Comprehensive Guide to MDR Clinical Evaluation & Clinical Evidence Requirements
16 min read

Introduction

The EU Medical Device Regulation (EU MDR) has shifted MDR clinical evaluation from a largely document-driven process to a dynamic, evidence-based approach that extends across the entire lifecycle of a medical device. Under the former Medical Device Directive (MDD), many manufacturers relied heavily on literature summaries, broad equivalence arguments, and static Clinical Evaluation Reports (CERs) MDR documents that might not be revisited for years. MDR has decisively changed that reality.

Today, clinical evaluation under MDR must be planned, systematic, transparent, and continuously updated. It must demonstrate that a device’s clinical performance and safety remain acceptable when compared with the current state of the art, and that clinical benefits are supported by data rather than assumptions. This philosophy is clearly articulated in the European Commission’s overview of the new medical devices regulations, which emphasizes patient safety, traceability, and robust post-market surveillance as key policy drivers.

MDR’s Evidence Paradigm: From “Having a CER” to “Defensible Clinical Evidence”

EU MDR 2017/745 replaced the MDD to address several longstanding concerns: device-related safety incidents, inconsistent evidence standards, and insufficient post-market vigilance. Clinical evaluation lies at the heart of this reform. Article 61 and Annex XIV make it clear that all devices requiring a conformity assessment must undergo clinical evaluation, and that this process must be systematic, planned, and updated throughout the device lifecycle.

MDR requires manufacturers to demonstrate:

  • That the device achieves its intended clinical performance in the real world.
  • That clinical benefits claimed for the device are supported by data.
  • That risks are reduced as far as possible and are acceptable when balanced with clinical benefits.
  • That evidence remains sufficient and current as scientific knowledge, competing technologies, and clinical practice evolve.

The consequence is that a medical device CER can no longer be a static technical file deliverable. It is now an evidence narrative, embedded within a broader clinical and performance evaluation, that connects the device’s intended purpose, the state of the art, clinical data, post-market information, and risk–benefit conclusions into a coherent whole.

Regulatory bodies and Notified Bodies have strengthened expectations for clinical evaluation reports for medical devices accordingly. Notified Body perspectives emphasize methodological transparency in the review of the MDR CER, evidence quality, and clear justification of “sufficient clinical evidence” for each indication and claim.

Regulatory Intent: Why MDR Demands Stronger Clinical Evidence

To understand how to construct defensible CERs, it is helpful to understand why MDR looks the way it does. Regulators observed that, under the MDD, clinical evidence was often:

  • Fragmented across documents,
  • Overly reliant on literature without a strong appraisal,
  • Based on weak or unverified equivalence to other devices, or
  • Outdated relative to advancements in science and technology.

A series of device-related safety events triggered greater scrutiny, leading to the recognition that clinical evidence for devices needed the same level of rigour that has long been expected for pharmaceuticals. MDR’s architects therefore set out to create a regime where:

  • Clinical evaluation is continuous, not episodic.
  • Evidence is proportionate to device risk but is always present.
  • Claims are specific and measurable, not vague or marketing-driven.
  • Post-market surveillance (PMS) and Post-Market Clinical Follow-up (PMCF) are integrated with CERs rather than separate processes.

This intent is reflected in key guidance documents endorsed by the Medical Device Coordination Group (MDCG). The MDCG endorsed guidance documents library includes MDCG 2020-5 on equivalence, MDCG 2020-6 on sufficient clinical evidence for legacy devices, and MDCG 2020-13 on clinical evaluation assessment, all of which translate MDR principles into practical expectations for manufacturers and Notified Bodies.

For manufacturers, this means that the focus should shift from “What is the minimum evidence required to get through an assessment?” to “What is the most robust clinical evidence story we can tell that accurately reflects device performance and safety over time?”

The Clinical Evaluation Lifecycle Under MDR

MDR envisions clinical evaluation as a lifecycle process, tightly integrated with PMS and PMCF. While Annex XIV describes discrete stages, in practice these stages form a loop that repeats throughout the device’s time on the market.

