Epidemiology & Prevention USMLE Step 1 Practice Question
A 45-year-old man with a 20-year smoking history presents to the emergency department with acute anterior wall myocardial infarction. His initial vital signs are BP 148/92 mmHg, HR 102/min, RR 18/min, and SpO2 94% on room air. Troponin I is 2.8 ng/mL. During admission, his serum total cholesterol is measured at 245 mg/dL. Epidemiologists studying a large cohort of patients with acute MI note a strong positive association between serum cholesterol level and risk of myocardial infarction in the overall population (p < 0.001). However, when the cohort is stratified by smoking status, the association between cholesterol and MI risk substantially weakens in both smokers and non-smokers (p = 0.15 in each stratum). Which of the following epidemiologic phenomena best explains the difference between the crude and stratum-specific associations?
Answer choices
- AEffect modification, in which smoking status alters the strength of the relationship between cholesterol and MI risk
- BSelection bias, as differential enrollment of high-cholesterol smokers versus non-smokers artificially strengthens the crude association
- CInformation bias, due to differential misclassification of smoking status based on cholesterol levels
- DConfounding, in which smoking status is associated with both cholesterol and MI risk, distorting their crude relationshipCorrect answer
- ERegression to the mean, as extremely elevated cholesterol values in the overall group regress toward baseline upon stratification
- FBerkson's paradox, whereby the association between cholesterol and MI appears stronger in the hospitalized population than in the general population
See the full explanation
Get the correct-answer rationale, why each distractor is wrong, the underlying mechanism, and high-yield associations — plus adaptive practice that targets your weak areas — with a free MedBoardPRO account.