- Published
- Updated
Comparing a new IVD with another test does not mean the other test is always correct. Describing its role and limitations helps avoid mislabelling performance. Link reference selection to the intended diagnostic question.
Define the assessment
Explain why the benchmark is trusted, how the target condition is defined and where uncertain outcomes belong. A reference standard and an alternative comparative method are not identical concepts. Qualified statistical and clinical reviewers should examine the basis for labelling outcomes correct or incorrect.
Worked example
For example, a new assay shows high agreement with a commercial test, but both may share an error source. Instead of immediately claiming clinical accuracy, the team examines comparator quality, explains what agreement establishes and records the absence of independent condition-status evidence as uncertainty.
Practical checks
- Document the target-condition definition alongside the process generating the reference decision.
- Check whether knowledge of comparator results influences interpretation of the new test.
- Keep additional investigation of disagreements traceable and distinct from the principal performance calculation.
Decision and limitations
The report should demonstrate that the measures fit the comparison design. Do not convert an agreement percentage directly into every clinical-accuracy claim. Explain possible reference errors or missingness and identify which analyses use results changed by additional testing.
Frequently asked questions
- Is another test always a reference standard?
- No. Assess its quality and the basis for establishing the target condition separately.
- What does high agreement establish?
- Primarily similarity to the selected comparator; broader accuracy interpretation depends on design and reference evidence.