Every medical device sold in the European Union requires a clinical evaluation. Not as a formality. As the primary mechanism through which manufacturers demonstrate that their device is safe, performs as intended, and generates clinical benefit that outweighs risk.
Under EU MDR (Regulation (EU) 2017/745), that demonstration must be systematic, documented, and updated over the full lifecycle of the device. The process that governs how you do it is described in MEDDEV 2.7.1 Revision 4, the guidance document issued by the European Commission that remains the definitive reference for clinical evaluators working under MDR.
This guide walks through the five stages of clinical evaluation as defined in MEDDEV 2.7.1 Rev 4: scope definition, data identification, data appraisal, clinical data analysis, and writing the clinical evaluation report (CER). It also covers what those stages require in practice, where manufacturers most often fall short, and how to manage the process without losing traceability.
Table of contents |
The five stages of clinical evaluation under MEDDEV 2.7.1 Rev 4 |
What is clinical evaluation for medical devices?
Clinical evaluation is a systematic, planned, and documented process for collecting, appraising, and analyzing clinical data relating to a medical device. Its purpose is to verify that the device achieves its intended clinical benefit, performs as the manufacturer claims, and does not pose unacceptable risks to patients or users.
Under Article 61 of EU MDR, every medical device requires a clinical evaluation regardless of risk classification. Class I devices are not exempt. The evaluation must be proportionate to the nature of the device and the clinical risks involved, but it cannot be skipped.
Two terms are frequently confused here. A clinical evaluation is not the same as a clinical investigation. Clinical evaluation is the overarching process of gathering and assessing all relevant clinical evidence. A clinical investigation is one specific method of generating that evidence, typically a prospective study conducted on human subjects. You may need a clinical investigation to support your clinical evaluation, but an investigation is not mandatory in every case.
The clinical evaluation must be carried out by qualified evaluators with the relevant medical, scientific, or clinical expertise. Their qualifications must be documented. MEDDEV 2.7.1 Rev 4 states that teams are generally preferable to individual evaluators given the breadth of expertise required.
The five stages of clinical evaluation under MEDDEV 2.7.1 Rev 4
MEDDEV 2.7.1 Revision 4 defines four numbered stages plus a fifth stage that it treats as a distinct output: writing the Clinical Evaluation Report (CER).
Stage 0: Scope. Define the device, its intended purpose, indications, target populations, and the technology it is based on. Identify the clinical claims you intend to support and the relevant general safety and performance requirements (GSPRs) under Annex I of EU MDR that the clinical evaluation must address.
Stage 1: Identification of pertinent data. Identify all sources of clinical data relevant to your device, including data generated by the device itself (direct data) and data from equivalent devices (indirect data). Establish a literature search protocol before any searching begins.
Stage 2: Appraisal of data. Assess the suitability, applicability, and scientific validity of each data source identified in Stage 1. Not every paper that appears in a literature search is usable. Suitability, applicability, population representativeness, and quality are the four criteria MEDDEV uses to filter results.
Stage 3: Analysis of clinical data. Analyze the appraised data against the device's claimed benefits, risks, and GSPRs. Determine whether the clinical evidence demonstrates the device's safety and performance, characterizes residual risks and uncertainties, and establishes the need for post-market clinical follow-up (PMCF).
Stage 4: CER. Document the entire evaluation in a Clinical Evaluation Report. The CER is both the output of the evaluation and a living document that must be updated over the device's lifecycle.
The clinical evaluation plan (CEP) governs how all five stages are carried out. MEDDEV 2.7.1 Rev 4 and Annex XIV of EU MDR both require a CEP before evaluation work begins. The CEP should describe the scope, evaluation methods, data sources, appraisal criteria, and the composition of the evaluator team.
One of the most common and costly planning mistakes is writing the CEP after design validation is complete. If the plan is not in place before verification and validation begins, you will often discover that your testing does not align with what the evaluation requires, and that means rework. Starting the CEP early is one of the most cost-effective decisions a manufacturer can make.
Establishing equivalence
Equivalence allows manufacturers to use clinical data from a device already on the market to support the clinical evaluation of their own device. For many Class IIa and Class IIb devices, equivalence is the primary route to satisfying clinical evidence requirements without conducting a new clinical investigation.
Under EU MDR, equivalence requires simultaneous demonstration across three dimensions: clinical, technical, and biological. The criteria are stricter than they were under the Medical Device Directive (MDD). In particular, MDR requires that there be no clinically significant difference in the performance and safety of the devices being compared, not just similarity.
For Class III devices and implantable Class IIb devices, MDR Article 61(4) and Annex XIV Section 3 require that the manufacturer have sufficient levels of access to the data from the equivalent device. In practice this often means a contract giving full access to the equivalent device's technical documentation. MDCG 2023-7 (December 2023) clarified the boundaries of this requirement: it defines specific exemption cases under Articles 61(4) to 61(6) where data from another manufacturer's device can be used without a full contract in certain circumstances, including legacy devices under Article 61(6)(a) and well-established technologies (WET) listed under Article 61(6)(b). Manufacturers working with Class III or implantable devices should read MDCG 2023-7 alongside MDCG 2020-5 to understand which pathway applies to their device.
