4. The Clinical Evaluation Report (CER)

What is a Clinical Evaluation Report?

The Clinical Evaluation Report (CER) is the comprehensive document where you compile all the findings from the clinical evaluation and make the case that your device is safe and performs as intended. In essence, the CER is the output of executing the Clinical Evaluation Plan. It includes the data you gathered (e.g. literature, clinical study results, post-market data) and your analysis of those data, leading to conclusions about the device’s clinical safety and performance. The CER is a crucial part of the technical documentation for MDR compliance and is required for devices of all classes (Class I to III). Even low-risk devices must have a CER, though the depth of data may differ. Regulators view the CER as evidence that the manufacturer has systematically assessed all relevant clinical information and confirmed that the device meets the applicable GSPRs.

According to one definition, an effective CER “describes a structured appraisal and analysis of all available clinical evidence to assess the safety and performance of a medical device.”

It’s essentially a detailed scientific report. While the clinical evaluation is an ongoing process over the device’s life, the CER is a snapshot of that process at a given time (typically prepared for initial CE-marking and updated periodically or when new information dictates).

Regulatory significance: The CER directly supports the declaration of conformity for the device. Per MDR Annex XIV, “the results of the clinical evaluation shall be documented in a Clinical Evaluation Report which shall support the assessment of the conformity of the device.” Notified Bodies use the CER as a primary resource to decide whether a manufacturer has enough clinical evidence for the device’s intended use. If the CER is inadequate or shows gaps in evidence, the NB will issue nonconformities – which can delay or prevent certification. Therefore, producing a high-quality CER is not just an academic exercise; it is central to obtaining and maintaining market approval.

What information must be included in a CER:

Unlike the CEP, which is about planning, the CER is about execution and results. While MDR Annex XIV doesn’t spell out an exact table of contents, it implicitly requires the CER to cover all the steps and findings of the clinical evaluation. Based on MDR, MEDDEV guidance, and NB expectations, a typical CER should include:

• Device description and context: A brief description of the device, its intended purpose, classification, how it works, and what it’s used for. Include identifying information (model numbers, etc.) and a summary of its regulatory status (new device or legacy device with prior approvals). This section sets the stage for the reader.

• Scope of the CER: State which device and indications are covered by this evaluation. If the CER is covering a family of devices or variants, clarify that. Also mention the date of the literature search/data cutoff so it’s clear how current the evidence is.

• Clinical Evaluation Plan summary: It’s wise to summarize the key points of the CEP in the CER – such as the clinical questions you set out to answer, the criteria for data inclusion, and whether equivalence or new studies were part of the strategy. You can even reference the CEP document. This assures the reader that the evaluation was planned and not retrospective cherry-picking. For example, if you pre-defined an objective to show non-inferiority to standard of care, state that in the CER.

• State of the art: A dedicated section that provides the clinical background. This includes the nature of the condition the device addresses, current treatment methods, known risks/benefits of those alternatives, and any medical guidelines or consensus documents. It essentially answers: “What is the gold-standard or usual care for this medical issue, and what are its limitations?” This section should draw on literature and data (cited) to establish benchmarks. Importantly, do not just stop at a general background – make sure to highlight quantitative benchmarks (e.g. “Standard treatment X has a 80% success rate and 5% complication rate.”). NBs have noted that sometimes the state-of-the-art section is treated as a formality and not used in later analysis. A strong CER will later compare the device’s clinical data to these state-of-the-art benchmarks in the discussion and conclusions.

• Clinical data (literature and other sources): This is typically the largest section. It includes the results of your systematic literature search and any other data sources:

  • Literature review results: Describe the search strategy (databases, keywords, time frame) and the selection process (how many articles found, how many included/excluded and why). Present the relevant clinical studies or reports on the device (or equivalent device, if applicable). For each included study or data set, provide a summary of the methodology and key outcomes. Organize this logically, e.g. by outcome or by study type. It’s often helpful to tabulate the evidence. Also include any unfavorable or contradictory evidence (regulators want to see that you didn’t ignore negative results). If the device is new and no direct literature exists, you’ll be focusing on equivalent device literature or possibly analogous device data.
  • Clinical investigation data: If you performed clinical studies on the device, summarize their protocol (or reference the full clinical investigation report) and present the results and statistical analyses. Ensure to cover safety outcomes (adverse events) and performance outcomes.
  • Post-market data: For devices already on the market (legacy devices or those with some field experience), include an analysis of post-market surveillance (PMS) data. This could be complaint data, vigilance/adverse event reports, registry data, published real-world evidence, etc. MDR requires integrating post-market clinical data into the evaluation. For example, if the device was sold under the MDD, summarize the PMS and any Post-Market Clinical Follow-up (PMCF) results available. Even for new devices, you might include any usability studies or observational data from pilot programs here.

