Which statement best describes immortal time bias in cohort studies?

• A. It arises from misclassification at baseline
• B. It results from loss to follow-up
• C. It occurs when the outcome prevents exposure assignment
• D. It occurs when exposure is assigned after start of follow-up, creating immortal time

Answer: (D)

Immortal time bias happens when there is a period after the study starts during which the outcome cannot happen, and that time is incorrectly counted as exposed person-time. This occurs when exposure status is defined or assigned based on events that occur after entry into the cohort.

In practice, if people are classified as exposed only after they start a treatment, the time from cohort entry up to that first treatment is “immortal”—they had to survive to receive the treatment to be counted as exposed. If that pre-exposure time is included in the exposed group’s follow-up, it artificially inflates the exposed group’s survival or reduces observed events, creating a spurious appearance that the exposure is protective when it may not be.

This is distinct from misclassification at baseline, which is about incorrect exposure status assigned at the start; from loss to follow-up, which biases results for different reasons; and from outcomes that prevent exposure assignment, which is a different selection issue. The essence is the misallocation of time between exposure groups due to the exposure being defined after follow-up begins.