How does start time and immortal time bias relate to exposure definition?

• A. If exposure is assigned at baseline, immortal time bias can occur
• B. If exposure is assigned after start of follow-up, immortal time bias can occur; ensure exposure classification is based on baseline information
• C. Immortal time bias only occurs in case-control studies
• D. Immortal time bias is not a concern in cohort studies

Answer: (B)

Immortal time bias happens when how you define exposure depends on future information or events that must occur to be classified as exposed. In a cohort, if you determine who is exposed based on events that happen after the study starts (for example, starting a medication during follow-up and then labeling people as “exposed” from the time of enrollment), there is a period before exposure during which the person would have to survive to become exposed. That interval is “immortal” with respect to the outcome, so people in the exposed group appear to have better outcomes simply because times before they actually received exposure are counted as exposed time. This artificially lowers event rates in the exposed group and biases results toward showing a protective effect.

Therefore, assigning exposure after the start of follow-up can create immortal time bias, and the way to avoid it is to classify exposure based on information available at baseline, so the exposure status is fixed from the start and there’s no immortal time created by future exposure definitions.

Other options are not correct because immortal time bias is not limited to case-control studies, and assigning exposure at baseline does not introduce immortal time bias.