• Appraisal of data and analysis: After listing out the data, the CER must appraise it – essentially, assess the quality and relevance of each data set – and then analyze what it all means for the device. An appraisal subsection might discuss the level of evidence (e.g. randomized trial vs. case series), bias or limitations in the data, and how confident we can be in the results. For example, you might use a scoring system or grading (some manufacturers use adaptations of academic evidence grading to show they critically evaluated each study). Then, the analysis part synthesizes the data: Does the totality of clinical evidence demonstrate the device’s performance claims? Is the safety profile acceptable compared to alternatives? This is where equivalence data would be explicitly brought in, if used: you must show that the equivalent device’s data are relevant and applicable to your device (because you have sufficiently proven the equivalence). According to MDR, you must “appraise all relevant clinical data by evaluating their suitability for establishing the safety and performance of the device”. MDCG 2020-13 (the Clinical Evaluation Assessment Report template used by NBs) expects to see a clear, unbiased appraisal – so consider structuring this section clearly (perhaps with subheadings for “Safety Data Appraisal” and “Performance Data Appraisal”).

• Benefit-Risk analysis: This critical section explicitly weighs the device’s benefits against its risks, in light of the state of the art. Summarize the proven clinical benefits (with supporting data) and the known risks/side effects (with rates or frequencies observed). MDR requires that the CER include conclusions about the benefit-risk profile. The analysis should discuss whether the benefits outweigh the risks for the intended patient population and indications. If there are different indications or sub-populations, do a benefit-risk assessment for each. NBs often expect to see a table or clear narrative that, for each key benefit, lists relevant supporting data, and for each risk, notes its severity/frequency, culminating in a reasoned conclusion that overall the device achieves a positive benefit-risk balance. This section should also compare the device’s outcomes to the state-of-the-art: for instance, “Device A reduced pain scores by 50% whereas the standard treatment typically achieves ~30%; complication rates were similar to standard of care.” This explicitly addresses whether the device is on par with, or better than, existing options (or if not, why it’s still justified). If the device is essentially equivalent in performance to existing devices, you might justify it by other considerations (e.g. if it’s cheaper or easier to use – though those are usually not primary clinical considerations for regulators, it’s mainly about clinical performance and safety). Also include here any risk mitigations or warnings that will be communicated (linking to labeling if needed) to ensure risks are as low as possible.

• Conclusions: Finally, the CER should end with a conclusion section that clearly states whether the clinical evidence is sufficient to declare conformity with relevant MDR requirements, and summarize how the device’s safety and performance have been established. This should tie back to the GSPRs you listed in the CEP. It’s good practice to explicitly mention that the device meets the criteria of Annex I based on the clinical evidence presented. For example, you might conclude that the device achieves its intended clinical benefits, that the performance objectives set out were met, and that any risks are acceptable and comparable to state-of-the-art alternatives. If there are any uncertainties or gaps, mention how they will be addressed (often through PMCF). Also, confirm that a PMCF plan is in place if required, to collect longer-term or additional data. Essentially, this section is your final argument that the benefits outweigh the risks and the device is fit for the market. Make it succinct and strong. NBs will read this carefully to see if you’ve ticked all the boxes.

Paths for clinical evaluation:

Now, within a CER, the type of clinical evidence used can come via different pathways. MDR recognizes a few paths to demonstrating clinical evidence for a device:

• Equivalence pathway: This path involves demonstrating that your device is equivalent to another device that already has clinical data (and often, already CE marked). If equivalence is established, you can leverage the existing clinical data of that similar device to support your own. The MDR allows using equivalent device data, but the bar for equivalence is high. You must show equivalence in technical characteristics, biological characteristics, and clinical characteristics (per MDR Annex XIV, Part A, 3). MDCG 2020-5 guidance “Clinical Evaluation – Equivalence” goes into detail on this, highlighting differences from the older MEDDEV criteria. Under MDR, any differences between the devices must not be clinically significant, and you need to justify equivalence for each of the three aspects. For example, devices should be of similar design and use, made of similar materials (especially if in contact with the body), and have the same clinical purpose.

