_(32290196795).jpg){kind=link}
- Published
- Updated
When the measurement method changes during follow-up, an observed difference may not solely reflect a change in the participant. Record changes in device, technique or assessment mode as potential interpretation issues, rather than merely logistical details.
Define the question and assumptions
Evaluate comparability between baseline and follow-up measurements. Unit conversion, calibration and measurement conditions address different problems; resolving one does not automatically resolve the others. Where possible, assemble appropriate bridging evidence and a statistical assessment plan before implementing the proposed method change.
Hypothetical study example
In a hypothetical multicentre trial, some sites switch to another device. Results use the same unit, but the methods have unknown distributional differences. The team links device dates to observations, examines comparability evidence and avoids interpreting the observed arm difference directly as a treatment effect.
Design review checks
- Maintain traceable method versions and the dates on which changes occurred.
- Check that identical units are not mistaken for identical measurement behaviour.
- Examine whether method changes coincide with treatment groups, sites or follow-up time.
Record the decision and its limits
The acceptance decision explains the potential scientific impact and remaining uncertainty. Plan sensitivity or supportive analyses when appropriate. If a transformation is proposed, document its rationale, validation and affected observations; producing a convenient result is not sufficient justification for changing the measurement scale.
Frequently asked questions
- Do identical units establish comparability?
- Not alone. The methods and their measurement properties must also be examined.
- Does every method change invalidate the data?
- No. Assess its impact using available evidence and the scientific objective.