Writing an IVDR Scientific Validity Report

How the three elements of performance evaluation connect

Annex XIII requires the performance evaluation to address scientific validity, analytical performance and clinical performance. These are not three independent documents that happen to share a cover page. The scientific-validity conclusion defines why the measured analyte is relevant; analytical performance shows how reliably the device measures it; clinical performance shows whether the result is associated with the target clinical condition or process in the intended population.

If the intended purpose changes, the scientific-validity question can change with it. A marker supported for one population, specimen or clinical role is not automatically supported for another.

Document a reproducible literature search

The search should be planned before conclusions are written. Databases, dates, concepts, limits, duplicate handling, inclusion criteria, exclusion criteria and screening stages need enough detail for another reviewer to understand what was searched and why evidence was retained or rejected.

A broad search may be useful for discovery, but the report needs a focused claim map. The reviewer should be able to trace each final statement back to the intended purpose, a defined evidence question and the cited source.

• Define analyte, clinical condition or physiological state, population, specimen and clinical role.
• Predefine the hierarchy of source types and the treatment of reviews, guidelines and primary studies.
• Record exclusion reasons at full-text review and retain the search evidence package.
• Assess relevance, methodological quality, consistency and residual uncertainty.

What if the literature is incomplete?

State what the available literature does not establish, assess whether another source can support the relationship and decide whether new evidence is needed. The performance evaluation plan should show how the gap affects analytical studies, clinical performance work, claims and post-market follow-up.

For an established analyte, the scientific-validity argument may rely heavily on peer-reviewed literature and accepted clinical knowledge. For a novel marker, combination or clinical role, the manufacturer may need its own studies or other justified evidence. The depth should be proportionate to novelty, risk and claim specificity.

Updating the report after certification

IVDR requires continuous performance evaluation. New literature, vigilance signals, PMPF findings, changes in clinical practice and changes to the intended purpose can all alter the scientific-validity conclusion. The report therefore needs a defined update trigger and an owner within the performance-evaluation process.

Record which conclusions, plans and reports are affected by each new finding, then update the corresponding controlled documents.

Separate a useful paper from a paper that supports this claim

A study may be scientifically sound yet address a different question. Suppose a marker is associated with disease severity in already diagnosed adults, while the proposed device is intended for screening an asymptomatic population. The paper may contribute background information, but it does not by itself establish the full relationship needed for that screening claim. Record the difference rather than discarding the paper without explanation or treating it as direct confirmation.

A practical appraisal table can identify the population, specimen, clinical role, reference definition and relevant findings, followed by an assessment of applicability and limitations. Keep the underlying study separate from reviews that discuss it so that the same evidence is not counted several times. Give contradictory findings a visible place in the evaluation.

Turn the conclusion into a usable evidence decision

The conclusion should distinguish what is established from what remains uncertain. If the relationship is supported only for a narrower population or clinical role, explain the boundary and its implications for the intended purpose. A report that says “more evidence is needed” without identifying the unanswered question gives the study team little direction.

For each gap, state whether it concerns the biological relationship, the device measurement or performance in the proposed clinical context. Those questions may lead to different work. A literature update, an analytical experiment and a clinical performance study cannot automatically substitute for one another. Assign ownership for resolving the gap and identify the document in which the resulting evidence will be evaluated.

Plan updates around changes that could affect the conclusion

Define how new literature, changes in clinical practice, new populations or changes to the intended purpose will be assessed. Retain the previous search strategy and appraisal decisions so the update can show what changed. An update should not silently replace an older conclusion without explaining the new evidence or reasoning.

During review, check that the report still agrees with the current performance evaluation plan, risk documentation and instructions for use. Consistency is especially important when a commercial description evolves faster than the scientific file. The aim is a maintained argument for the actual product claim, with accessible evidence and explicit limits.