A cohort study shows a crude relative risk of 2.0 for coffee intake and pancreatic cancer. After stratifying by smoking status, the relative risk is 1.2 among nonsmokers and 1.25 among smokers. The most appropriate conclusion is:
- A Smoking confounds the coffee and pancreatic cancer association and the data should be adjusted for smoking ✓
- B The association is causal since stratum-specific risks remain above 1
- C Smoking modifies the effect of coffee, so interaction should be reported
- D Stratification was inappropriate because the crude estimate should always be preferred
Correct answer: A. Smoking confounds the coffee and pancreatic cancer association and the data should be adjusted for smoking
Explanation
Confounding is suspected when the crude measure differs meaningfully from the stratum-specific measures, and the stratum-specific estimates are similar to each other. Here both strata show about 1.2 while the crude value is 2.0, indicating smoking accounted for much of the apparent association and adjustment is warranted. Effect modification requires the stratum-specific estimates to differ substantially from each other, which they do not.
Reference: Gordis Epidemiology, 6th ed.
High-yield for: NEET PGINI-CETNExTFMGEUSMLEPLABMRCP
Written and medically reviewed by the StethoPrep medical team.