  One major challenge of the equivalence route is access to data. MDR Article 61(5) and MDCG guidance state that if you are claiming equivalence to a device not owned by your company (i.e. a competitor’s device), and your device is Class III or an implant, you must have a contract in place to obtain the competitor device’s full technical documentation and clinical data. This is rarely feasible in practice – most companies will not share proprietary data. As a result, demonstrating equivalence to a competitor’s high-risk device is nearly impossible unless there is some agreement (e.g. licensed technology or a partnership). For lower-risk devices, you may use literature on competitor devices without full tech file access, but you still need enough publicly available information to justify equivalence on all three characteristics. Because of these hurdles, many manufacturers opting for equivalence use their own predicate device (e.g. an earlier generation product) as the equivalent device, since they have full access to its data.

  Requirements and challenges: If you pursue equivalence, be prepared to provide a thorough equivalence demonstration in your CER. This usually includes a side-by-side comparison table of your device vs. the equivalent device, covering design, materials, principles of operation, indications, performance specifications, etc., with justifications that any differences do not adversely affect safety or performance. MDCG 2020-5 provides a template of sorts for how to structure such comparisons. All equivalence criteria from both MDR and MEDDEV 2.7/1 rev.4 need to be met. A common pitfall is to claim equivalence but provide insufficient evidence or reasoning – NBs will reject such attempts. For example, claiming equivalence to a device that has a different intended use or a different material without a strong justification would likely fail. Another challenge is that if the equivalent device is a competitor product, the NB might question how you obtained certain detailed information (lack of evidence could be a showstopper).

  Equivalence ≠ skipping clinical evaluation: It’s important to note (and MDCG 2020-5 emphasizes) that using an equivalence strategy does not eliminate the need to perform a full clinical evaluation of your device. Equivalence just means you are allowed to incorporate the other device’s clinical data into your analysis. You still must critically evaluate that data as if it were your own. You also need to address any gaps – for example, if some risks are specific to your device (maybe due to a design difference) that weren’t present in the equivalent device’s data, you may need additional evidence or testing for those. In many cases, Post-Market Clinical Follow-up (PMCF) studies are expected when equivalence is used, especially for high-risk devices. MDR Article 61(4) essentially mandates that if you didn’t do your own pre-market clinical study for a new Class III or implant device (because you used equivalence), you must conduct a PMCF study after CE mark to further confirm safety and performance.

  In summary, the equivalence pathway can be a viable route to gather clinical evidence without duplicating clinical trials, but it requires meticulous documentation and often additional post-market commitments. Manufacturers should follow MDCG 2020-5 guidance closely and be prepared for intense NB scrutiny on whether equivalence is truly justified.

• Clinical investigations (clinical trials): This is the most direct pathway: generating new clinical data through a clinical investigation (or more than one) on your device. Under MDR, conducting a clinical investigation is expected for many new devices, particularly those in higher risk classes. Article 61(4) of MDR states that for Class III and implantable devices, a clinical investigation shall be performed unless the device falls under certain exceptions. Those exceptions (in Article 61(6)) mainly cover devices that have been CE marked before (legacy devices) or that can claim equivalence to such devices, or are well-established technologies (WET) – we’ll discuss these in “Exceptions” below. For novel Class III/implant devices, you should plan a clinical study to collect data on safety and clinical performance, because literature alone is unlikely to suffice. Similarly, for innovative lower-risk devices where no relevant equivalent data exists, a targeted clinical study may be needed (even if not strictly required by 61(4)).

  When and why a clinical investigation is needed: Essentially, if you cannot fully satisfy the MDR’s requirement for “sufficient clinical evidence” using existing data, you need to do a clinical investigation. NBs will look for clinical investigations especially when a device involves new technology, new clinical claims, or high risk to patients. For example, a new implantable device for which only preclinical (lab and animal) testing exists must undergo a human clinical study to gather data on safety (e.g. complication rates) and performance (clinical outcomes). The purpose is to generate device-specific evidence of how it works in the intended patient population. Even if not mandated by class, any time there is a significant clinical question unanswered by literature (like “does the device actually improve outcome X in patients?”), a clinical investigation is the most robust way to answer it.

