15 - Case-Control Study

Question 1

In terms of analysis, what is the difference between case-control and cohort studies?

Answer 1

Case-control cannot do IRR but cohort can do ORR and IRR

Question 2

Why would you not investigate leukaemia with a cohort study?

Answer 2

It is a really rare outcome, would have to collect lots of person years. Better to use case-control

Question 3

Briefly outline a case-control study and why we use them?

Answer 3

Question 4

How do you work out an odds ratio?

Answer 4

Question 5

How do you work out 95% CI?

Answer 5

Question 6

Using the following data, work out the IRR?

Answer 6

Question 7

How is the precision of OR affect?

Answer 7

Question 8

What is a nested case-control study?

Answer 8

Collection of data from evolving outcome and exposure of a concurrent or prospective cohort study

Question 9

What are the different types of cohort and case-control study?

Answer 9

Question 10

What are the different types of information bias and what effects do they have on analysis?

Answer 10

Question 11

What are the different types of information bias and what effects do they have on analysis?

Answer 11

Question 12

How can confounding be minimised?

Answer 12

• Matching using important confounders e.g age and gender
• Adjusted for by analysing with logisitic regression so adjusted OR’s

Question 13

Compare cohort and case-control studies.

Answer 13

Question 14

What are the issues for case control and cross-sectional studies?

Answer 14

Question 15

Why would you choose a case-control study over a cohort study for the association of vitamin K with chronic leukaemia?

Answer 15

• Long time for outcome if cohort
• Can explore lots of different exposures in the study, not just one
• Rare disease

Question 16

What is the advantage of having more controls than cases in a case control?

Answer 16

Both healthy and cases affect error factor and therefore 95% CI and it is easier to acquire controls

Question 17

How would you interpret this data?

Answer 17

• There is a 25% increased chance of developing childhood leukaemia when there is pethidine/vitamin K administration
• Null hypothesis of 1 is in the 95% CI so p>0.05 therefore you cannot reject the null hypothesis
• There is also a chance that the risk of developing cancer is decreased by up to 11% or increased by up to 73% with the administration of the medication

Question 18

How could you account for confounding when looking at child birth and pethidine administration over time?

Answer 18

Categorise into 5 year age groups and this can account for confounding