EU MDR Clinical Evaluation: “State of the Art” Means the Alternatives, Not Just Your Device

A Clinical Evaluation Report can have exhaustive data on the subject device — every study, every complaint trend, every PMCF report — and still draw a major nonconformity from a Notified Body reviewer. The most common reason is not missing device data. It is a “state of the art” section that describes the device's own history and calls that the benchmark, when MDR Article 61 and Annex I ask for something else: how the device compares to the alternatives a patient could actually receive instead.

What Article 61 actually asks the state of the art to do

Article 61 requires a manufacturer to confirm conformity with the general safety and performance requirements in Annex I on the basis of clinical data providing sufficient clinical evidence, including an acceptable benefit-risk ratio. That benefit-risk judgment is relative — acceptable compared to what a patient's condition is currently managed with, not acceptable in isolation. The state-of-the-art analysis is the section that is supposed to establish that comparator, which is also why the equivalence route and the state-of-the-art analysis are frequently confused: equivalence lets you borrow another device's clinical data under specific conditions, while state of the art sets the bar your own data has to clear regardless of which route you use to get there.

The finding pattern reviewers keep citing

Two gaps account for most of the state-of-the-art deficiencies Notified Bodies raise, and neither is a data-volume problem:

• The search stops at PubMed. A systematic, reproducible search run only against PubMed misses the European literature a reviewer expects; MEDDEV 2.7/1 Rev 4's methodology calls for at least PubMed and Embase, with a documented PICO strategy and inclusion/exclusion criteria a reviewer can retrace step by step.
• The benchmark has no numbers. Describing the state of the art qualitatively — “several alternative treatments exist” — without pulling quantitative performance and safety thresholds from current clinical guidelines gives a reviewer nothing to compare the subject device's own event rates and success rates against.
• The comparison set is the wrong device. Benchmarking only against the manufacturer's prior model, instead of the full currently accepted range of alternative treatments and competitor devices for the condition, understates the bar the Clinical Evaluation Report has to clear.
• It was written once and never revisited. A state-of-the-art section frozen at first submission misses a comparator's updated data or a newly available treatment option — exactly the kind of shift a periodic refresh exists to catch.

The device's own file can be complete and the benefit-risk conclusion can still be unsupportable, because the comparator it was measured against was never the right one. Why state-of-the-art findings recur even on data-rich submissions

Where this connects to the rest of the file

None of this sits in isolation. A state-of-the-art analysis that correctly identifies the current alternatives also sharpens the intended purpose and risk-management inputs feeding the wider EU MDR & IVDR strategy, and it is exactly the section a PMCF plan should be designed to keep current — new PMCF data on the subject device rarely moves the benchmark, but a competitor's new clinical data or a newly approved alternative treatment does, and the plan needs a trigger for catching that rather than waiting for the next full CER cycle to notice.

Treating the state-of-the-art section as a formality — a paragraph restating the device's own indication before the literature review begins — is how a data-rich CER still draws a major finding. Treating it as the section that sets the comparator the rest of the benefit-risk argument has to clear is what keeps the review from stalling on a section that, on paper, looked finished.