  Designing the clinical investigation: MDR has its own requirements for how clinical investigations are to be conducted (see MDR Article 62-80 and Annex XV for detailed provisions). Additionally, ISO 14155:2020 is the internationally recognized standard for GCP (Good Clinical Practice) in device trials, and compliance with it is effectively expected. When planning a clinical study, manufacturers should develop a Clinical Investigation Plan (CIP) that meets Annex XV and ISO 14155 guidelines. Key elements include:

  • Study design – clearly define if it’s randomized, controlled, blinded, etc., or an observational study, and why that design is appropriate.
  • Study endpoints – what clinical outcomes will be measured to demonstrate performance/benefit, and what safety endpoints will be monitored (e.g. adverse event rates). These should align with the device’s intended clinical benefits and the state of the art (for example, use endpoints common in the literature for that medical field).
  • Patient population – inclusion/exclusion criteria defining the target population. Ensure it matches the intended use (e.g. if device is intended for adults, don’t only study it in healthy volunteers or a different group). MDR also expects that subjects represent the European population if the device is for EU use. That could mean including European study sites or justifying that data from elsewhere (e.g. a US trial) is applicable to Europe.
  • Sample size – how many subjects will be enrolled and the statistical rationale for that number. An underpowered study can be a big problem, so statistical justification is critical.
  • Follow-up duration – the length of time patients will be followed to observe outcomes. This should correspond to the device’s risks and expected life; for instance, an implant might need 12-month or longer follow-up to capture healing and any complications, whereas a diagnostic test might only need short-term follow-up.
  • Study locations – where the study will be conducted. If outside the EU, you should justify that the clinical practice and population are relevant to the EU context.
  • Data analysis plan – statistical methods to be used, success criteria (e.g. hypothesis tests, non-inferiority margins, etc.), and handling of missing data.
  • Overall ethical and GCP compliance – statements on following ISO 14155, obtaining ethics committee approvals, informed consent, etc.

  In the CER, you will include the results of any such investigation. But for the purposes of MDR compliance, it’s important that the quality of the clinical investigation is high. NBs often check the CIP or summary of it. In fact, MDCG has been working on templates for CIP and Clinical Investigation reports to standardize this. As a manufacturer, demonstrating that your trial was well-designed and executed per ISO 14155 gives confidence in the data. MDCG 2020-6 (which deals with legacy devices’ clinical evidence) also reminds that new clinical data might be needed to meet MDR’s “sufficient clinical evidence” standard, and it provides a hierarchy of evidence to consider. If your evidence falls short, a new investigation might be top of that hierarchy to add.

  Using investigation data in the CER: Ensure that the CER includes a summary of the clinical investigation results and a critical appraisal of that data. Highlight how the study outcomes support the device’s safety/performance and clinical claims. If the study had any shortcomings (e.g. a missed endpoint or an adverse event occurred), discuss how that is addressed or mitigated. Remember, even with a clinical study, the MDR expects you to consider if the data proves the device’s benefits in the context of current therapy.

• Well-established technologies and other exceptions: The MDR provides some exceptions where extensive new clinical evidence might not be required, primarily for certain devices deemed well-established or for legacy devices that were already on the market under the old directives. Let’s break these down:

  • Legacy devices: These are devices that were CE marked under the previous directives (MDD/AIMDD) and are now transitioning to MDR. MDR Article 61(6)(a) basically says that if a device has been marketed under the old directives and there are no significant changes in its design or intended use, it may not require a new clinical investigation for MDR – provided that sufficient clinical data exist. However, manufacturers must still perform a clinical evaluation of the device against MDR’s requirements, using the data they have (e.g. years of post-market data, published studies, etc.). MDCG 2020-6 is dedicated to guiding such manufacturers on how to demonstrate sufficient clinical evidence for these legacy devices so they can meet MDR standards without starting from scratch. The key is that the evidence that supported the device under MDD needs to be reassessed: Is it still current? Does it cover all new MDR aspects (like clinical benefits, risk/benefit, etc.)? Often, legacy devices have a lot of post-market data that can be leveraged. The CER for a legacy device should focus on compiling the historical clinical data (pre-market studies, post-market surveillance, any literature) and showing that this body of evidence satisfies MDR’s criteria. If there are gaps (for instance, maybe under MDD the device never had a formal clinical study because it was low-risk, but MDR now expects more), the manufacturer might need to collect some new data or perform a literature search to update the state of the art. In any case, legacy devices aren’t exempt from clinical evaluation; they just might rely more on existing data, and MDCG 2020-6 helps interpret what is “sufficient” in that context.

