Case-Control Studies Flashcards
What are the two groups in a case control study?
Cases - those individuals who have the disease,
Controls - individuals who do not have the disease.
State two ways in which case-control studies are better than cohort studies.
They are less expensive, quicker and they a are better for studying rare outcomes.
What do we use in a case-control study to measure excess risk between cases and controls?
Odds ratio which is calculated AD/BC
How can we lower the error factor for an odds ratio?
Increase the sample size - it is worth having 4-6 times as many controls as cases.
What is a conventional case-control study?
This is where historical data is used and collected retrospectively about individuals who become cases.
What is a nested case-control study?
This involves collection of data for an outcome which is evolving, e.g. Exposure in a concurrent cohort study.
In what type of case control study is the population for sampling already defined?
Nested case control study.
What is selection bias in a case control study and what effect will this have on the odds ratio?
This is where the way in which people are chosen for the study makes it unrepresentative of the general population. This will lead to an underestimate of the odds ratio.
What is the effect of non differentiated mis classification on an OR?
This will cause shrinkage of the odds ratio towards the nulls.
State a way in which confounding can be reduced in case control studies.
By matching individual cases with controls (say by date of birth) this can reduce the risk of confounding.
State the effect of systematic mis classification on the OR. How can this arise?
Is can lead to shrinkage to the null or for it to get further away. This can be caused by recall bias, assessor bias, or a difference in data collection methods.
What is a case-control study?
This is where we look at individuals who have a disease and look backwards to find their exposure.
