- Published
- Updated
Repeated measurement of one specimen can help assess variability under specified conditions. Repeats are not independent clinical participants. Unclear specimen, repeat and condition definitions can distort sample counts and apparent precision.
Define the assessment
State the purpose, timing, operator and instrument for repeats. Planned replication and retesting prompted by an uncertain result answer different questions. Replacing failed attempts with only the final successful measurement can discard information about workflow and measurement behaviour.
Worked example
For example, ten specimens are measured three times and reported as thirty clinical samples. The team separates measurement counts from people and specimen counts. It discusses an assessment accounting for related repeats with the statistician and retains all raw results rather than selecting only the closest pair.
Practical checks
- Identify people, specimens, aliquots and repeat measurements as separate units.
- Reconcile planned replication with actual repeats and their reasons.
- Define result-selection or summarisation rules before reviewing measurements.
Decision and limitations
The report should explain which conditions the variability assessment represents. Good repeat agreement under one condition does not automatically cover different sites or users. More repeats do not eliminate limited clinical representation; keep the contributions of technical and clinical evidence distinct.
Frequently asked questions
- Is every repeat a new clinical sample?
- No. People, specimens and measurements are different units and should be reported separately.
- May only the most consistent repeats be selected?
- Avoid arbitrary outcome-driven selection. Define and justify the rule in advance.