- Published
- Updated
If membership in a group requires remaining event-free for a period, the comparison can appear artificially favourable. This problem can arise when treatment received later or future recording information is used to classify an earlier follow-up period as already exposed.
Define the question and assumptions
Place selection, exposure classification and follow-up origin on one timeline. Ask which events can occur during the waiting period required to enter a group. Discuss suitable design approaches or time-varying exposure methods with a specialist; merely adding total follow-up time to a model does not automatically repair incorrect classification of earlier person-time.
Hypothetical study example
In a hypothetical registry, the treated group includes people receiving an intervention months after diagnosis. Those people necessarily survived until treatment. The team checks whether time between diagnosis and administration was incorrectly credited to the treated group, creating an advantage before the intervention could have had any effect.
Design review checks
- Check whether group membership requires information from a future event.
- Show explicitly how time before exposure contributes to the compared groups.
- Test eligibility and follow-up rules using representative individual event timelines.
Record the decision and its limits
Acceptance reproducibly explains how time and events contribute. Naming the bias is not evidence that it has been corrected; inspect actual derivations. If an alternative design creates a different target population, explain that change. Report remaining temporal uncertainty and the limits it places on causal interpretation of the resulting comparison.
Frequently asked questions
- Does longer follow-up remove this bias?
- No. Incorrect classification and time-origin rules may remain problematic.
- Does it concern only death outcomes?
- No. Event-free requirements for group entry can affect other outcomes too.