Which approach specifically helps prevent immortal time bias in the analysis?

• A. Exclude immortal time from analysis.
• B. Fix exposure status after baseline.
• C. Discard late follow-up data.
• D. Use time-varying exposure definitions or proper time-at-risk allocation.

Answer: (D)

Immortal time bias arises when a period of follow-up during which the outcome cannot occur is misclassified as exposed time, making the exposure look spuriously protective. The way to prevent this is to align exposure status with the actual time at risk by using time-varying exposure definitions or proper time-at-risk allocation. By updating exposure status as time progresses and splitting follow-up into intervals where exposure is known, each unit of person-time is attributed to the correct exposure category. This approach ensures the risk set at every moment reflects true exposure and avoids artificially favorable outcomes for the exposed group. In practice, you’d model exposure as a time-dependent covariate (or allocate person-time accurately so unexposed time ends when exposure starts). Excluding immortal time or fixing exposure at baseline either discards valid data or imposes incorrect exposure classification, and discarding late follow-up loses information and can bias results.