Equivalence as a Basis for Using Clinical Data in MDR Conformity Assessment

So, How Does Clinical Evaluation Actually Work?

Clinical evaluation is part of the conformity assessment of a medical device under Regulation (EU) 2017/745 (MDR). Because of the importance of this area within the overall structure of the technical documentation, it very often constitutes a separate expert process. This is hardly surprising, because clinical evaluation brings together almost all of the most important elements of the documentation: data from the use of the device under real-world conditions, risk management, laboratory studies, literature, PMS and PMCF.

When I talk with Manufacturers about clinical data, what I hear most often is not so much the question “are clinical investigations needed?”, but rather: “do we really have to start from zero, given that similar devices have been used for many years?”. A new investigation means time, cost and patient involvement, and duplicating knowledge that has already been reliably gathered is not always scientifically or ethically justified. That is precisely why the MDR permits the use of data from an equivalent device. This is not a gap in the regulations, nor is it a shortcut. It is a fully legitimate approach recognised in Article 61(3)(a) and Annex XIV, Part A, Section 3, but it requires a robust link to be established between data from another device and the device under evaluation. Unlike the previously applicable MDD and AIMDD directives, the MDR has significantly tightened the conditions under which equivalence can be used as a basis for clinical evaluation.

Let us start with the fact that, in order to confirm the conformity of a medical device with the MDR, the Manufacturer must demonstrate that the device meets the General Safety and Performance Requirements (GSPR) set out in Annex I to the MDR. The first GSPR already states that the device must achieve the performance intended by the Manufacturer, be safe and effective, and that the risk associated with its use must be acceptable when weighed against the benefit to the patient, taking into account the current state of the art. Article 61 of the MDR generally addresses conformity with these requirements, setting out the general principles applicable to clinical evaluation. Article 61(1) states that confirmation of conformity with the relevant GSPR, the evaluation of undesirable side-effects, and the acceptability of the benefit-risk ratio must be based on clinical data providing sufficient clinical evidence. This raises two important questions: what exactly are clinical data, and what does SUFFICIENT mean?

As regards the first issue, it seems fairly simple. Clinical data are defined in Article 2(48) of the MDR. Put most simply, they are information concerning safety or performance that is generated from the use of a device. It must be remembered, however, that the MDR indicates specific sources of such data. They may come from clinical investigations of the device under evaluation; from clinical investigations or other studies reported in scientific literature concerning a device for which equivalence has been demonstrated; from reports published in peer-reviewed scientific literature on other clinical experience with either the device under evaluation or an equivalent device; and also from clinically relevant information from PMS, in particular from PMCF.

However, not every piece of information from the market automatically becomes strong clinical evidence. The number of devices sold, considered alongside the number of complaints or incidents, is of course needed, but the mere statement “we sold a million units and received three complaints” does not, in itself, say very much. One needs to know what those complaints concerned, their severity and causal relationship, how many patients actually used the device, over what period of time, and one must also take into account the fact that not all problems are reported to the Manufacturer. The absence of reported incidents for implants may reflect underreporting, a long latency of complications, or a lack of systematic follow-up, and therefore is not, in itself, positive evidence of clinical performance. Nevertheless, clinically relevant information from PMS may fall within the definition of clinical data. MDCG 2020-6 also notes that complaints and incident data are generally low-quality sources and, on their own, are usually not sufficient to confirm safety, performance and the clinical benefits achieved.

Even before the MDR era, PMCF was sometimes informally compared to a phase IV study, although the division into phases I–IV is terminology that primarily applies to studies of medicinal products. Today, PMCF covers a much broader spectrum of activities and is a continuous process that updates the clinical evaluation after a device has been placed on the market. Its basic purpose is the active collection and evaluation of data from the use of the Manufacturer’s own CE-marked device in accordance with its intended purpose. The PMCF plan also takes into account information on equivalent or similar devices if their data are to be further evaluated and presented in the PMCF report. It should be stressed here, however, that the data intended to confirm the continuing safety and performance of our device should come from the device under evaluation. Data concerning equivalent or similar devices may, on the other hand, be used, among other things, to update the SOTA and to assess safety signals. Monitoring a competing device does not thereby constitute direct PMCF for the device under evaluation (MDCG 2020-7). If equivalence has indeed been demonstrated, clinical data on the equivalent device obtained from the sources indicated in Article 2(48) may support the clinical evaluation report (CER) for our device.

