Case-Control Studies

Question 0

Explain the Odds Ratio (OR).

Answer 0

(ad)/(bc) = a/b divided by c/d, or a/c divided by b/d. Compares the odds of having the the disease if exposed with the odds of having the disease if unexposed. Or the odds of being exposed if you have the disease with the odds of being exposed if you don’t have the disease. Assumes disease is rare i.e. (A«b>«D) so OR=IRR.

Question 1

How is a case-control study performed?

Answer 1

Recruit diseased (cases) and disease-free (controls) individuals and then determine their exposure.

Question 2

What are the advantages of case-control studies?

Answer 2

Quicker and cheaper than cohorts, good for rare diseases, can study a range of exposures, no losses to follow up.

Question 3

What are the disadvantages to case-control studies?

Answer 3

Not good for rare exposures, prone to selection and information (recall) bias, only nested studies can ensure exposure precedes outcome and measure incidence or absolute risk.

Question 4

What is selection bias?

Answer 4

When the cases in the study are not representative of the general population from which the cases came (eg. choose from cardiac ward).

Question 5

What is recall bias?

Answer 5

Incorrectly determining exposure status by looking back at history. Maybe systematic, eg. cases looking for explanation for disease, or non-differentiated misclassification - shrinkage to null.

Question 6

How do you deal with confounding in case-control studies?

Answer 6

Matching cases and controls with similar details. Logistic regression adjusts for it.

Question 7

How is the error factor calculated for an odds ratio?

Answer 7

e^(2x rt(1/a + 1/b + 1/c + 1/d)). Therefore it is worth increasing the number of controls to reduce the error factor (up to about 6 x cases).