Equivalence in the CER: How to Build the MDCG 2020-5 Evidence Table

In the transition from MDD to MDR, one part of the CER breaks more often than any other: equivalence. Under the MDD, the phrase "a similar device used for the same purpose" was accepted for years and the literature route leaned on it. MDR Annex XIV Part A(3) narrows that door; MDCG 2020-5 defines how to pass through the narrower door, which table is accepted and which claims are no longer valid.

This article covers the three dimensions of equivalence, the single-device rule, the data-access requirement, the limits set by Articles 61(5) and 61(6), and how the comparison table a Notified Body accepts is built.

Three dimensions: technical, biological, clinical

Annex XIV Part A(3) defines equivalence across three sets of characteristics and requires all three to be demonstrated:

• Technical: Similar design, similar conditions of use, similar specifications and properties (physicochemical properties, energy intensity, tensile strength, viscosity, surface characteristics, wavelength, software algorithms), similar deployment method and similar principles of operation.
• Biological: The same materials or substances in contact with the same human tissues or body fluids for a similar kind and duration of contact, including release characteristics and degradation products.
• Clinical: Use for the same clinical condition or purpose, at a similar severity and stage of disease, at the same site in the body, in a similar population; the same kind of user; similar critical performance.

The rule is explicit: "similar" is not enough. For each characteristic the difference is defined and it is shown, with scientific justification, that the difference is not clinically significant for safety and clinical performance. A single difference that cannot be justified defeats the equivalence claim.

The MDCG 2020-5 table

MDCG 2020-5 turns the manufacturer's comparison into a table. Each row is a characteristic; the columns are the device under evaluation, the equivalent device, the differences, the assessment of the clinical significance of the differences, and the supporting source. The Notified Body reads this table row by row.

Three principles decide how the table is built:

• Sourced cells: Behind every cell sits technical documentation, an instructions-for-use document, a material certificate or a publication. "As is well known" is not accepted.
• Clinical relevance of the difference is justified: What is shown is not the absence of a difference but the insignificance of it. That justification rests on literature, bench testing or preclinical data and is consistent with the appraisal section of the CER.
• No empty rows: If information on a characteristic is unavailable, the row is not left blank; the gap is stated openly and its effect on the conclusion discussed. An empty row makes the assessor doubt the rest of the table.

The single-device rule

MDCG 2020-5 expects equivalence to be built to a single device. If more than one device is used, equivalence in all three dimensions is demonstrated for each one separately. Assembling a virtual "equivalent device" from the convenient characteristics of different devices is not accepted. In practice this rule is where CERs fall most often: one device is taken for the technical side and another for the clinical side, and a merged narrative is written instead of two separate tables.

Data access and Article 61(5)

Annex XIV Part A(3) requires the manufacturer to demonstrate clearly that it has sufficient access to the data on the equivalent device. Published literature is part of that access, but technical and biological characteristics can rarely be shown from publications alone.

Article 61(5) makes the limit firm for implantable and class III devices: if equivalence is claimed to another manufacturer's device, there must be a contract between the two manufacturers granting full and ongoing access to the other device's technical documentation, and that device's clinical evaluation must have been performed in accordance with the MDR. Without such a contract the equivalence route is closed for these classes; your own clinical investigation is required.

Article 61(4) treats separately the case where the manufacturer claims equivalence to its own device already on the market; Article 61(6) provides an exemption from the clinical-investigation requirement for certain well-established technologies (listed product groups such as sutures, staples and dental filling materials). These exceptions are interpreted narrowly and their justification is written into the CER.

To have your equivalence table, and the way the equivalent device's literature is linked to the CER, reviewed from the Notified Body's perspective, request a free 15-minute scoping call.

Equivalence for software devices

For software medical devices the technical row includes "software algorithms", and an equivalence claim is especially hard there: two algorithms serving the same clinical purpose may have been trained on different data, operate at different thresholds and have different performance profiles. MDCG 2020-1 defines clinical evaluation for software through the triad of valid clinical association, analytical performance and clinical performance; an equivalence claim must be justified at all three levels. In practice the equivalence route is rarely sufficient on its own for software devices; standalone performance data of your own is needed. See how to design a SaMD clinical validation study for that design.

What happens to the literature if equivalence cannot be built

Publications on a device whose equivalence cannot be shown are not discarded from the CER; they change place. They contribute to the definition of the state of the art, to the performance and safety range of alternative treatment options and to the benefit-risk comparator. They do not count as evidence for the device's own performance and safety claims. How that distinction is made in the literature section is covered in Notified Body findings on the CER literature search.

Common Mistakes

• Updating an MDD-era CER without touching the equivalence section: The most frequent finding on legacy devices; MDCG 2020-6 defines sufficient clinical evidence for legacy devices separately and expects the equivalence claim to be rebuilt to the MDR standard.
• Hiding the difference: The assessor sees the difference anyway; justifying its clinical insignificance is always stronger than not writing it down.
• Not aligning the intended-purpose sentences: Clinical equivalence starts with a comparison of the two devices' intended-purpose statements; if indication, population and user differ, the row is red.
• Claiming access to a competitor's device: Building technical equivalence from the instructions for use and brochures alone does not satisfy the data-access requirement.

Related Articles

• Notified Body Findings on the CER Literature Search
• Literature Appraisal in Clinical Evaluation: MEDDEV 2.7/1 Appendix A6
• How to Write the CER Statistics Section
• The SaMD Clinical Evidence Package: A Single-File Map

Under the MDR, equivalence is not a shortcut but a separate evidence burden. When the table is built row by row and every difference is justified, the literature route opens; when it cannot be built, seeing that early lets you start collecting your own clinical data in time. The experience behind the method is on the about page.

Scope, timeline and budget differ for every file; we settle them in a free scoping call:

Free Scoping Call