EU MDR clinical evaluation

EU MDR Article 61 requires every device’s safety and performance to rest on sufficient clinical evidence, planned, gathered and reported under EU MDR Annex XIV Part A and kept current by post-market clinical follow-up under Part B. You set and justify the level of evidence. Implantable and Class III devices normally need clinical investigations.

The requirement in plain English

Sufficient evidence, justified by you (EU MDR Article 61(1)). Conformity with the relevant safety and performance requirements, the side-effects and the benefit-risk balance must rest on clinical data. You decide what level of evidence your device and claims need, and you justify it.

A defined method (EU MDR Article 61(3)). The evaluation draws on three sources: a critical review of the literature, valid only for an equivalent device and adequate data; the results of all available clinical investigations; and the alternative treatments available.

The plan, the appraisal and the report (Annex XIV Part A). You write a clinical evaluation plan that names the requirements needing clinical support, the intended purpose and target groups, the clinical benefits with measurable outcomes, the methods for safety and benefit-risk, and a clinical development plan. You identify data through a systematic literature review, appraise each item for quality and relevance, generate new data where gaps remain, and analyse the whole. Favourable and unfavourable data both go in. The clinical evaluation report records it and belongs in the technical documentation (EU MDR Article 61(12)).

Equivalence (Annex XIV section 3). You may rely on another device’s data only if the two are similar in technical, biological and clinical characteristics, with no clinically significant difference. For a device made by someone else, the implantable and Class III route needs a contract giving full, ongoing access to its technical documentation (EU MDR Article 61(5)).

Clinical investigations for higher risk (EU MDR Article 61(4)). Implantable and Class III devices need them unless the device modifies your own marketed device and the notified body endorses equivalence. Two exemptions in EU MDR Article 61(6) cover devices lawfully marketed under the old directives with sufficient data, and a list of well-established devices that Delegated Regulation (EU) 2026/1451 lengthened in 2026, adding items such as needles, guidewires, bone fillers and reusable surgical instruments.

When clinical data are not appropriate (EU MDR Article 61(10)). You may show conformity on non-clinical testing alone only with a justification grounded in risk management, how the device interacts with the body, the clinical performance intended and your claims, set out in the EU MDR Annex II documentation.

Kept current (EU MDR Article 61(11)). Post-market clinical follow-up (PMCF), planned under Annex XIV Part B, and post-market surveillance data update the evaluation for the device’s life. The PMCF plan sets the methods and objectives, and the PMCF evaluation report records the findings. For Class III and implantable devices that report is updated at least annually.

Expert panels (EU MDR Article 54). For Class III implants and certain Class IIb active devices, the notified body asks an expert panel to review your clinical evaluation.

Medical Device Coordination Group guidance helps here: MDCG 2020-6 on sufficient clinical evidence for legacy devices, and MDCG 2020-13 on how notified bodies assess a clinical evaluation.

This is guidance to help you understand the requirement, not legal advice.

What good looks like

A 35-person company makes Class IIa software that triages photographs of skin lesions for dermatologist review. Its clinical evaluation plan states the intended purpose, users and patient group, and defines clinical benefit as sensitivity and specificity for melanoma against histology, with acceptance thresholds drawn from the state of the art.

The literature review runs to a written protocol: databases, search strings, dates, inclusion criteria and an appraisal score for every paper, so another evaluator could repeat it. Published data on similar algorithms cannot prove this algorithm’s performance, so the company runs its own clinical performance study and records why equivalence was not claimed.

The clinical evaluation report is written by a dermatologist and a regulatory scientist whose CVs sit in the file. It includes unfavourable results, such as lower sensitivity on darker skin types, and states how the labelling and the PMCF plan respond. The PMCF plan monitors real-world performance on a sample referred for biopsy, and each PMCF evaluation report feeds the next periodic safety update report and the risk file.

Evidence an auditor expects

Clinical evaluation plan (required): the notified body checks it covers Annex XIV section 1 and sets measurable benefits.
Literature search protocol and appraisal: usual evidence that the review was systematic and repeatable.
Clinical evaluation report (required, in the technical documentation): conclusions that follow from the data, with favourable and unfavourable results.
Equivalence justification: for each claimed device, the three characteristics compared and, where required, the access contract.
Clinical investigation or performance study reports: where the evidence needed them.
PMCF plan and evaluation report (required unless justified): methods, objectives, timetable and findings.
Evaluator qualifications: CVs showing relevant clinical and methodological expertise, which notified bodies ask for.