As for SUFFICIENCY, matters are not so simple. There is neither a clear definition nor a single table showing that, for example, five case-series publications are sufficient for a class IIa device, while two clinical investigations are required for class III. The MDR defines clinical evidence in Article 2(51), while MDCG 2020-6 explains that the quantity and quality of the data, together with the results of their evaluation, should permit a qualified assessment of whether the device is safe and achieves the intended clinical benefit(s) when used as intended. Put more simply, what counts is the total weight of the clinical evidence presented in the context of the device under evaluation.

From my experience as a clinical expert in medical technologies, I can say today, with regard to the level of sufficiency: it all depends…

It depends on the type of device, its interaction with the body, its impact on the course of therapy or diagnosis, the type of claim made by the Manufacturer, the degree of innovativeness, and the possible consequences of malfunction. The class of the device itself is important, but it does not yet answer all of these questions.

If analysis of these issues shows a limited impact of the device on the course of therapy or diagnosis, limited interaction with the body, and limited possible consequences of malfunction, then the clinical evaluation may be based on data lower in the hierarchy of evidence, for example an observational study, registry data, a high-quality PMCF survey, or appropriately analysed PMS data. This does not, of course, mean that every survey or every case series will be sufficient. The methodology, group size, method of patient selection, endpoints, follow-up time and risk of bias still need to be assessed. The same approach will also apply to data collected for the claimed equivalent device.

For devices for which demonstrating conformity on the basis of clinical data is indeed not appropriate, Article 61(10) may be considered — but that is a topic for a separate chapter. What is important is simply not to treat that provision as a general exemption for every device that does not come into contact with the body or has only a limited impact on the course of therapy. Such an approach must follow from risk management, the nature of the device’s interaction with the body, the intended clinical performance, and the Manufacturer’s claims. Under MDCG 2020-6, it does not apply to implantable devices or class III devices, and for other devices it should be exceptional in nature.

If, on the other hand, we have a device that significantly affects the course of therapy, interacts strongly or over a long period with the body, and the possible harm to the patient is serious, the clinical data should be based on stronger evidence — often on properly designed clinical investigations concerning the device under evaluation or an equivalent device. What counts is the sum of the evidence gathered, but it must be remembered that many pieces of weak data do not automatically become strong evidence merely because there are many of them. It is also important not to confuse the source of the data with their quality. A literature publication may describe a large, randomised clinical investigation, while PMCF data for the device under evaluation may come merely from an uncontrolled survey in a small group. “Literature” cannot therefore automatically be ranked below “the Manufacturer’s own investigation”. Every data set must be assessed in terms of methodology, risk of bias, adequacy of the population, endpoints, follow-up time, and its relationship to the device under evaluation.

If we have weaker PMCF data for the device under evaluation but, at the same time, strong literature data for a genuinely equivalent device, the combined body of evidence may be sufficient. The condition, however, is that equivalence itself has been properly demonstrated and that the remaining gaps are minor and appropriately monitored.

It should also be remembered that the manner of selecting clinical evidence and assessing its sufficiency should always be related to the current state of the art (SOTA) in the relevant field of medicine. SOTA does not mean solely the most modern solution available anywhere in the world, but rather what is currently and generally recognised as good practice in medicine and technology. If the SOTA shows that, for a given technology, the standard is studies with a control group, objective endpoints and a longer follow-up period, then basing the evaluation solely on a small number of complaints and an uncontrolled survey will be difficult to consider sufficient — put plainly, such data may simply be insufficient to confirm the relevant GSPR. On the other hand, “literature” must not automatically be treated as a weak source. After all, it is precisely in the literature that large, randomised clinical investigations may be published. The source of the data and the quality of the data are not the same thing.

So, What About the Literature?

It is worth distinguishing between two functions of literature: the first concerns knowledge in the field of a given medical technology, and the second concerns data on the device under evaluation or an equivalent device. This does not mean that these always have to be two completely separate reviews, but one must know which question a given publication is answering.

