.jpg){kind=link}
- Published
- Updated
Some events prevent another event from occurring later or change its interpretation. Treating such an event as ordinary end of follow-up can inadvertently change the question answered by a clinical trial analysis.
Define the question and assumptions
Define the event of interest and competing events clinically. Probability, event rate and an effect under a specified condition are different summaries. Agree with the statistician which quantity is needed for the decision; do not let a software default determine the intended clinical interpretation of the endpoint.
Hypothetical study example
In a hypothetical study, time to a particular complication is monitored while another event creates a state in which that complication can no longer occur. The team records distinct event types. It justifies an approach matching the clinical objective rather than labelling the competing event as loss to follow-up.
Design review checks
- Specify event-type rules, temporal ordering and any mutually exclusive classifications.
- Present competing-event frequencies separately for each treatment group under comparison.
- Explain whether the selected summary describes a probability, rate or another quantity.
Record the decision and its limits
Acceptance requires reproducible classification and analytical contribution for example event histories. Discuss method assumptions and alternative scientific questions explicitly. Results from different methods should not be compared as if they estimate the same target; treatment differences in competing events belong in the main interpretation, not merely a technical footnote.
Frequently asked questions
- Is a competing event the same as missing data?
- No. The event itself is known and changes the scientific question.
- Is there one universally correct method?
- Method selection depends on the target quantity and clinical objective, with specialist assessment.