Common mistakes

Equivalence on looks alone

Claiming equivalence because two devices are technically alike, with no biological or clinical comparison and no access to the other device’s data, fails quickly. Compare all three characteristics and state what you can actually see.

A literature review nobody could repeat

Ad hoc searches with no protocol give the notified body no way to test completeness. Write the protocol first and record every exclusion.

Using Article 61(10) as a shortcut

The exception for non-clinical evidence needs a reasoned justification. “Software does not touch the patient” is not one; the decisions it informs do.

A report that never changes

A clinical evaluation report signed at certification and untouched since is out of date at the first surveillance audit. Tie its update cycle to the PMCF evaluation report and the PSUR.

“No PMCF needed” without a reason

Deciding PMCF is unnecessary is allowed only with a justification in the plan. Without one, auditors treat it as missing.

How Sted helps

Sted’s knowledge base includes a Clinical Evaluation and PMCF Procedure (PRC-CEP), tagged to EU MDR clinical evaluation, covering the clinical evaluation plan for each device, how sufficient evidence is shown from your own data or equivalence, the literature search and appraisal, the clinical evaluation report and its update cycle, and the post-market clinical follow-up plan and evaluation report.
For EU MDR Sted’s knowledge base carries five procedures, the Clinical Evaluation and PMCF Procedure (PRC-CEP), Post-Market Surveillance and Vigilance Procedure (PRC-PMSV), Enhanced Summative Evaluation Procedure (PRC-ESE), Risk Management Procedure (PRC-RMP) and Incident Management Procedure (PRC-IMP), and eight record templates: the Post-Market Surveillance Plan (REC-PMS), Post-Market Surveillance Review Record (REC-PMSR), Periodic Safety Update Report (REC-PSUR), Medical Device Incident Record (REC-MDI), Medical Device Risk Management File (REC-MDRM), Human Factors Engineering Report (REC-HFE), Enhanced Summative Evaluation Requirement Review (REC-ESER) and Problem Report (REC-MDPR), with six seeded requirement rows and no control catalogue.
Sted drafts your policies and procedures from its knowledge base, adapted to your organisation, for you to review.
Approving sets an effective date and a review date, which starts at the effective date plus the document’s review frequency; a review task opens 30 days before the review date and turns overdue after it.
Under the Post-Market Surveillance and Vigilance Procedure (PRC-PMSV) Sted’s knowledge base carries a Post-Market Surveillance (PMS) Plan (REC-PMS) for each device, setting out what is gathered, from where, by whom and how often, a Post-Market Surveillance (PMS) Review Record (REC-PMSR) and a Periodic Safety Update Report (PSUR) (REC-PSUR) covering the benefit-risk balance, clinical follow-up, sales and use, and serious incidents for each reporting period.
Records are part of the Full Management System plan.

Sted does not run a clinical evaluation, appraise clinical data or decide that evidence is sufficient. Qualified evaluators do, and the notified body assesses their work.

EU MDR: where clinical evaluation sits among the regulation’s steps.
EU MDR post-market surveillance and vigilance: the data that keep the evaluation current.
EU MDR technical documentation: where the report and the PMCF plan sit.
ISO 13485 design and development: design validation, which clinical evaluation completes.
Document control: versioning the plan and report as they change.
Evidence: what counts as proof for an auditor.

Frequently asked questions

Can I use equivalence to a competitor’s device under EU MDR?

Only with difficulty for implantable and Class III devices: you need a contract giving you full, ongoing access to the other manufacturer’s technical documentation, which competitors rarely grant. For other classes you must still show equivalence in technical, biological and clinical characteristics and have enough access to the other device’s data to justify it.

Do I need a clinical investigation for a Class IIa device?

Not always. The regulation requires clinical investigations by default for implantable and Class III devices; for a Class IIa device you can build the evidence from literature, equivalence and your own data, if together they are sufficient for your claims. Many software manufacturers still run a clinical performance study because published data on their own algorithm do not exist.

How often must a clinical evaluation report be updated?

Throughout the device’s life, as post-market and follow-up data arrive. For Class III and implantable devices the PMCF evaluation report, and the summary of safety and clinical performance where one exists, are updated at least once a year. For other classes the update cycle follows your post-market surveillance and periodic safety update reports.