The criteria for equivalence are set out in Annex I of MEDDEV 2.7.1 Rev 4. MDCG 2020-5 addresses specific areas where MDR and MEDDEV 2.7.1 Rev 4 diverge on equivalence requirements. Teams claiming equivalence should read MDCG 2020-5 alongside MEDDEV.
Three equivalence failures come up most often in notified body review. The first is lack of data access: equivalence is claimed to a competitor device but the manufacturer cannot demonstrate actual access to that device's technical and biological data. For Class III implantable devices where equivalence is claimed to a different manufacturer's device, a contract granting explicit data access is required unless one of the MDCG 2023-7 exemption cases applies. The second is superficial comparison, where visual similarity between devices is presented without specification-level analysis. Technical equivalence requires the same design principles, materials, specification ranges, and similar manufacturing processes compared point by point, not a shape overlay. The third is indication mismatch, where the intended use is close but not identical, a common example being an adult device applied to a pediatric population.
If you cannot establish equivalence and need to generate direct clinical data, a clinical investigation plan (CIP) is required before any study begins. The CIP describes the study design, objectives, endpoints, subject population, and risk mitigation measures, and it must comply with ISO 14155 and MDCG 2024-3, which provides detailed guidance on CIP content requirements under MDR Annex XV.
Literature review and data identification
The literature review is the primary method for identifying indirect clinical data. A literature review for clinical evaluation is not the same as a general background search. MEDDEV 2.7.1 Rev 4 is explicit that the review must be systematic, meaning it follows a pre-specified protocol, uses defined search terms across identified databases, and documents inclusion and exclusion criteria before any results are reviewed.
The protocol must be written before the search is conducted. A protocol written after the fact to justify data already selected does not satisfy the MEDDEV or MDR requirements. Notified bodies scrutinize literature search protocols closely during technical documentation review, and a poorly documented or post-hoc protocol is a common finding in audit reports.
Data sources typically include published literature (via databases such as PubMed, EMBASE, and CINAHL), clinical registries, incident reports, vigilance data, and data from equivalent devices in post-market surveillance. Direct data from the device under evaluation, including premarket clinical investigations and real-world data from post-market follow-up, should also be incorporated where available.
Once the search is complete, the appraisal process filters the results. MEDDEV recommends applying four criteria to each source: suitability (does it address a relevant question), applicability (does it relate to the device under evaluation or an equivalent), population (is the population representative of the intended patient population), and quality (does it meet scientific validity standards). Appendix III of MDCG 2020-6 provides a suggested hierarchy of clinical evidence that should guide how different data types are weighted.
A common pitfall at this stage is under-documenting the appraisal rationale. Each included or excluded source should have a documented reason. Reviewers should be prepared to defend inclusion decisions, particularly for literature from equivalent devices where the basis for equivalence has not yet been formally established.
Clinical data appraisal and analysis
Analysis takes the appraised dataset and asks whether it is sufficient to demonstrate that the device meets its intended clinical claims, satisfies the relevant GSPRs, and presents an acceptable benefit-risk profile.
The connection between clinical data analysis and risk management is direct and mandatory. ISO 14971 risk management and the clinical evaluation are not separate workstreams; the outputs of each inform the other. Clinical data from post-market surveillance should be fed back into the risk management file to update benefit-risk characterizations over the device lifecycle.
This connection runs deeper than it first appears. Your risk management process identifies hazards and assigns severity and probability to risks. Clinical evaluation is how you validate those assignments. Did the risks you anticipated appear? At the frequency you expected? Were there risks you did not anticipate? Without clinical data feeding back into your risk assessment, you are working from assumptions that have never been tested in practice. That is why the CER update cycle and the risk management file review cycle need to be synchronized in your QMS, not managed as separate processes that occasionally reference each other.
MEDDEV 2.7.1 Rev 4 describes four elements of the analysis phase: using sound, scientifically valid methods; making a comprehensive analysis that addresses all relevant aspects of safety and performance; determining whether additional clinical investigations are needed to resolve gaps or uncertainties; and establishing PMCF needs based on residual risks and unanswered questions from the evaluation.
The analysis should address each clinical claim individually. Generic statements that the device is safe and performs as intended are not sufficient. Notified bodies expect evaluators to demonstrate how the available evidence supports each specific claim, and to explain how any gaps or uncertainties have been addressed or will be addressed through ongoing PMCF activities.
For EU MDR clinical evaluation, the analysis also feeds directly into the Summary of Safety and Clinical Performance (SSCP) that is required for Class III and implantable Class IIb devices. The SSCP is a public-facing document, so the quality of the underlying clinical evaluation analysis determines the quality of what is ultimately visible to regulators, healthcare professionals, and patients.
.png)
The clinical evaluation report (CER)
The CER is the document that captures and presents the entire clinical evaluation. It is not a summary. It is the record that your Notified Body will review during conformity assessment, and it must be complete enough for an independent evaluator to understand the rationale behind every conclusion.
