Clinical Evaluation of Medical Devices Under the MDR

Demonstrating a medical device’s conformity with the relevant general safety and performance requirements of the Medical Device Regulation (MDR), formally Regulation (EU) 2017/745, requires a documented clinical evaluation. Many manufacturers underestimate the effort involved, especially when robust clinical data is lacking, an equivalent product cannot be plausibly justified, or individual product claims are not supported by evidence. This article outlines the requirements, processes, and typical pitfalls.

The clinical evaluation is essential for market access

The clinical evaluation is mandatory for all risk classes under Article 61 of the MDR and Annex XIV, Part A. It is not a one-time document but an ongoing process throughout the entire product lifecycle, from pre-market planning to post-market clinical follow-up (PMCF).

The Clinical Evaluation Plan (CEP) and the Clinical Evaluation Report (CER) are among the most rigorously reviewed parts of the technical documentation. Incomplete or insufficiently substantiated evidence leads to additional requests, delayed certification, or, in the worst-case scenario, rejection. For software classified as a medical device and AI-based medical devices, there is the additional challenge that demonstrating clinical benefit for algorithmic products is often more difficult to substantiate through literature or equivalence than it is for established product categories. Furthermore, every claimed clinical performance or benefit statement must be supported and reviewed by appropriate clinical evidence.

This requirement is particularly unfamiliar with startups and software companies that are new to the medical device market. Unlike in traditional software testing, it is not enough for an algorithm to function technically correctly. What is crucial is demonstrating that it delivers the claimed medical benefit for the intended user group under the intended clinical conditions of use.

Fundamentals: terminology and legal framework

Clinical evaluation is generally based on the totality of available clinical data and may combine clinical investigations, scientific literature, data from PMS/PMCF activities, and other appropriate clinical findings. Furthermore, Article 61 of the MDR requires that available alternative treatment options be taken into account when conducting a review of safety, performance, and clinical benefits. MDCG guidelines specify these requirements.

Definition	Meaning
Clinical Evaluation Plan (CEP)	Outlines the scope, methodology, and data sources of the Review
Clinical Evaluation Report (CER)	Documents the results and review of all clinical data used
Equivalence	Reliance on data from a technically, biologically, and clinically equivalent product. Under the MDR, sufficient access to the technical documentation of the reference product must also be demonstrated on a regular basis.
Clinical Claim	Any medical effect or performance attributed to the Product and derived from its intended purpose. This term is frequently used in regulatory practice but is not explicitly defined by the MDR.
Clinical Benefit	Clinical benefit, as defined in Article 2(53) of the MDR, refers to the positive effects of a product on a person’s health or on patient management, including diagnostic or therapeutic benefits.
PMCF	Systematic post-market clinical follow-up to continuously attest to safety and performance

Article 61(10) of the MDR permits, in justified exceptional cases, the waiver of clinical data in favor of non-clinical evidence. However, the waiver of clinical data is an exception that must be interpreted narrowly. It applies only if evidence based on clinical data is deemed unsuitable and requires justification based on risk management. This exception is not available for implantable devices or Class III products. Even when the exception is applied, a clinical evaluation remains necessary.

Relevant MDCG guidelines include MDCG 2020-5 (Equivalence), MDCG 2020-6 (clinical evidence for existing devices), MDCG 2020-13 (template for clinical evaluation reports by Notified Bodies), and MDCG 2024-10 for the clinical evaluation of orphan medical devices. For software medical devices, MDCG 2020-1 contains a separate framework for the Clinical evaluation or Performance evaluation of medical device software (MDSW). The MEDDEV 2.7/1 Rev. 4 guideline no longer has independent regulatory binding force under the MDR but continues to be used by many manufacturers and Notified Bodies as a methodological reference framework. For the post-market phase, MDCG 2025-10, published at the end of 2025, specifies the requirements for post-market surveillance. Furthermore, as things stand, ISO 18969 is a draft standard that systematically describes the clinical evaluation process throughout the product lifecycle. Subject to further standardization work, no binding requirements have yet been established based on this draft.

Clinical evaluation is closely linked to risk management in accordance with ISO 14971. Findings from clinical evaluation, PMS, PMCF, and vigilance must be continuously incorporated into the risk-benefit assessment and risk control measures.