The lifecycle typically begins with a Clinical Evaluation Plan (CEP), which forms the foundation of the CEP process. The CEP sets out the intended purpose of the device, the clinical claims to be supported, the methodology for identifying and appraising clinical data, and the role of literature and post-market data in that process. A well-constructed CEP is not a template exercise; it should reflect the device’s risk class, technological characteristics, clinical context, and novelty.

From there, manufacturers proceed to systematic data identification. Under MDR, this is no longer a casual literature search; it is a structured exercise akin to a systematic review in academic medicine. The requirements described in MDCG 2020-13 and related clinical evaluation templates expect clear documentation of databases searched, search terms, date ranges, inclusion and exclusion criteria, and screening processes.

Identified data such as clinical studies, registries, PMS records, PMCF outputs, and other sources are then subjected to critical appraisal. Here, the manufacturer evaluates methodological quality, bias risk, relevance of populations and endpoints, and the consistency of results. The goal is not to show that all data are perfect, but to show that conclusions in the CER are transparent, balanced, and rooted in evidence, including discussion of limitations.

The resulting synthesis is documented in the Clinical Evaluation Report, which connects:

  • A description of the device and its intended clinical use,
  • An analysis of the state of the art,
  • A summary and appraisal of clinical data,
  • A risk–benefit assessment,
  • A discussion of evidence sufficiency in line with MDR and MDCG expectations, and
  • A description of PMS and PMCF activities that will address residual risks and uncertainties.

Throughout the device lifecycle, PMS and PMCF provide new information. Trend analyses, vigilance reports, registry data, user feedback, and PMCF studies help confirm whether the assumptions made at the time of initial evaluation still hold true. MDR expects manufacturers to feed this information back into the CER and risk management file, updating conclusions and, where necessary, updating the CEP and PMCF plans. This lifecycle view is a natural bridge to medical device lifecycle management, where evidence planning and governance are embedded into broader quality and regulatory systems.

Anatomy of a Defensible Clinical Evaluation Report

A CER that satisfies MDR and Notified Body expectations is more than a compilation of documents. It is a logical, defensible argument supported by data and grounded in recognized standards and guidance. White papers such as Clinical evaluation under EU MDR from BSI offer helpful insight into how Notified Bodies interpret requirements in practice.

A strong CER typically includes:

1. Device and Clinical Background

The CER begins by describing the device in enough depth that a technically literate reader can understand how it works, where it is used, and for whom it is intended. This section should explain the device’s mechanism of action, critical design features, intended anatomical and clinical use, and any meaningful variants or configurations. It should also provide a concise clinical background, explaining the medical condition or indication, existing treatment pathways, and where the device fits in those pathways.

This context is important because MDR expects claims to be evaluated in light of clinical reality. A diagnostic device used in a screening setting, for example, operates in a very different risk and evidence environment than a therapeutic implant used in high-risk surgery. These nuances should be explicitly acknowledged and carried through into later sections of the CER.

2. Clinical Claims and Endpoints

MDR is explicit that manufacturers may only make claims supported by evidence. This requires a shift from broad or marketing-style statements (“highly effective,” “improves outcomes”) to clear, measurable claims linked to specific endpoints, such as reduction in complication rates, improvement in diagnostic accuracy, or specific usability benefits.

Each device claim should clearly map to one or more clinical endpoints, and those endpoints must be traceable to the evidence presented in the CER. Notified Bodies consistently evaluate this alignment as part of their CER review, assessing whether the clinical data truly substantiates the claims made in labeling, IFUs, and promotional materials. Ensuring this traceability is a foundational expectation highlighted across Notified Body guidance and a critical element of a compliant, defensible CER.

3. State of the Art (SOTA)

The state of the art is not a rhetorical concept; it is the foundation of a defensible risk–benefit justification. It establishes the clinical and technological context by answering essential questions: What does good clinical care look like for the relevant condition or indication? Which devices, procedures, or therapies are currently used? And what are their known benefits, limitations, and safety considerations?

