The Clinical Evaluation Report (CER) is where a medical device's clinical story either holds together or falls apart in front of a notified body. Under the EU Medical Device Regulation (MDR 2017/745), clinical evaluation is a continuous, documented process required by Article 61 and Annex XIV, and MEDDEV 2.7/1 Rev 4 remains the working methodology reviewers expect to see followed. Most CER findings are not about missing data. They are about method: searches that cannot be reproduced, appraisal that cannot be traced, and conclusions that outrun the evidence.
Start with the Clinical Evaluation Plan, not the report
MDR Annex XIV Part A requires a Clinical Evaluation Plan (CEP) before the evaluation itself, and notified bodies read the CEP first. The plan defines the device and its intended purpose, the General Safety and Performance Requirements (GSPRs) that need clinical evidence, the clinical benefits and their measurable outcome parameters, the state of the art in the clinical field, and the methods for identifying, appraising, and analyzing data. A CER written without a real CEP reads as a justification assembled after the fact, and reviewers recognize it immediately.
The MEDDEV 2.7/1 Rev 4 stages
Stage 1: Identification of pertinent data
Clinical data comes from three directions: data generated by the manufacturer (clinical investigations, post-market clinical follow-up, complaints and vigilance), data from the scientific literature on the device itself, and data on equivalent devices where equivalence is claimed. The literature search needs a documented protocol: databases searched, search strings, dates, inclusion and exclusion criteria, and a PRISMA style flow of what was found and excluded. The most common notified body finding in this stage is a search that cannot be re-run to produce the same results.
Stage 2: Appraisal
Each data source is appraised for methodological quality, relevance to the device and intended purpose, and weighting. A published appraisal method with explicit criteria, applied consistently and recorded in tables, is the difference between an appraisal and an opinion. Suitability criteria should address whether the data concern the actual device or an equivalent, the population match, and the outcome relevance.
Stage 3: Analysis
The appraised evidence is analyzed against the requirements set in the CEP: does it demonstrate conformity with the relevant GSPRs for safety and performance, does it support every claim in the labeling and instructions for use, and does the benefit risk profile remain acceptable against the state of the art? Any gap between what the evidence shows and what the intended purpose claims must be identified honestly here, with a plan to close it, typically through post-market clinical follow-up (PMCF).
Stage 4: The report
The CER documents all of the above with the qualifications of the evaluators, described against the MEDDEV expectation of relevant degree level scientific or clinical expertise and documented experience. Conclusions must be specific: which GSPRs are met by which evidence, what residual risks remain, and what the PMCF plan will address.
Equivalence under MDR: narrower than it used to be
Claiming equivalence to another device requires demonstrating technical, biological, and clinical equivalence to a level that MDR made deliberately harder than the directive era. All three dimensions must hold for the same device, differences must be shown not to affect clinical safety and performance, and for implantable and Class III devices Article 61(5) additionally requires a contract giving access to the equivalent device's technical documentation on an ongoing basis, unless the devices are from the same manufacturer. Equivalence built on marketing brochures and a comparison table stopped working when MDR arrived; if your clinical strategy depends on it, the access question decides feasibility before any writing starts.
State of the art: the section teams underestimate
State of the art in MDR usage means current, generally accepted good practice in the clinical field: the available treatment options, their outcomes, applicable standards and guidance, and where the device sits among them. It is established from clinical guidelines, systematic reviews, and comparative data, and it sets the benchmark the device's benefit risk profile is judged against. A CER whose state of the art section is a paragraph of background text gives the notified body no benchmark, and reviewers respond by questioning the benefit risk conclusion itself.
PMCF: where the CER's honesty becomes visible
Post-market clinical follow-up is the mechanism MDR provides for closing the evidence gaps the clinical evaluation identifies, and notified bodies read the CER and the PMCF plan as one document in two parts. A credible PMCF plan names the specific residual questions (long term performance, rare complications, performance in subpopulations underrepresented in premarket data, off label use signals), matches each to a proportionate method (registries, surveys of users, focused clinical investigations, structured literature surveillance), and sets timelines for feeding results back into the CER, the risk management file, and for Class III and implantables the Summary of Safety and Clinical Performance. A PMCF plan that could apply to any device in the catalogue tells the reviewer that the gap analysis in the CER was not sincere.
Practical notes on the literature search itself
Three habits prevent most search findings. First, search more than one database: MEDLINE alone misses device relevant engineering and European literature that Embase and specialty sources capture, and the protocol should say why each source was chosen. Second, search for safety, not only performance: adverse event terms, vigilance databases such as MAUDE, and field safety notices belong in the identification stage because reviewers will check whether known issues with the device type appear in your appraisal. Third, keep the audit trail: export the raw results, record the deduplication, and file the excluded studies with reasons. When a notified body reruns your search string, and they do, the numbers need to reconcile.
The findings that stall CE marking
- Literature searches without a reproducible protocol or with unexplained exclusions.
- Claims in the IFU and marketing materials that the clinical evidence never addresses.
- Equivalence claims missing one of the three dimensions, or missing the Article 61(5) access contract.
- PMCF treated as a formality, with no connection between identified evidence gaps and the PMCF plan.
- Evaluator qualifications undocumented, or a single author where the required expertise plainly spans several disciplines.
- A CER frozen in time: MDR expects the clinical evaluation to be updated throughout the device lifetime, with the CER, risk management file, and post-market surveillance feeding each other.
How EvySaif writes CERs that pass
Our clinician led team plans and writes CERs to MDR Annex XIV and MEDDEV 2.7/1 Rev 4, from CEP through reproducible literature searches, documented appraisal, GSPR mapped analysis, and PMCF planning, with the same discipline carried into IVD performance evaluations under IVDR. If your CER is due for its next update, or your notified body has raised findings, see our clinical evaluation services, our post-market surveillance support, or talk to a clinician about your file.