Clinical claim: definition and scope

Through the intended purpose, the manufacturer specifies the medical purpose for which the product is to be used. The clinical claim is also derived from the intended purpose; that is, the specific medical effect or performance attributed to the product, such as a certain level of diagnostic accuracy, a reduction in complications, or an improvement in the course of a disease. Clinical claims typically arise from the intended purpose, but also from the label, instructions for use, performance descriptions, and other manufacturer information regarding the Product.

The clinical claim must be distinguished from two related terms:

Intended purpose:
Describes the intended use of the product as a whole and serves as the basis for qualification, classification, and risk management. The clinical claim specifies a single, verifiable statement of efficacy based on this.
Marketing claim:
Aimed at users or patients, it is subject to labeling and advertising requirements (including Art. 7 MDR and the prohibition against misleading claims). A marketing claim must not go beyond what is supported by the clinically substantiated claim.

Each clinical claim must be individually supported by clinical data. A general, undifferentiated statement of benefit is not sufficient. If new claims are added or existing ones expanded during the course of product development, the clinical evaluation must be updated accordingly.

Practical approach

Define the intended purpose and clinical claims:
The starting point is the precise intended purpose. Clinical claims are derived from this, each of which requires its own evidence.
Create a Clinical Evaluation Plan (CEP):
The plan specifies, for each claim, the search strategy for the literature review, potential equivalent products, and evaluation criteria.
Systematically collect and review data:
Literature searches, the evaluation of existing clinical investigations, and, where applicable, the justification of equivalence is conducted according to a predefined, systematic, and transparently documented methodology and reported separately for each claim.
Draft a Clinical Evaluation Report (CER):
The CER summarizes the data for each claim and reviews whether it provides evidence of safety and clinical benefit.
Integrate the PMCF Plan:
Any remaining uncertainties, including those related to individual claims, are addressed in the PMCF plan and incorporated into the ongoing updates to the CER. Not every medical device necessarily requires PMCF measures. However, any decision to waive such measures must be justified.

The CER is not a static document and is regularly updated throughout the product lifecycle, particularly in light of new findings from vigilance or PMCF, as well as whenever there is a change to the intended purpose or claims.

Challenges and risks

Justification of equivalence is rigorously tested:
Notified bodies require detailed access to technical data for the comparator product, which is often not available in practice.
Lack of proprietary clinical data for new product categories:
For example, with AI-based software, it is often not possible to rely on established equivalent products.
Claims without individual substantiation:
If multiple functions or value propositions are substantiated in a general manner rather than individually, the evidence may be considered incomplete.
Literature review without documented methodology:
A non-systematic or incompletely documented search is a common deficiency identified during the review.
PMCF is treated as a secondary consideration:
Without a robust PMCF plan, the clinical evaluation remains incomplete and will be revisited during recertification.

FAQ

Is clinical evaluation required for every medical device?

Yes, it is mandatory for all risk classes under Article 61 of the MDR. The scope and depth depend on the risk class, intended purpose, and number of clinical claims made.

What is the difference between intended purpose and a clinical claim?

The intended purpose describes the overall purpose for which the device is intended. The clinical claim is the individually verifiable statement regarding a medical effect or performance derived from the intended purpose.

Is literature data sufficient without conducting a clinical investigation?

This is possible provided that the available literature or equivalence data supports the demonstration of safety and performance for each claim. If the available data is insufficient, a separate clinical investigation may be required.

How often must the CER be updated?

The MDR does not specify a fixed deadline. The key factor is a risk-based update interval that takes into account PMCF results, new findings, and changes to claims. For products in higher risk classes, Notified Bodies often expect annual or risk-based updates, particularly when new PMS, vigilance, or PMCF data are available. For certain product categories - especially Class III and implantable devices - shorter update intervals may be required due to additional MDR requirements and the link to the PMS/PSUR process.

Does the clinical evaluation also apply to software and AI products?

Yes. The MDR makes no exception here. The challenge lies in selecting suitable data sources and equivalent products, as well as in providing separate evidence for each claim.

Conclusion

The clinical evaluation is one of the most technically demanding parts of the technical documentation and, at the same time, an ongoing process - not a document to be created once. It is crucial that each clinical claim is individually supported by evidence. The effort required for the review and preparation of the CER is often underestimated. Early resource and time planning, a documented methodology, and a robust PMCF plan reduce the risk of additional requests and delays in the certification process.

We will review whether your clinical evaluation and claims meet the requirements of the Notified Body and assist you with planning, literature research, and CER preparation. Please feel free to contact us.