Constructing SOTA requires engagement with clinical practice guidelines, authoritative position statements, high-quality reviews, and key clinical trials. The MDCG endorsed guidance documents include materials that touch on SOTA concepts for clinical evaluation and PMCF, and professional guidelines from medical societies provide the clinical dimension.

A well-developed SOTA section establishes the benchmark against which a device must be assessed—demonstrating that its safety and performance are at least consistent with, and ideally superior to, what is considered current best practice. Our dedicated discussion on state of the art in CERs explores how to frame and document this benchmark effectively.

4. Literature Review and Clinical Data Appraisal

Literature-based evidence under MDR must be handled with academic-level discipline. MDCG 2020-13 and related clinical evaluation templates outline expectations that echo the structure of systematic reviews, such as: defined search strategies, clearly justified inclusion/exclusion criteria, and documented appraisal of methodological quality.

In practice, this means that the CER should not simply list studies, but should explain:

  • Why particular publications were considered relevant,
  • How they were appraised,
  • What limitations they have, and
  • How, taken together, they support or fail to support specific clinical claims.

This approach aligns naturally with structured literature search protocol and review methods. Data extraction tables, bias assessments, and structured summaries help transform literature from a narrative backdrop into a quantitative and qualitative evidentiary base.

5. Equivalence and Legacy Data

MDR dramatically tightens the use of equivalence. The MDCG 2020-5 guidance on clinical evaluation, equivalence and MDCG 2020-6 guidance on sufficient clinical evidence for legacy devices make it clear that equivalence is only acceptable if the manufacturer has access to detailed technical, biological, and clinical data for the other device.

For many manufacturers, especially those relying on competitor devices as alleged equivalents, this access simply does not exist. As a result, equivalence arguments that were once acceptable under the MDD are now frequently rejected. Legacy devices must often be supported instead by strengthened PMS and PMCF activities, new clinical data, or refined indications. These issues are explored more fully in Article 61 clinical evaluation.

6. Risk–Benefit Analysis

Risk–benefit analysis under MDR is expected to be data-driven, explicit, and iterative. It should draw on risk management documentation (typically aligned with ISO 14971), clinical data, SOTA benchmarking, and PMS/PMCF findings.

Rather than describing general statements such as “benefits outweigh the risks,” a robust analysis will explain why this is the case, referencing specific event rates, complication profiles, comparative performance against alternative therapies, and insights from post-market experience. If specific risks remain uncertain, the CER should explain how PMCF activities will address those uncertainties.

7. PMS and PMCF Integration

The CER is also the place where manufacturers show how post-market surveillance and PMCF activities support ongoing evidence generation. Templates such as MDCG 2020-7 and 2020-8, available via the MDCG guidance library, are particularly relevant here.

This is where the CER directly connects with MDR PSUR, PMS, and PMCF requirements, reinforcing the lifecycle-focused approach central to effective medical device lifecycle management.

Legacy Devices Versus New MDR Devices

MDR’s expectations differ in important ways for legacy devices and new MDR devices, but both must ultimately meet the same standard: demonstrable, sufficient clinical evidence.

For legacy devices, MDCG 2020-6 offers detailed guidance on what constitutes “sufficient” clinical evidence. It acknowledges that manufacturers may rely to some extent on historical data, but stresses the need for:

  • A current, MDR-compliant clinical evaluation,
  • Updated SOTA assessment,
  • Reinforced PMS and PMCF where gaps are identified, and
  • Alignment with MDR General Safety and Performance Requirements.

Legacy device CERs built under MDD are not grandfathered in; they must be refreshed to MDR expectations. Devices that used older equivalence-based approaches often face the greatest burden, as equivalence must now satisfy far stricter evidentiary criteria.

For new devices placed on the market directly under MDR, the expectation is that clinical evidence will be more explicitly device-specific. Clinical investigations may be required, especially for higher-risk, implantable, or novel technologies. Study design must meet standards for scientific quality and ethical conduct, reflecting the principles of ISO 14155 and aligning with the methodologies regulators expect in modern clinical research.

