A case-control study of coffee drinking and pancreatic cancer yields a crude odds ratio of 2.0. When the data are stratified by smoking status, the odds ratio is 1.1 in smokers and 1.1 in non-smokers, and smoking is strongly associated with coffee drinking in the controls. The crude estimate differs from the stratum-specific estimates mainly because of:
- A Effect modification by smoking
- B Selection bias in case ascertainment
- C Confounding by smoking ✓
- D Recall bias regarding coffee intake
Correct answer: C. Confounding by smoking
Explanation
Confounding is diagnosed when the crude estimate differs materially from the homogeneous stratum-specific estimates; here both strata show OR near 1.1 while the crude OR is 2.0, so smoking distorts the coffee-cancer association. Smoking qualifies as a confounder because it is associated with the exposure (coffee), independently associated with the outcome, and not on the causal pathway. Effect modification is excluded because the stratum-specific estimates are nearly identical rather than divergent.
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.