Knowledge in the field of the technology (SOTA) requires a broad view of publications, guidelines, reports and current clinical practice. If, for example, we have a dressing and declare specific clinical indications, we need to describe the cases in which such a dressing is used, what treatment methods are currently available, what alternatives exist, e.g. pharmacological or surgical, what the guidelines of professional societies say, and what the clinical benefits and risks are for dressings based on the same technology. The assessment of a simple dressing for superficial breaches of the skin will look different from the assessment of a specialist dressing for diabetic wounds. Only once we have gathered this information can we assess whether there is a device on the market sufficiently similar to ours for its data to be relevant to our evaluation.

This approach applies to most types of devices: implants, substance-based devices, dental devices and active devices, as well as software. For software, the literature must additionally confirm the scientific validity of its medical application (as mentioned in MDCG 2020-1), i.e. the relationship between the input data, the output generated by the software, and the clinical condition to which that output relates.

It may turn out that there is a device on the market that is almost identical to ours. At this point, however, we need to pause for a moment, because a similar device is not the same as an equivalent device. A similar device is not yet an equivalent device.

This is one of the most important distinctions in this entire subject. “Similar” describes certain common features but does not reflect the specific characteristics of the devices. “Equivalent” is a conclusion arising from a formal comparison of all the required technical, biological and clinical characteristics indicated in Annex XIV, Part A, Section 3 of the MDR. Here we are already talking about identifying the specific properties of the device.

Data on similar devices are highly relevant. They help us understand the SOTA, identify hazards and possible adverse effects, establish clinically relevant endpoints, design PMCF or a clinical investigation, and set benchmarks against which our own device should be compared. In principle, however, they do not become clinical data pertaining to the device under evaluation within the meaning of Article 2(48) merely because the technology looks similar.

Nor is it always appropriate to force a conclusion of full equivalence. First, it is advisable to establish what clinical data on that device actually exist and how they will contribute to our CER. The mere absence of clinical data does not yet determine that the device cannot be equivalent — that is decided by a comparison of its characteristics — but it does mean that demonstrating equivalence may have limited value for the CER. If, on the other hand, we cannot demonstrate that all the equivalence criteria are met, it is more appropriate to treat the device as similar and use its data for the SOTA, risk management and PMS.

The situation is considerably better where published clinical investigations exist for the selected device. In that case, it is worth considering demonstrating equivalence and, if possible, incorporating those data into the clinical evaluation of our device. This can be done for both class IIa and class IIb devices and — subject to additional conditions — for an implantable device or a class III device. Clinical evaluation for higher-risk devices may indeed also be based on data from an equivalent device. A publication alone, however, is not enough to transfer the conclusions. First, it must be demonstrated that the investigated device is genuinely equivalent to the specific version of our device, and only then can the quality of both the favourable and unfavourable data be assessed.

A common oversimplification in this area is: “if we use another Manufacturer’s data, a contract is always required”. That is incorrect. Article 61(4)–(6) primarily describes the cases in which an implantable device or a class III device may be exempted from the default obligation to carry out a clinical investigation. These are not general conditions for using data on an equivalent device in every clinical evaluation.

Put simply:

- Article 61(4) covers, among other things, a new device designed by modifying a device already placed on the market by the same Manufacturer. Equivalence must be demonstrated and endorsed by the notified body, the clinical evaluation of the earlier device must provide sufficient evidence, and appropriate studies must be planned after the new device has been placed on the market.

- Article 61(5) concerns a special pathway for a device equivalent to a device of another Manufacturer, where the Manufacturer wishes to make use of the exemption from the obligation to carry out a clinical investigation. In that case, a contract is required to ensure full access to the technical documentation on an ongoing basis, the original clinical evaluation of the equivalent device must comply with the MDR, and the relevant conditions of Article 61(4) must also be met. This applies to implantable devices and class III devices where the device is new to the Manufacturer and the Manufacturer wishes to use this pathway instead of carrying out a clinical investigation before placing it on the market (see MDCG 2023-7).

- Article 61(6)(a) concerns devices lawfully placed on the market under the earlier directives, i.e. legacy devices. Their clinical evaluation must be based on sufficient clinical data and, where they exist, must comply with the relevant common specifications concerning clinical evaluation for that type of device.