These strategic questions how much clinical data is needed, in what form, and over what timeframe, are explored in detail in Article 61 clinical evaluation, which focuses on sufficiency of evidence across both legacy and new devices.

Clinical Evaluation for SaMD and AI-Based Devices

Software as a Medical Device (SaMD) and AI-driven systems present a unique set of evidence challenges. Algorithms may be updated frequently, performance may depend on data quality and representativeness, and risks such as bias or model drift must be managed carefully.

The IMDRF document “Software as a Medical Device (SaMD): Clinical Evaluation” provides a widely used global framework for SaMD evidence, and regulators such as the FDA have adopted it in part through guidance converging around similar principles.

Under MDR, SaMD and AI-based devices must demonstrate:

  • Clinical validity: the relationship between the software’s output and the targeted clinical condition or outcome.
  • Clinical performance: evidence that the software achieves its intended purpose in the target population and use environment.
  • Clinical benefit: measurable impact on decisions, workflows, or outcomes relative to SOTA.

Evidence may take the form of diagnostic accuracy studies, reader studies, simulated environment testing, usability and human factors evaluations, and PMCF data from real-world use. Manufacturers must also show how algorithm updates are controlled, validated, and documented over time, and how cybersecurity and interoperability risks are mitigated.

These topics are examined more thoroughly in clinical evaluation for SaMD and AI-based devices, which focuses on translating MDR principles into algorithm-specific strategies.

Establishing and Using the State of the Art

The concept of state of the art is often misunderstood as a descriptive backdrop, but in MDR it plays a central evaluative role. It defines what is currently regarded as acceptable clinical performance and risk within a given indication.

Establishing SOTA typically involves reviewing:

  • Clinical practice guidelines from professional societies,
  • Landmark clinical trials and high-quality meta-analyses,
  • Authoritative health technology assessment reports, and
  • Regulatory evaluations and safety communications from competent authorities.

Resources such as the MDCG guidance pages and curated compilations like Medical Device Regulation (MDR) resources help contextualize regulatory expectations and harmonized standards, which in turn influence what regulators consider to be SOTA.

By clearly describing SOTA, a CER can:

  • Benchmark device performance and safety,
  • Justify benefit–risk conclusions,
  • Expose where the device is innovative or where evidence gaps remain, and
  • Guide PMCF study design to address unanswered questions.

Our focused discussion in state of the art in CERs explores how to construct this section so that it supports, rather than weakens, the CER’s overall argument.

Literature Review Excellence Under MDR

The quality of the literature review is often a determining factor in whether a CER is accepted at first review or sent back with major non-conformities. Regulators and Notified Bodies now expect literature methodologies to resemble academic systematic reviews in transparency and discipline.

That means starting with clearly formulated questions, often PICO-based (Population, Intervention, Comparator, Outcome), and then designing searches that can credibly capture the relevant evidence body. The MDCG endorsed guidance documents, especially MDCG 2020-13, provide insight into assessment expectations for clinical evaluation.

A strong literature review section will typically:

  • Explain the databases and sources used,
  • Describe the search strings and timeframes,
  • Present a high-level screening flow (similar to PRISMA diagrams in scientific publications),
  • Justify why certain data sets were prioritized, and
  • Show how study quality was assessed and integrated into the weight of evidence.

This is precisely where robust literature search protocol and review practices become invaluable. They not only support better CERs but also create a reusable framework that can be applied to multiple devices and updated efficiently.

Clinical Benefit, Performance, and Data Traceability

Under MDR, the notion of “sufficient clinical evidence” is tightly tied to the concept of traceability. Each claim must be traceable to the data that supports it, and each piece of data should be traceable through the CER back to its source.

A well-structured CER often uses tables, cross-references, and narrative explanations to show how:

  • Claims about effectiveness or performance map to specific endpoints,
  • Endpoints are measured in clinical investigations or real-world data, and
  • Those measurements align with SOTA expectations.

This transparent mapping strengthens the credibility of the clinical evaluation report for medical devices and ensures that the evidence package meets MDR scrutiny.

