The IVDR Performance Evaluation Report Is One Synthesis, Not Three Sub-Reports

IVDR Article 56 and Annex XIII require three independently evidenced pillars — scientific validity, analytical performance, clinical performance. Most IVD manufacturers know this by now and have commissioned all three. Fewer have noticed that Annex XIII does not stop there: it requires a further document, the performance evaluation report, that states the device's classification, intended purpose, technology, and evidence-approach justification by drawing on all three. Three well-built sub-reports stapled together are not that document. They are its raw material.

Three evidenced pillars, one open question

Scientific validity, analytical performance, and clinical performance each answer a distinct question, and manufacturers who have internalized that distinction — see why scientific validity is not a literature-review formality — are usually doing solid evidentiary work on each pillar individually. What that work does not automatically produce is the document Annex XIII actually asks for: a performance evaluation report that takes classification, intended purpose, technology, and the evidence-approach justification and states them as one reasoned conclusion, not three conclusions filed in sequence.

What the synthesis has to say

Section 1.3.2 is specific, and specificity is exactly what a stapled-together PER tends to lack. The synthesis has to state the device's classification and the rationale behind it — taking the device's own safety and performance characteristics into account, not just citing a rule number. It has to state the intended purpose and the claims made for the device, the technology the device relies on, and a justification for the level and type of clinical evidence chosen. None of that content lives natively inside a scientific validity report, an analytical performance report, or a clinical performance report. It has to be written, as its own argument, by someone reading across all three.

• Classification is a rationale, not a citation. The synthesis has to show the classification follows from the device's actual safety and performance characteristics — this matters most acutely for Class D devices, where the evidence bar and EU reference laboratory involvement raise the cost of an unreasoned classification statement.
• Intended purpose and claims have to match what was evidenced. If analytical performance data supports a narrower population or cut-off than the clinical performance report assumes, the synthesis is where that gap gets reconciled — or missed.
• The evidence-approach justification is the argument a Notified Body actually tests. It explains why this mix of literature, analytical data, and clinical data is proportionate to this device's risk — not a given once each pillar exists.
• Proportionality cuts both ways. A higher-risk device needs deeper synthesis, not just more pages in each pillar; a lower-risk device's synthesis can be correspondingly brief, but it still has to exist.

Three independently evidenced reports do not average into a conclusion. Someone has to write the conclusion, and Annex XIII says what it has to contain. Why the synthesis is not optional

Where ownership breaks down

The gap rarely comes from weak evidence. It comes from organization: scientific validity sits with a medical or scientific affairs lead, analytical performance sits with the lab or assay development team, clinical performance sits with clinical affairs, and each delivers a competent, independently defensible report on schedule. What nobody is explicitly accountable for is the PER as a single document — so when post-market performance data updates the clinical performance conclusions, the classification rationale and evidence-approach justification written months earlier do not get revisited, because reconciling them was never anyone's named job. That is also where the PER's status as a living document bites hardest: it has to be updated as new post-market performance follow-up data arrives, and a document with three owners and no synthesis owner tends to update one section while the other three quietly go stale.

None of this adds a fourth evidence-generation workstream to an already demanding regulation. It adds one accountable step: reading across scientific validity, analytical performance, and clinical performance, and writing down the classification, intended purpose, and evidence-approach argument those three reports are supposed to support. Manufacturers who treat that step as its own deliverable — owned, reviewed, and updated on its own trigger — are the ones whose EU MDR & IVDR technical documentation holds up when a Notified Body asks not for the three reports, but for the argument that connects them. If your last PER update touched one pillar and left the synthesis as written, that is where to look first; our IVD regulatory strategy work exists for exactly that gap.