When Equivalent Device Evidence Can Support an EU MDR Clinical Evaluation

Key points#

• Equivalence is a claim about one identified device configuration and one identified comparator configuration, not about a broad product family.
• The EU MDR requires technical, biological, and clinical characteristics to be considered together. A match in only one or two domains is incomplete.
• Every difference needs a scientific explanation of whether it could create a clinically significant difference in safety or clinical performance.
• Data access is a threshold question. If the comparison cannot be checked with sufficiently complete and accurate information, it cannot support an equivalence claim for conformity assessment.
• Equivalent-device data may strengthen a clinical evaluation. They do not replace the clinical evaluation, risk management, post-market clinical follow-up, or device-specific evidence where gaps remain.

Begin with the conclusion the evidence is supposed to support#

An equivalence analysis is an evidence bridge. Clinical data were generated with Device B, while the manufacturer wants to use them to support Device A; the bridge is sound only if the characteristics that could influence safety or clinical performance are sufficiently alike and the important differences have no clinically significant effect.

That framing prevents several common errors. Two products are not equivalent merely because they share a trade description, classification rule, intended specialty, or mechanism in a marketing summary, and a device generation is not interchangeable with every earlier model in its family. A CE mark on the comparator does not prove that its data apply to the new product; the analysis must name the precise device, model, version, accessories, intended purpose, and use conditions on both sides.

The intended claim sets the needed resolution. If the proposed evidence supports long-term implant durability, then a comparison limited to short-term procedural success is not enough, and if the claim includes home use, evidence from a professionally operated hospital device may not transfer unless differences in users and settings are addressed. The question is always claim-specific: which result is being carried across, and which device characteristics could change that result?

Build one comparison across three connected domains#

Annex XIV of the MDR organizes equivalence into technical, biological, and clinical characteristics. These are not three alternative routes. They are parts of one justification.

Technical characteristics describe how the product produces its effect#

The technical assessment covers design, specifications, and operating principle. It covers conditions of use, deployment method, and critical performance. Depending on the product, it may also require dimensions, tolerances, and energy output. It may require surface properties, mechanical strength, and accessories. It may require calibration, software logic, cybersecurity-related function, or manufacturing features that affect performance.

The comparison needs more than a row marked "similar." It should identify the source record for each value, describe any difference, and connect the difference to possible clinical consequences, and a small dimensional change may be unimportant for one external device and decisive for an implant placed in a narrow anatomical space. A revised algorithm may preserve the user interface while changing which patients are classified as positive. Context decides materiality.

Software deserves version-level precision. Matching a product name is weak if the model, input specification, decision threshold, or clinical workflow changed. The analysis should establish which release generated the cited data and whether that release represents the product under evaluation.

Biological characteristics concern the finished device in contact with the body#

For biological equivalence, Annex XIV calls for the same materials or substances in contact with the same tissues or body fluids, with a similar kind and duration of contact and similar release characteristics. MDCG 2020-5 emphasizes that the material-contact requirement is not satisfied by a generic statement that both products use a familiar polymer or metal.

Processing can change what reaches the patient. Sterilization, coatings, and colorants can matter even when the base material name matches. So can adhesives, manufacturing residues, and surface treatment. So can degradation and leachables. The comparison therefore concerns the finished device and its contact conditions. The record should show material identity, contact location, contact duration, and relevant release behavior, then explain the clinical significance of any difference.

Clinical characteristics test the real intended use#

The clinical domain includes the condition or purpose, disease severity and stage, and anatomical site. It includes patient population, user type, and clinically relevant performance. Similar technical output does not establish clinical equivalence if it is produced in a different population or used to make a different decision.

Age, anatomy, and physiology can change both benefit and risk. So can comorbidities, care setting, user training, and disease spectrum. A clinician-operated system and a patient-operated system may differ even if their hardware is close. A performance value measured in a tertiary referral population may not represent a screening population with a lower event rate. Clinical comparison should therefore follow the full use pathway rather than stop at the device specification.

Treat differences as the central work product#

A useful equivalence table makes your differences visible. For each required characteristic, record:

1. the specification and evidence for the device under evaluation;
2. the corresponding specification and evidence for the proposed equivalent device;
3. the exact difference or confirmation of no difference;
4. the mechanism by which that characteristic could affect safety or performance;
5. supporting verification, validation, nonclinical, clinical, or literature evidence; and
6. a reasoned conclusion on clinical significance.

The table is an index to the argument, not the argument itself. "The new design is better" is not a scientific resolution. An improvement can alter failure modes, user behavior, or clinical effect. Likewise, an absence of reported problems is not proof of no meaningful difference when surveillance or version information is incomplete.

More than one comparator may be assessed, but each asserted equivalent device must satisfy the complete comparison. Selecting the materials from one product, performance from a second, and intended use from a third creates a hypothetical composite rather than an equivalent device. Data from products that are merely similar may still inform the state of the art, risk analysis, endpoint selection, or a literature review, but they should be labeled as similar-device evidence, not promoted to equivalence.

Resolve data access before the strategy depends on it#

You must have sufficient access to the data needed to establish the technical, biological, and clinical characteristics. Public sources can be useful. But they often omit exact composition, manufacturing details, or tolerances. They omit software versions, corrective changes, or unfavorable findings. Marketing material rarely supplies a verifiable comparison on its own.

MDCG 2023-7 clarifies two points that are easy to blur. First, sufficient access does not always mean access to the comparator's entire technical documentation. It means access adequate to establish the characteristics relevant to the equivalence claim. Second, a higher level of access cannot rescue devices that are materially different. Complete records improve confidence in the comparison; they do not make a failed comparison pass.

The guidance describes a hierarchy of possible access arrangements and their limitations. Document what you can inspect, how current the data are, which version they describe, how accuracy was checked, and how any missing information was addressed. The notified body evaluates whether that access is sufficient. If missing information compromises completeness or accuracy, the equivalence claim cannot be used for conformity assessment.

Keep the high-risk-device exception separate from ordinary equivalence#

Article 61 gives implantable and class III devices a strong expectation for clinical investigation, subject to specified exceptions. One route concerns a modified device already marketed by the same manufacturer. Another defined route in Article 61(5) concerns reliance on a device made by another manufacturer and requires a contract that allows full, ongoing access to the technical documentation, along with the other statutory conditions.

That contractual rule should not be generalized into either of two incorrect claims: that every equivalence analysis always requires a contract, or that sufficient access without a contract always removes the clinical-investigation expectation for a high-risk device. Device class, manufacturer relationship, and regulatory history all matter. So do common specifications, notified-body endorsement, and the precise Article 61 route. The legal exemption analysis and the scientific equivalence analysis are connected but distinct.

Use the data inside a complete clinical evaluation#

Once equivalence is demonstrated, the comparator's relevant clinical data can be appraised within the clinical evaluation. You must still assess methodological quality, relevance to the intended purpose, and favorable and unfavorable findings. You must assess bias, duration of follow-up, and whether the total evidence supports the claimed benefit-risk profile.

The next decision is a gap analysis. Does the assembled evidence cover every important claim, population, user, duration, and risk? Novel features, weak comparator studies, short follow-up, uncertain long-term safety, or an expanded intended purpose may still require a clinical investigation of Device A. Nonclinical testing can answer some questions, but it cannot automatically substitute for a missing clinical outcome.

Finally, equivalence is not permanent. Product changes, new evidence, complaints, vigilance signals, and post-market clinical follow-up can alter the conclusion. A defensible file is version-controlled and updated when either device or the state of the art changes. The lasting principle is straightforward: evidence follows the product only as far as a complete, verifiable, clinically reasoned comparison can carry it.