- Article 61(6)(b) covers certain device types expressly listed in the MDR as well-established technologies. Here too, provided that there are sufficient clinical data and that the relevant common specifications are complied with, where available, there is no need to initiate a new clinical investigation.

At this point, the provisions are easy to confuse. Article 61(4) sets out the general obligation to carry out clinical investigations for implantable devices and class III devices, while the remainder of paragraph 4 and paragraphs 5 and 6 describe four separate cases in which exemption from that obligation is possible. These cases are considered separately: the pathway under Article 61(5) refers to the relevant conditions of paragraph 4, while the cases under paragraph 6 are independent of it. These are not general conditions stating when data on an equivalent device may be used, but rather when exemption from carrying out a new clinical investigation is possible.

A contract with the Manufacturer of the equivalent device is required in the specific situation described in Article 61(5): namely, a new implantable device or class III device, equivalent to a device of another Manufacturer, where the Manufacturer wishes to make use of the exemption from the obligation to carry out its own clinical investigation. The contract must ensure full access to the technical documentation on an ongoing basis, the original clinical evaluation of the equivalent device must comply with the MDR, and the demonstration of equivalence, together with the remaining conditions, must be endorsed by the notified body.

The situation of legacy devices under Article 61(6)(a), and of devices listed in Article 61(6)(b), is different. MDCG 2023-7 clearly explains that the contract requirement of Article 61(5) should not automatically be applied to these cases. However, if the Manufacturer bases the evaluation on data from an equivalent device, it must still demonstrate equivalence in accordance with Annex XIV and show sufficient levels of access to the data needed to compare the specific technical, biological and clinical characteristics of the devices. This does not always require access to the full technical documentation. If, however, publicly available information does not make it possible to establish the relevant characteristics of the specific model or generation of the device, equivalence may not be reliably demonstrable.

Legacy-device status itself also does not mean that the data are, by definition, sufficient. Data collected under the earlier directives may be valuable, but it must be checked whether they correspond to the current intended purpose, population and claims, and whether they meet the current quality requirements of the MDR.

If, in practice, the burden of proof for an implantable or class III device rests on the clinical data from the equivalent device, it is crucial that equivalence be demonstrated rigorously and be supported by sufficient levels of access to the data within the meaning of MDCG 2023-7. This does not always mean access to the full technical documentation — apart from the case under Article 61(5) — but the available information must allow a reliable comparison of all relevant technical, biological and clinical characteristics. A mere list of publications is not enough if we do not know the parameters of the specific model or generation of the device.

Equivalence also does not remove PMCF obligations. For a legacy device under Article 61(6)(a), there is no automatic rule that a new observational study must be completed before the MDR certificate is issued. At the stage of assessing the documentation, however, the clinical evidence must already be sufficient. If a significant gap exists that cannot be closed without new data, the notified body may require the missing data to be generated before certification — future PMCF cannot serve as a promise that replaces missing evidence. If, on the other hand, the body of evidence is sufficient and the remaining uncertainty can be safely monitored, a prospective PMCF activity may be planned for implementation after certification. It is worth remembering in this regard that a legacy device is already a device placed on the market under an earlier directive, so a study conducted before MDR certification may still qualify as PMCF.

For devices other than implantable devices and class III devices, there is likewise no general obligation to conclude a contract with the Manufacturer of the equivalent device. This does not mean, however, that the instructions for use (IFU) of a competing device will always be sufficient. The IFU may help in comparing the intended purpose, the user or the population, but it usually does not provide complete information on material composition, the manufacturing process, the properties of the finished device, or the specific software version. Comparative testing performed by the Manufacturer on samples, supplier data, standards, patents or other reliable information can fill these gaps. If, however, the relevant characteristics cannot be established, equivalence simply cannot be assumed.

So, How Do We Demonstrate Equivalence?

Annex XIV, Part A, Section 3 of the MDR provides the applicable framework. It indicates which characteristics are to be compared, and to what extent. Technical, biological and clinical characteristics must be compared. All three sets of criteria must be fulfilled, and differences may be accepted only where they do not result in a clinically significant difference in safety or clinical performance.

