MDR Article 74: The PMCF Study That Still Needs 30 Days’ Notice

A device already carries a CE mark. The post-market clinical follow-up study your team is scoping is, on paper, just evidence generation — confirming what the technical documentation already claims. That framing is exactly where Article 74 of the EU Medical Device Regulation catches teams out. The moment a PMCF protocol asks a subject to do something beyond the device's normal conditions of use — an extra visit, an extra blood draw, an extra imaging session — and that something is invasive or burdensome, the study stops being routine surveillance. It becomes a clinical investigation with its own notification clock and its own substantive requirements.

The three-way split PMCF protocols actually fall into

Manufacturers tend to treat PMCF as a single activity with a single governance answer. The regulation splits it three ways, and the split turns entirely on what the protocol asks participants to do and whether the device stays inside its approved use.

• Within intended purpose, no added invasive or burdensome procedures. Registry analysis, questionnaires, review of data already generated in routine care. Outside Article 74(1) entirely — though national law can still impose its own requirements.
• Within intended purpose, added procedures that are invasive or burdensome. Article 74(1) applies: a 30-day notification through the electronic system referenced in Article 73, with Annex XV Chapter II documentation, plus the imported substantive requirements below.
• Outside the device's intended purpose. Article 74(1)'s notification route does not apply at all. Article 74(2) routes the study through the full Articles 62–81 clinical-investigation regime — the same process, and the same timeline, as a device under investigation for the first time.

The protocol decides the regulatory pathway before the first subject is enrolled. Teams that scope the added procedures after the budget and timeline are set are the ones who discover the 30-day clock too late. Why the classification step comes first

What the notification route still requires

It is tempting to read "notification" as lighter than "application" and assume the substantive bar drops with it. It does not. Article 74(1) carries over points (b) through (k) and (m) of Article 62(4) — the core of what a sponsor must demonstrate for any clinical investigation — along with Articles 75 to 77, which govern how competent authorities assess and act on the submission, and Article 80(5) and (6), which layers the investigation's safety reporting into the incident-vigilance framework. A PMCF study that trips Article 74(1) is a notification procedurally and a full clinical investigation substantively. Build the PMCF evaluation report and the underlying protocol with that in mind from the first draft, not after a competent authority asks why the dossier is thin.

Where this sits next to Article 70

Article 74 is the post-market counterpart to the pre-market route most teams already know: a brand-new clinical investigation application under Article 70, assessed before enrollment can begin at all. Article 74(1) exists precisely because a PMCF study on an already-approved device does not need that full pre-assessment cycle to be safe — but "lighter procedure" was never the same claim as "lighter requirements." Scoping the protocol against the intended purpose and the invasive-or-burdensome test, before the study is budgeted, is what keeps a post-market surveillance program from colliding with a clinical-investigation clock nobody planned for.

None of this is a reason to avoid invasive or burdensome PMCF procedures when the clinical question needs them — some post-market evidence genuinely cannot be collected any other way. It is a reason to make the Article 74 classification a scoping-stage decision rather than a discovery made mid-protocol. A PMCF program built around that discipline treats the notification window as a planning input, not a surprise, and arrives at the competent authority with a dossier that already answers the Article 62(4) questions instead of reconstructing them under deadline pressure.