Why is loss to follow-up problematic in cohort studies, and how can you address it?

• A. It can introduce attrition bias if related to exposure or outcome; address with sensitivity analyses, multiple imputation for missing data, inverse probability weighting, and ensuring good follow-up.
• B. It generally has no impact on bias if exposure is rare; address by increasing the study duration.
• C. It is not a concern in cohort studies if the exposure is measured accurately.
• D. It biases results in the opposite direction and cannot be mitigated.

Answer: (C)

Loss to follow-up introduces attrition bias in cohort studies because those who drop out may differ in important ways related to both the exposure and the outcome. If participants with a certain exposure are more likely to leave the study, or if those who develop the outcome are more likely to be lost, the remaining group may no longer be representative of the original cohort. This can distort the estimated association between exposure and outcome, sometimes exaggerating or masking true effects, and the direction of bias depends on how the missingness relates to exposure and outcome.

To address it, focus on both prevention and analysis. Preventively, implement good follow-up procedures, keep contact information up to date, and maintain engagement so fewer participants are lost. Analytically, use methods that account for missing data and differential follow-up: multiple imputation for missing follow-up data under reasonable missing-at-random assumptions, inverse probability weighting to reweight the observed data to reflect the full cohort, and sensitivity analyses exploring how results change under different missing data scenarios. Also compare baseline characteristics of those followed vs. those lost to assess potential bias, and report results with and without adjustments for follow-up loss when appropriate.