The technical characteristics to be compared include, among other things, the design; conditions of use; specifications and properties, including physicochemical properties such as intensity of energy, tensile strength, viscosity, surface characteristics, wavelength and software algorithms; deployment methods; principles of operation; and critical performance requirements (this should be treated as a non-exhaustive list). For the biological characteristics, the MDR is more explicit: materials or substances in contact with the body should be the same, be in contact with the same human tissues or body fluids for a similar kind and duration of contact, and exhibit similar release characteristics of substances, including degradation products and leachables. For the clinical characteristics, we compare the same clinical condition or purpose, including a similar severity and stage of disease; the same site in the body; a similar population, including age, anatomy and physiology; the same kind of user; and similar relevant critical performance in view of the expected clinical effect for a specific intended purpose.

All three categories must be satisfied. “Similar” here means similar enough that the differences do not give rise to a clinically significant difference in safety or clinical performance.

More than one equivalent device can be identified, but each of them must independently satisfy all the criteria for equivalence. One cannot create a “virtual equivalent” by taking technical properties from device A, material composition from device B, and clinical investigations from device C. The correct approach is to select device X from among those available on the market, assess whether its technical, biological and clinical characteristics can all be identified, identify the available clinical data for device X, and incorporate those data into the CER for the device under evaluation. Only such an approach allows equivalence to be addressed correctly in the clinical evaluation.

In order to conduct an equivalence assessment properly, it is also advisable to establish which characteristics actually affect the safety and clinical outcome of the device. This does not mean that the remaining criteria can be omitted. They must be analysed, and any lack of relevance must be clearly justified. A detailed comparison of the housing material of a device will not be meaningful if it does not come into contact with the patient and does not affect flow, dosing, ergonomics or performance. It is not enough, however, simply to write “no impact” — one must show why there is no such impact.

A Few Practical Examples

In the case of aqueous vaginal gels or lubricants, osmolality [mOsm/kg] may be significant. WHO/UNFPA materials refer to a value of 380 mOsm/kg or lower as desirable, and to a value below 1200 mOsm/kg as an interim procurement recommendation. These are, of course, not automatic equivalence criteria under the MDR, nor are they values that can be directly transferred to every vaginal gel, but they constitute an important point of reference for epithelial safety. A product with an osmolality of 300–400 mOsm/kg and a product exceeding 2000 mOsm/kg may have different biological effects. On the other hand, merely comparing a range of 800–1200 mOsm/kg with a range of 300–400 mOsm/kg does not automatically establish a lack of equivalence. Composition, pH, contact time, the results of irritation and cytotoxicity studies, and clinical data must all be assessed.

If a device analyses a thermal image of the skin, what may matter is not just the measurement range itself, but also accuracy, thermal sensitivity, detector resolution, sampling frequency, and the method of image processing and the decision thresholds. A difference in resolution will be significant when it changes the ability to detect a clinically important area or leads to a different decision.

For substance-based devices, the primary packaging may also be significant. Suppose one aqueous saline solution for the nose is packaged in a Type I glass bottle, and another in a medical-grade HDPE bottle. The mere difference in material need not exclude equivalence if the Manufacturer demonstrates comparable stability, protection against oxygen, light and moisture, no significant sorption, an acceptable extractables and leachables profile, and a comparable dose delivered by the applicator. For an olive-oil-based spray, the packaging difference may already carry greater significance owing to oxidation, gas permeation and possible interactions with the plastic. Packaging can therefore be neither automatically deemed irrelevant nor automatically deemed an obstacle. Its actual impact on the finished device over the entire shelf life must be assessed.

For higher-risk devices, manufacturing processes and their effect on the final properties of the device must also be taken into account. For example, for a bone plate or screw made of 316LVM stainless steel, different heat-treatment conditions, the degree of cold working, or the method of passivation may affect the microstructure, residual stresses, fatigue strength, corrosion resistance and the release of metal ions. This, in turn, may affect the risk of mechanical failure or a local tissue reaction. However, these differences should not be linked to osseointegration — here, the material and the properties of the implant surface itself will be more directly relevant. Osseointegration can be affected by different surface functionalisation, for example a titanium-oxide layer on one implant and a hydroxyapatite coating on another. Here, however, it must be clearly stated: since these are different materials in contact with tissue, such devices will, as a rule, not satisfy the biological equivalence criterion of Annex XIV. Even if the SOTA reveals publications showing similar clinical outcomes, this does not replace the requirement to use the same materials or substances. We may speak of similar devices, but not of full equivalence.