Article 61 of EU MDR requires a CER for every medical device. For guidance on how to structure and write the CER, MEDDEV 2.7.1 Rev 4 remains the primary reference, supplemented by MDCG guidance documents that address specific aspects of MDR compliance.
The CER should include: the scope and objectives of the evaluation; a description of the device and its intended purpose; the evaluation team's qualifications; the clinical evaluation plan; a description of the literature search methodology and databases searched; the appraisal results for included and excluded sources; the clinical data analysis; the benefit-risk conclusion; identification of residual risks and uncertainties; and the PMCF plan or justification for why PMCF is not required.
How often must the CER be updated?
The CER is a living document. For Class III devices and Class IIb implantables, it must be updated at least annually. For other device classes, the update frequency must be justified based on the risk classification, the maturity of the technology, and any new information that emerges from post-market surveillance or vigilance reporting.
The update obligation is not a formality. Notified bodies check that the clinical evidence base has been reviewed against recent literature and post-market data. A CER with a literature search cutoff more than one or two years old, without documented justification for why an update was not required, is a common nonconformity finding.
The relationship between design validation and clinical evaluation is also worth understanding when planning CER update cycles. Changes to device design or intended purpose trigger a reassessment of the clinical evaluation, not just the verification and validation activities.
Post-market clinical follow-up and the living CER
Post-market clinical follow-up (PMCF) is not optional for most devices under EU MDR. It is the mechanism by which manufacturers generate the ongoing clinical evidence needed to update the CER and maintain the benefit-risk characterization over the device's commercial life.
The PMCF plan must be included in the clinical evaluation documentation before the device reaches the market. The plan describes the methods and activities the manufacturer will use to proactively collect clinical data after launch: registries, surveys, literature monitoring, sponsored studies, or real-world evidence programs.
PMCF data flows back into the clinical evaluation through the post-market surveillance (PMS) system. The PMS system aggregates data from multiple sources including complaints, vigilance reports, literature, and PMCF activities, and the clinical evaluator uses that aggregate output to determine whether the CER conclusions remain valid or require updating.
For manufacturers who entered the market under the Medical Device Directive (MDD) and are transitioning to MDR, the PMCF requirements represent a material increase in documentation and evidence obligations. Devices that were previously covered by a literature review and a light-touch CER now require active evidence generation plans with defined timelines, methods, and endpoints. The MDR transitional period was extended by Regulation (EU) 2023/607, with end dates of December 31, 2027 or December 31, 2028 depending on device risk class and subject to applicable conditions. Manufacturers still under MDD certification should factor these deadlines into their PMCF planning, since meeting the MDR clinical evidence standard at the point of transition requires a clinical evidence base built over years, not months.
Managing the clinical evaluation process
Clinical evaluation generates a large body of documentation: the CEP, literature search protocols, appraisal records, analysis worksheets, the CER itself, and the PMCF plan. Every document must be version-controlled, traceable to the device it supports, and accessible to reviewers during conformity assessment and post-market inspections.
The connection between the clinical evaluation and other quality system elements, including risk management, design controls, and post-market surveillance, means that changes in any one area can trigger an obligation to update the others. Managing those interdependencies manually across disconnected file stores is the single most common source of clinical evaluation nonconformities. Manufacturers working in spreadsheets or shared drives regularly find that their literature appraisal records do not reflect the current device description, or that the CER references a risk management version that has since been superseded.
If your CER currently gets updated when a submission deadline forces it, you are not running a clinical evaluation program. You are running a documentation catch-up cycle, and the compounding cost of that tends to show up at the worst possible time. The CER update schedule needs to be a standing event in your QMS, the risk management file should be reviewed every time your clinical evidence changes, and your labeling and intended use documentation should reflect what your clinical evaluation actually concludes. None of that happens automatically. It requires deliberate process design built into your QMS from the start.
Greenlight Guru Clinical is purpose-built for manufacturers who need to manage clinical data as part of their device lifecycle, not as a separate compliance function. It connects PMCF data collection, literature management, and clinical evidence directly to the device technical file, surfaces update obligations when connected documents change, and maintains an audit trail that satisfies both the documentation requirements of MEDDEV and the traceability expectations of EU MDR. If your clinical evaluation documentation lives in disconnected files with no automated linkage to your quality system, the cost of each CER update cycle will compound over time.
If you want to see how Greenlight Guru supports the full clinical evaluation lifecycle from CEP to CER to PMCF, get a demo of Greenlight Guru Clinical.
Keep reading:
free eBook:
Ultimate Guide to Clinical Evaluation of a Medical Device in the EU
Ultimate Guide to Clinical Evaluation of a Medical Device in the EU
%20The%20Ultimate%20Guide%20to%20Clinical%20Evaluation%20of%20a%20Medical%20Device.png?width=180&name=(cover)%20The%20Ultimate%20Guide%20to%20Clinical%20Evaluation%20of%20a%20Medical%20Device.png)
%20The%20Ultimate%20Guide%20to%20Clinical%20Evaluation%20of%20a%20Medical%20Device-1.png?width=250&name=(cover)%20The%20Ultimate%20Guide%20to%20Clinical%20Evaluation%20of%20a%20Medical%20Device-1.png)