The MDCG 2020-6 guidance on sufficient clinical evidence gives concrete examples of how evidence sufficiency might be judged for legacy devices, but the conceptual framework applies more broadly: evidence must be relevant, robust, and coherent.

If evidence is incomplete, for example, if clinical trials were small, observational, or limited to certain subpopulations, the CER should acknowledge these limitations and show how PMCF activities or post-market data collection will mitigate them. This is the bridge to MDR PSUR, PMS, and PMCF requirements, where continuous evidence development is planned and executed throughout the device’s lifecycle.

PMCF and Lifecycle Evidence Generation

PMCF medical device activities are MDR’s way of ensuring that devices continue to perform as expected once widely used in real-world settings. PMCF is required unless a well-reasoned justification demonstrates that it is not necessary, a high bar for most devices, especially in moderate and higher risk classes.

PMCF can take many forms i.e. prospective studies, registry participation, targeted follow-up of specific patient groups, real-world performance analytics, or structured user surveys. The common thread is that PMCF must be purposeful and proportionate, designed to address specific residual risks, uncertainties, or evidence gaps identified in the CER or risk management documentation.

MDCG 2020-7 and 2020-8 provide templates for PMCF plans and evaluation reports, available via the MDCG guidance and templates page.

Information generated through PMCF feeds into:

  • Periodic Safety Update Reports (PSURs),
  • Risk–benefit reassessments,
  • Labelling and IFU updates where needed, and
  • Regular CER revisions.

This continuous loop is at the heart of MDR’s lifecycle vision and is central to the ideas discussed in medical device lifecycle management.

Conclusion

The transition to the EU MDR has fundamentally reshaped expectations around MDR clinical evaluation and the evidence manufacturers must generate to support their devices. What once could be approached as a periodic documentation exercise is now a continuous, analytically driven process that demands rigor, transparency, and scientific discipline. A defensible CER must reflect not only the device’s pre-market evidence but also its real-world performance, evolving risk profile, and alignment with the state of the art.

For organizations navigating this landscape, success depends on adopting an approach that treats clinical evidence as a living system. This means establishing structured methodologies for literature review, critical appraisal, benefit–risk reasoning, and lifecycle evidence management that extend far beyond regulatory minimums. It requires intentional integration of PMS and PMCF activities, clear articulation of clinical claims, openness about uncertainties, and proactive strategies for maintaining evidence sufficiency as technologies, clinical practices, and regulatory expectations evolve.

Ultimately, MDR’s clinical evidence framework is not merely a compliance requirement—it is an opportunity. Manufacturers who embrace this paradigm gain deeper visibility into device performance, strengthen patient safety outcomes, and build trust with regulators and clinicians. They also create more resilient product portfolios, supported by evidence ecosystems that can adapt to future scientific, technological, and regulatory shifts.

In a healthcare environment increasingly driven by data, transparency, and accountability, robust clinical evaluation is both a regulatory imperative and a strategic asset. By investing in sound clinical evidence methodologies and lifecycle-aligned processes, manufacturers can ensure that their devices not only meet MDR expectations today but remain safe, effective, and relevant in the years to come.

How Freyr Can Help

Freyr’s global regulatory experts support manufacturers in developing MDR-aligned Clinical Evaluation Reports (CERs), literature review frameworks, Clinical Evaluation Plans (CEPs), and lifecycle-based clinical evidence systems that stand up to Notified Body scrutiny. Additionally, Freyr helps manufacturers to build EU IVDR-ready Performance Evaluation Packages that can withstand scientific scrutiny and remain credible as the IVD device is used across diverse populations, clinical settings, and evolving standards of care.

If you need assistance strengthening your clinical evaluation approach, closing evidence gaps, or ensuring MDR-compliant CER submissions, speak to our experts. Our team is equipped to guide you through every stage of the clinical evidence lifecycle, from planning and data appraisal to robust documentation and post-market evidence generation.

Blog Tags

Frequently Asked Questions (FAQs)