Drugdu.com expert's response:

EU MDR stipulates in Annex XIV that if a manufacturer wishes to cite the clinical data of a marketed device (predicate device) to support its own clinical evaluation, it must first prove equivalence between the two devices. MDR assesses this on a dimension-by-dimension basis across technical characteristics, biological characteristics, and clinical characteristics, and the overall requirements are significantly stricter than the old MDD regulation. The core principle is: the two devices must be "used under similar conditions", meaning there are no significant clinical differences in safety and clinical performance.

I. Technical Characteristics

MDR requires the manufacturer to prove that the two devices are highly similar in terms of technology, with the core being "used under similar conditions." The specific items to compare include:

Design and specification performance: The design principles and physical/chemical properties should be similar. Key performance parameters such as energy intensity, tensile strength, viscosity, surface characteristics, wavelength, etc., must all be comparable. Evidence such as technical drawings, specification comparison tables, and performance test reports should be provided. The operating principles and key performance requirements must also be identical or highly similar.

Software algorithms — this is the biggest upgrade in MDR compared to MDD. MDR explicitly stipulates that the software algorithms in the device under evaluation and the predicate device must be similar. This covers two categories: first, the algorithms in software that drive or influence the use of the device; second, the algorithms in software intended to be used as a standalone medical device (SaMD). When proving software equivalence, the argument must be made from three angles: functional principle, clinical performance, and intended purpose. Evidence such as algorithm flowcharts, source code comparisons, and functional test reports can be provided. However, there is one exception: if the software is used solely for device configuration or calibration and has no medical purpose whatsoever, algorithm similarity is not required — you only need to prove that it will not have a negative impact on usability, safety, or clinical performance.

II. Biological Characteristics

MDR is far more detailed and strict in this area than MDD. If the device has direct or indirect contact with the human body, the following must be compared item by item:

Materials/substances: The same materials or substances must be used. Note that the requirement here is "identical," not "similar."

Contact targets: The tissues or body fluids in contact with the human body must be the same.

Contact mode and duration: The type of contact (e.g., surface contact, extracorporeal contact, implant, etc.) and the duration (transient, short-term, long-term) must be similar. This is a new explicit requirement in MDR that did not exist in MDD with this level of detail.

Substance release characteristics: The substance release characteristics must be similar, including degradation products and leachables. MDR specifically emphasizes the distinction between "identical materials" and "similar substance release characteristics" — even if the raw materials are exactly the same, differences in processing, design, or use environment can cause minor changes in release characteristics, so this item must be demonstrated separately. The biological evaluation can be conducted following the principles of the ISO 10993 series of standards.

Several special cases: For devices composed of substances or combinations of substances intended to be introduced into the human body and that are absorbed by or locally dispersed within the human body, MDR additionally requires that the substances must be completely identical and must comply with the requirements of Annex I of Directive 2001/83/EC. For drug-eluting medical devices (Class III), if a manufacturer wishes to claim equivalence with another manufacturer's device, MDR explicitly requires that the two manufacturers must sign a contract ensuring continuous and full access to the predicate device's technical documentation.

III. Clinical Characteristics

This is the area where MDR has changed the most compared to MDD and has the strictest requirements.

User type must be identical — this is a brand-new explicit requirement in MDR. MDR defines "user" as all healthcare professionals who use the device, as well as lay users (i.e., individuals who have not received formal education in healthcare or medicine, such as a patient self-testing blood glucose at home). When arguing equivalence, the manufacturer must consider whether the expected user's ability and knowledge level would affect safety, clinical performance, and outcomes. For example: if your device is intended for use by ordinary people at home, but the predicate device you cite can only be operated by a doctor in a hospital, the equivalence basically does not hold.

"Same clinical conditions or purposes" specifically covers: same medical condition, gender, duration of use, similar disease severity and disease stage, same anatomical site of use, similar population characteristics (age, anatomy, physiology), and similar key performance indicators.

Expected clinical outcomes: For a specific intended purpose, the expected clinical outcomes and related key performance of the two devices must be similar.

IV. Several Hard-and-Fast Rules That Apply Throughout

First, all three dimensions must be fully satisfied. If there is a significant difference in any one of the technical, biological, or clinical dimensions, you cannot directly cite the other device's data. Having differences is not necessarily a dead end — you can supplement with scientific justification proving that the differences do not affect safety and performance, and at the same time provide your own clinical data to fill the gap. But this will significantly increase the workload.

Second, the predicate device itself must also comply with the current state of the art (SOTA). If the technology of the reference device is already outdated, even if your device is identical to it, you cannot cite it.

Third, no "patchwork equivalence" is allowed. MDR explicitly prohibits you from borrowing one part from Device A and another part from Device B to piece together equivalence. You must be fully equivalent to each selected predicate device across all three dimensions.

Fourth, equivalence is not a one-time thing. As technology develops and SOTA updates, an equivalence relationship that was previously recognized may no longer hold and must be re-evaluated.

Fifth, documentation requirements. The Clinical Evaluation Report (CER) must contain a dedicated section clearly stating: which predicate device(s) were selected and why, how the three dimensions were compared item by item, whether there are any differences, how those differences are explained, and finally, which data from the predicate device was cited.

One-Sentence Summary

MDR equivalence = technical similarity (including software algorithms) + biological similarity (including materials and release characteristics) + clinical similarity (including user type and conditions of use). All three dimensions are indispensable, any difference requires scientific justification, and the overall threshold is a full level higher than MDD — especially software and user type, which are the two areas most likely to get you stuck in a review today.