Relying on equivalence will not, however, always be possible. If we have an innovative device, a breakthrough technology, new materials, or a solution based on AI, and there is no sufficiently similar device on the market, it will not be possible to identify an equivalent device. In that case, clinical data for the device under evaluation, appropriate to the specific gap, will need to be generated.

For software, the question of equivalence is related, among other things, to algorithms. MDCG 2020-5 indicates that the functional principle of the algorithm, its intended purpose and clinical performance should be compared. It is not reasonable to require equivalence to be demonstrated at source-code level, provided the software has been developed and validated in accordance with the relevant principles.

In practice, for software, particularly software based on AI, one must compare much more than the mere declaration that both devices “use artificial intelligence”. What may matter are the input data, the sensors or cooperating hardware, the population, the method of data preparation, the logic and version of the model, the decision thresholds, the type of output, the role of the user, and the place of the software within the clinical pathway. For a proprietary algorithm, or one that changes over time, demonstrating equivalence will be very difficult, primarily because of limited access to this information. This does not mean that it is legally impossible, but in practice the Manufacturer may not have the data needed to carry out a reliable comparison.

For legacy devices, we can identify another specific situation in which the device belongs to a group of well-established technologies (WET). It is worth stating immediately, however, that the MDR does not provide for a separate certificate, nor for any formal “granting” of WET status. It is for the Manufacturer to justify that a given group of devices has a simple, common and stable design that has changed little, a well-known safety profile and clinical performance, an established place in the standard of care, and a long history on the market. These criteria are described in MDCG 2020-6.

For legacy devices belonging to WET, it is possible, in exceptional cases, to justify a lower level of evidence and to base the clinical evaluation on accumulated information from multiple sources. Data on similar devices may then support confirmation of conformity with the GSPR. It must be remembered, however, that MDCG 2020-6 clearly states that data on similar devices are not clinical data pertaining to the device under evaluation within the meaning of the MDR, and that relying on complaints and incident data alone is not sufficient.

A potential example of WET could be simple cotton-gauze dressings, provided that the Manufacturer genuinely demonstrates the stability of their design, technology and safety profile. Similarly, one might consider standard monopolar surgical electrodes with an active tip or a wire made of stainless steel or tungsten. This does not mean, however, that every electrode or every gauze will automatically be WET — material, geometry, the method of energy delivery, composition, the manufacturing process and the intended purpose may still materially change the clinical profile.

In endodontics, an approximately 17% aqueous EDTA solution has been used for many years for chelating and removing the smear layer, together with sodium hypochlorite solutions, usually at concentrations of around 0.5–5.25%, for irrigation, disinfection and dissolving organic tissue. The mere fact that a given substance has been used for several decades is likewise not sufficient to classify every device containing it as WET. The full composition, concentration, form, site and duration of contact, method of application, mechanism of action, and the Manufacturer’s claims still need to be taken into account.

If a significant change is introduced to such a device, the existing data may cease to be sufficient. However, not every change automatically deprives a technology of its WET character, and not every change automatically requires a clinical investigation. If a Manufacturer changes the bleaching agent used in gauze manufacture from hydrogen peroxide to a different substance, it should primarily assess the impact of the process on residues, chemical characteristics, biological safety, and the performance properties of the finished device. If these issues can be reliably resolved by chemical, biological and performance testing, a new clinical investigation need not be necessary.

The situation may look different for the addition of silver-plated yarn or an active coating and the associated antimicrobial claim. Such a change may introduce a new mechanism of action, new exposure and new risks, thereby moving the device away from being a simple and stable technology. In that case, the continued applicability of the WET concept, the possibility of demonstrating equivalence with an existing device, and the need to generate clinical data for the device must be reassessed. Nevertheless, the type of new data should still correspond to the actual gap — a large clinical investigation will not always be the only solution.

