A Directed Acyclic Graph (DAG) shows: Smoking → Lung Cancer, Smoking → Cardiovascular Disease, and Air Pollution → Lung Cancer. If one adjusts for Cardiovascular Disease in a study of Smoking → Lung Cancer, this introduces:

• A Confounding bias — cardiovascular disease confounds the smoking-lung cancer relationship
• B Measurement bias — cardiovascular disease misclassifies smoking status
• C Selection bias — adjusting for cardiovascular disease restricts the study population
• D Collider bias — cardiovascular disease is a collider on the path between smoking and lung cancer ✓

Explanation

A collider is a variable that is caused by two or more variables on a causal path; adjusting for a collider opens a spurious non-causal path between its causes. Cardiovascular disease is caused by both smoking and (potentially) air pollution; conditioning on cardiovascular disease creates a spurious association between smoking and air pollution (M-bias or collider stratification bias). This is a DAG-based concept in modern causal inference: conditioning on colliders introduces, rather than removes, bias.

Reference: Park's Textbook of Preventive and Social Medicine, 27th ed.

High-yield for: NEET PGINI-CETNExTFMGEUSMLEPLABMRCP

Written and medically reviewed by the StethoPrep medical team.