  • Well-established technology (WET) devices: The MDR uses the term “well-established technologies” in Article 61(6)(b) as a subset of devices that might be exempt from the requirement of conducting clinical investigations, even if they are Class III or implantable. However, “well-established technology” is not explicitly defined in MDR – it’s generally interpreted to mean devices that have been used in clinical practice for many years with well-known safety and performance characteristics. Examples might include things like surgical sutures, hypodermic needles, or bone screws – devices that aren’t novel or high-risk in the sense that their clinical function is well understood and has been proven over time. According to MDCG 2020-6, the term WET is not strictly defined by MDR, but such devices are expected to have a long history of safe use and possibly a wealth of literature available. If your device qualifies as a WET, you may not need to perform a new clinical study if you can rely on clinical data from similar devices and other sources to demonstrate safety and performance. For instance, a manufacturer of a basic surgical instrument could gather published clinical data on that type of instrument (perhaps from many years of surgeries documented in literature) as evidence.

    However, being a well-established technology does not remove the need for a CER or clinical data. You still have to compile the evidence that shows the device conforms. The difference is that regulators acknowledge that for WET devices, a lower level of clinical evidence might be justified – meaning you might not need as large or as new a data set as you would for an innovative device. For example, bench testing and decades of literature could suffice to confirm a certain device’s safety, where a brand new device would require fresh clinical trials. But caution: manufacturers should not self-declare their device as “well-established” without solid reasoning. The NB will expect an explanation of why the device is considered WET (e.g. “Device X has been in clinical use since the 1980s with only incremental changes, and its clinical performance is documented in 50+ publications”). Even then, the clinical evaluation must be done and documented. Often, NBs will scrutinize WET device CERs to ensure that any claim of not doing new studies is backed up by strong existing evidence.

  • Other special cases: MDR Article 61(10) allows, in exceptional cases for devices other than Class III/implants, that equivalence might not even be needed if the manufacturer can justify that clinical data is not deemed appropriate (for example, purely physical principle devices where bench/testing yields all necessary info). However, this is rarely invoked and would require a very solid justification in the CER and risk management. Also, devices that incorporate medicines or biologics have their own additional evidence requirements (but those are beyond our scope here).

In the CER, if you are using any exception (legacy device using prior data, WET device using literature, etc.), be explicit about it and cite the MDR clause (e.g. “This device qualifies for the exception in Article 61(6)(b) as a well-established technology. Therefore, no new clinical investigation was conducted; instead, the clinical evidence consists of literature and PMS data.”). Then, prove that the evidence is sufficient. MDCG 2020-6 Appendix III provides a “hierarchy of clinical evidence” which can help in arguing sufficiency – e.g. high-quality randomized trials on similar devices might be nearly as good as having data on your own device. Always err on the side of more evidence and more analysis, because even legacy and WET device CERs under MDR have frequently been found lacking if they rely on outdated or inadequate data.

Relevant guidelines for CER preparation:

In addition to MDR and MDCG guidance, manufacturers still often refer to MEDDEV 2.7/1 rev.4 (the clinical evaluation guidance from under the old directives) for practical instruction on how to write a CER. MEDDEV rev.4 provides a recommended outline for a CER and methods for literature review and appraisal. Many of those principles remain valid under MDR (systematic approach, appraisal of data quality, etc.). However, keep in mind MEDDEV has not been updated for MDR, so where there are differences (for example, MEDDEV’s equivalence criteria were slightly looser than MDR’s), MDR requirements prevail. Another useful document is MDCG 2020-13, which is the Clinical Evaluation Assessment Report (CEAR) template used by NBs. This template basically lists all the points an NB auditor will check in your CER. Reviewing MDCG 2020-13 while writing your CER can be very helpful to ensure you didn’t miss anything expected. For instance, the CEAR template asks if the CER considered state-of-the-art, if it addressed each GSPR that needed clinical data, if PMS data were included, etc. Aligning your CER with these points can preempt many NB questions.

To summarize, the CER is where all your planning and data collection come together to demonstrate your device’s compliance from a clinical perspective. Whether through new clinical trials, equivalence to an existing product, or literature on a well-established technology (or most likely a combination of these), the CER must convincingly show that the device is at least as safe and effective as the current state of the art. It should be a stand-alone document that regulators and NBs can read and understand the device’s clinical profile without needing to ask for a lot of additional explanation. Clarity, completeness, and correctness (with evidence to back every claim) are the hallmarks of a good CER.