In summary, the clinical evaluation of a medical device can be based on different approaches to demonstrating conformity with the GSPR, but each of them comes down to the same general principle: reliable evidence must be presented that the device achieves its intended performance, that its risks are acceptable, and that its clinical benefits outweigh the residual risks. It is not about demonstrating that the device poses no risk whatsoever, since in the case of medical devices this would often be impossible. It is about demonstrating that the risks have been reduced as far as possible, are known, and remain acceptable in relation to the benefits and the current state of the art.

Equivalence is not a way of avoiding clinical evaluation. It is a way of demonstrating that knowledge obtained for one device can genuinely be applied to another. The better we know both devices, and the stronger the justification for the absence of clinically significant differences, the more robust this link is. If, on the other hand, we lack information about the composition, the process, the specific version, the algorithm or the population, we should not try to paper over these gaps with the word “similar”.

I understand Manufacturers’ concerns about being required to carry out yet another clinical investigation, especially where the technology has been used for years and a great deal of data is available. Good clinical evaluation, however, consists neither in demanding an investigation merely as a box-ticking exercise nor in taking someone else’s data at face value. It consists in showing honestly:

1.      what we know directly about our device,

2.      what we know from data on a genuine equivalent device,

3.      what follows only from knowledge of a similar technology, and

4.      what data is still missing.

Only then can a proportionate way of closing the gap be chosen — sometimes this will be an analysis of existing data, sometimes a targeted PMCF activity, and sometimes indeed a new prospective clinical investigation.

A Manufacturer should not be forced to repeat work that genuinely relates to its device. Conversely, patient safety should not rest merely on the assumption that, because two devices look similar, they must behave in the same way. A well-conducted equivalence assessment is about identifying precisely that boundary.

Finally, I will outline the most common shortcomings that, from the perspective of a notified body, I observe in clinical documentation in the context of equivalence:

1.         Superficial justification of technical equivalence — limited to the statement “the same mechanism of action”, without a detailed comparison of design, materials and performance parameters.

2.         Insufficient access to the underlying data — relying on a competitor’s device on the basis of nothing more than incomplete, publicly available data.

3.         Underestimating differences in the patient population — overlooking subtle differences in indications, contraindications or population characteristics that are clinically significant.

4.         Treating equivalence as an end in itself rather than as an element of a broader clinical-evaluation strategy — even where the equivalent device’s data support the CER, the PMCF plan should address the remaining gaps and risks concerning the device under evaluation. For implantable devices and class III devices, explicit reference must additionally be made to the post-market studies referred to in Article 61(4) and MDCG 2020-7; for a legacy device, their scope and timetable should take into account the legal basis of the given pathway, the available evidence, the remaining uncertainties, and the existing PMCF obligations.

5.         Failure to update the equivalence analysis — a comparison carried out once, at the stage of initial certification, and not updated despite design changes to the equivalent device or the device under evaluation.

 

__________________________________________

 

Sources

1. Regulation (EU) 2017/745 of the European Parliament and of the Council, in particular Article 2(48) and (51), Article 61, Annex I and Annex XIV

2. MDCG 2020-5, Clinical Evaluation – Equivalence. A guide for manufacturers and notified bodies:

3. MDCG 2020-6, Clinical evidence needed for medical devices previously CE marked under Directives 93/42/EEC or 90/385/EEC:

4. MDCG 2023-7, Guidance on exemptions from the requirement to perform clinical investigations pursuant to Article 61(4)–(6) MDR and on sufficient levels of access to data needed to justify claims of equivalence:

5. MDCG 2020-7, Post-market clinical follow-up (PMCF) Plan Template:

6. MDCG 2025-10, Guidance on post-market surveillance of medical devices and in vitro diagnostic medical devices:

7. MDCG 2020-1, Guidance on Clinical Evaluation (MDR) / Performance Evaluation (IVDR) of Medical Device Software:

8. WHO/UNFPA, materials on the osmolality of lubricants - www.unfpa.org/sites/default/files/pub-pdf/UNFPA_Safe%20Lubricants%20For%20All%20%28Web%29%202022.pdf

www.who.int/docs/default-source/medicines/norms-and-standards/guidelines/trs1025/trs1025-annex11.pdf