Questions 1-5 are based on the following scenario:
In a cohort study, men were classified as having high or low cholesterol levels. Among the group with low cholesterol levels there were 50 heart attacks in 5000 person-years of follow-up and among the group with high cholesterol levels there were 150 heart attacks in 6000 person-years of follow-up.
Q1: What was the incidence of heart attack among men with low cholesterol levels?
Q2: What was the incidence of heart attack among men with high cholesterol levels?
Q3: What was the incidence of heart attack all men?
Q4: How strong was the association between high cholesterol and heart attack?
Q5: If high cholesterol levels are a cause of heart attack, what proportion of heart attacks in men could theoretically be prevented if nobody had a high cholesterol level?
Q6: Name 5 potential sources of selection bias in epidemiological studies.
Questions 7-10 are based on the following scenario:
A case-control study was conducted to evaluate the relationship between physical activity andcoronary heart disease (CHD) in men. A total of 406 men newly diagnosed with CHD were included together with 406 men of similar ages who did not have CHD. The risk of CHD (measured by the OR) was higheramong men who were inactive or only moderately active (collectively called ‘inactive’) than among those who were physically active:
Odds ratio for inactive men compared to active men= (299 x 136)/(270 x 107)=1.41
Suppose there is no misclassification in these data.
What would the observed OR have been…?
Q7: If 20% of all the inactive men had been misclassified as active?
Q8: If 10% of all the active men had been misclassified as inactive?
Q9: If 20% of inactive controls had been misclassified as active?
Q10: If 10 % of inactive cases had been misclassified as active?
Questions 11-14 are based on this table:
Q11: Is the adjusted relationship between body mass index and ER-positive breast cancers statistically significant? Give the reason for your answer in one short statement (no more than 10 words)
Q12: What is the most likely adjusted estimate for the RR for body mass index and ER-positive breast cancers? (Provide a numerical answer with 2 decimals)
Q13: What is the highest plausible value for adjusted RR for body mass index and ER-positive breast cancers? (Provide a numerical answer with 2 decimals)
Q14: If you were to calculate a Rate Difference for the relationship of body mass index and ER-positive breast cancers and test for statistical significance, what is your Null Hypothesis.
