another manufacturer’s documentation (MDCG 2023-7). Check them early and document them.
  • Unclear risk logic. The 2026 edition separates risks of using the device from risks of required procedures outside routine care. Plans that blend the two tend to draw questions at review.
  • Loosely defined endpoints. Objectives should translate directly into pre-specified primary endpoints, with biostatistical input. Later changes require documented amendments.
  • Incomplete submission packages. Ethics committees receive at least the CIP, investigator’s brochure, informed consent form, recruitment materials and the principal investigator’s CV. Gaps cause clock-stops and extra review rounds.

How Qserve can support

Qserve has focused exclusively on medical devices and IVDs since 1998. Its clinical work is delivered by Qserve Clinical Services, a specialized medical device and IVD contract research organization (CRO), together with the Qserve regulatory consultancy, so strategy and execution can be planned together. The table shows where that support typically fits.

Final thoughts

Premarket clinical evidence is a planning exercise before it is a study. Manufacturers who move fastest decide early what the clinical evaluation can and cannot support, design any investigation from that gap, and prepare submissions with the same care as the protocol. With the fourth edition of ISO 14155 now published, it is also a good time to check whether plans and templates reflect the updated expectations.

If you would like to discuss your own evidence strategy, or have a clinical evaluation, protocol or submission package reviewed, the Qserve clinical team is happy to talk. Qserve offers a free 30-minute consultation through its clinical strategy support page.

This article is a general overview and does not replace the text of the regulation, the standard or advice on a specific device.