A cohort study reports a crude relative risk of 1.8 for coffee consumption and pancreatic cancer. When the data are stratified by smoking status, the relative risk is 1.2 among smokers and 1.2 among non-smokers. The crude estimate differs from the stratum-specific estimates mainly because:
- A Coffee consumption causes smoking
- B Smoking is an effect modifier of the association
- C Smoking is a confounder of the coffee-pancreatic cancer association ✓
- D The study suffers from differential misclassification
Correct answer: C. Smoking is a confounder of the coffee-pancreatic cancer association
Explanation
Confounding is suspected when the crude measure differs meaningfully from the stratum-specific measures, commonly judged by a change of more than 10 percent after adjustment. Here smokers drink more coffee and also have higher pancreatic cancer risk, so smoking distorts the crude association upward. Effect modification requires the stratum-specific estimates to differ substantially from each other, which they do not. Causation of smoking by coffee is irrelevant to why the crude estimate is inflated.
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.