Immortal time bias in a cohort study occurs when a period during which the outcome could not occur is misclassified as exposed time. This bias most directly leads to which effect on the observed risk in the exposed group?

• A. Exposure time is always accurately measured, preventing bias.
• B. It describes bias from differential loss to follow-up.
• C. A period during which the outcome could not occur is misclassified as exposed time.
• D. The exposure classification is never biased by time.

Answer: (C)

Immortal time bias happens when a period in which the outcome cannot occur is counted as exposed time. In practice, this means times before someone truly becomes part of the exposed group are still attributed to exposure. For example, if exposure starts after cohort entry, the interval before exposure begins is often included as exposed person-time. Events cannot occur during that immortal interval, so it adds time to the exposed group with no events. This makes the exposed group look to have fewer events per person-time, artificially lowering the observed risk in the exposed group and giving a spurious impression that exposure is protective.