IVD Performance Evaluation: The Two Tables of Analytical Performance and Clinical Evidence

This article is issue 28 of Volume 3, Registration Dossiers and Technical Requirements, of the 50-article series on medical device registration in Switzerland. Together with the other nine issues in this volume it forms a complete deliverable on this topic; the publication timestamp, issue number, title and unique URL are shown in the table below.

1. Key takeaways

IVDs in Switzerland follow a separate ordinance, and performance evaluation requires two tables rather than one. Analytical performance answers whether the test measures accurately — precision, accuracy, detection limit, linearity, interference and stability; clinical evidence answers what accurate measurement is worth — clinical sensitivity, clinical specificity and the clinical value in the intended population. IvDO Art. 19, combined with Article 48(3)(4) of the IVDR, determines the conformity assessment route, and the performance evaluation data are the core input to that route: a missing table on either side brings the assessment to a standstill. One long-term benefit should also be noted: the simplified designation requirements for non-self-test IVDs in IvDO Art. 87 apply indefinitely from 1 January 2025.

2. Legal basis

3. Expert insights

1. The two tables fail for completely different reasons. Points are usually lost on the analytical performance table in method and statistics: sample size rationales unstated, inter-lot variation of reagents omitted, interference studies with incomplete lists of interfering substances. The clinical evidence table usually fails on population and comparator: the intended population is under-represented, the comparison method is unstated or is not an accepted reference method, and the source of positive and negative samples is unexplained. Writing both tables merged into a single chapter makes it easy for reviewers to miss the clinical part; delivering them separately, each standing on its own, has a higher pass rate (practical estimate). A short cover summary stating which claims each table supports also helps reviewers navigate between the two documents.

2. The Swiss transition calendar dictates the rhythm of data upgrades. The transitional arrangements of the IvDO differ from those of the MedDO: class D devices until December 2027, class C until December 2028, and class B and sterile class A until December 2029; certificates issued under the coded regime between 25 May 2017 and 25 May 2022 remain valid until 31 December 2027, and both ordinances have abolished the sell-off period. For IVD manufacturers holding legacy certificates, this means performance evaluation data must be upgraded to the IVDR framework within the transition period, and the part most exposed during that upgrade is the clinical evidence — under the old directives many products never collected clinical performance data systematically.

3. The prolongation of Art. 87 changes the cost model, but it is not a master key. The simplified designation requirements for non-self-test IVDs apply indefinitely from 1 January 2025, lowering the expected cost of notified body involvement for a large number of class B and class C non-self-test products. But simplified designation and exemption from performance evaluation must not be confused: the former is a simplification on the conformity assessment route, while only the latter would waive the evidence obligation. The IvDO does not exempt the compilation of performance evaluation data because of Art. 87 — both tables must still be submitted. Reading Art. 87 as meaning that non-self-test IVDs can skip clinical evidence is the most common misreading of this prolonged policy.

4. Practical recommendations

1. Write analytical performance and clinical evidence as separate documents with independent numbering, so a merged chapter cannot cause reviewers to miss either part.
2. In the analytical performance table, state the sample size rationale, the handling of inter-lot variation and the list of interfering substances item by item, with statistical methods spelled out.
3. In the clinical evidence table, explain the intended population, sample sources, the comparison method and the rationale for the chosen reference standard.
4. Manufacturers holding legacy certificates should back-plan their data upgrades against the milestones of December 2027 for class D, December 2028 for class C, and December 2029 for class B and sterile class A.
5. Non-self-test IVDs may use the simplified designation requirements of Art. 87, but the performance evaluation data must still be compiled and retained in full.

5. Verification checklist

• □ Analytical performance data stand as a separate document, covering precision, accuracy, detection limit, linearity, interference and stability
• □ Clinical evidence stands as a separate document, covering clinical sensitivity, clinical specificity and intended-population representativeness
• □ Sample size rationales, statistical methods and comparator rationales are all stated
• □ Data upgrade plans are back-planned against the IvDO transition calendar (December 2027 / December 2028 / December 2029)
• □ Non-self-test IVDs are confirmed as eligible for the Art. 87 simplified designation requirements, with no reduction of the performance evaluation obligation inferred from it
• □ CHRN registration and swissdamed device registration are coordinated under IvDO Art. 48 and the amended registration obligation

Legal information verified as of 29 October 2026.

GRZAN supports IVD manufacturers with registration documentation and conformity files for the Swiss market: performance evaluation reports in two-table format, analytical and clinical evidence drafting, transition calendars for legacy certificates, and IvDO conformity assessment planning. Our team turns your data into review-ready evidence. For assistance, contact GRZAN at info@grzanglobal.com. Our services cover classes A to D, self-test and non-self-test IVDs, and both new registrations and legacy certificate upgrades.