L3 Case-control studies

Question 1

Define a case-control study.

Answer 1

The observational epidemiologic study of persons with the disease of interest and a suitable control group of persons without the disease.

Question 2

How are case-control studies retrospective?

Answer 2

They start by knowing who has the disease.

Then they work back to find who was exposed.

Question 3

Case-control studies are always the design of choice if a disease is ___, or if a quick answer is required.

Answer 3

Rare.

Question 4

Name the 2 major sources of cases in a case-control study.

Answer 4

Population-based.

Hospital-based.

Question 5

What are the advantages of population controls in case-control studies?
(3)

Answer 5

Exclusion criteria easy to apply.

Cases and controls from same study base.

Representative of whole study base.

Question 6

What are the disadvantages of population controls in case-control studies?
(4)

Answer 6

Low participation rates could lead to bias.

Inconvenience of finding controls.

Recall bias.

May not be motivated to take part.

Question 7

Define neighbourhood controls in case-control studies.

Answer 7

Controls selected from residences in same geographical areas as the cases.

Question 8

What are the advantages of neighbourhood controls in case-control studies?
(2)

Answer 8

Selection doesn’t need a roster of people.

Possible risk factors which vary geographically may be more balanced between cases and controls.

Question 9

What are the disadvantages of neighbourhood controls in case-control studies?
(5)

Answer 9

Costly.

Possibly not representative of study base.

Households as sampling units (increased chance of selection for people living alone).

Difficult to document non-response.

Possible over-matching if exposure is related to residence.

Question 10

What are the advantages of hospital controls in case-control studies?
(5)

Answer 10

Easy to identify controls.

Readily available in sufficient numbers.

More likely to be aware of previous exposures or events than healthy individuals.

Likely to be subject to same selection factors determining hospitals attended.

More likely to cooperate than healthy individuals which reduces bias due to non-response.

Question 11

What are the disadvantages of hospital controls in case-control studies?
(2)

Answer 11

By definition, they are ill:

• Disease may have a common aetiology or be on the causal pathway
• More likely to smoke
• More likely to be from lower socio-economic status
• More likely to be heavy drinkers
• Bias

Question 12

Name 2 minor control types in case-control studies.

Answer 12

Friends.

Relatives.

Question 13

How many control groups should you have in a case-control study?

If multiple, what stipulation must they follow?

Answer 13

One is standard, but you can have more.

Control groups should be independent (i.e. no overlap).

Question 14

Define matching in case-control studies.

Answer 14

When cases are compared with a group of controls who are comparable in other respects other than disease/case status.

Question 15

What are the disadvantages of matching in case-control studies?
(4)

Answer 15

Increased effort to find suitable controls.

Exclusion of cases if no match found.

Longer study duration (control selection delayed until case identified).

Overmatching.

Question 16

Define overmatching in case-control studies.

How will it effect the relative risk?

Answer 16

A matching variable is involved in, or is closely related to the mechanism whereby the exposure influences disease risk:

• Could be on causal pathway
• Could be a surrogate
• Could be just strongly associated with both disease and exposure

Overmatching will bias the relative risk towards 1.

Question 17

Give an example of overmatching.

Answer 17

Study of smoking and lung cancer. Matching on carrying a cigarette lighter or not would be classed as overmatching.

Question 18

What is the general rule for matching in case-control studies?

Answer 18

Only consider matching on risk factors whose confounding effects need to be controlled for, but are not of scientific interest in the study.

Question 19

For matching in case-control studies, what are the design considerations for number of controls per case?
(3)

Answer 19

1:1 matching is fine.

But increasing the number of controls per case increases statistical power.

1: 4 matching is as far as a study should usually go.
- Inefficient beyond this point for the increases in statistical power.

Question 20

Describe a nested case-control study?

Why are they useful?
2

Answer 20

Cases and controls drawn from a cohort study.

Particularly useful to adjust for confounding effect of time.

Also very useful when wanting more detailed measurements on a subset of the cohort.

Question 21

What are the advantages of case-control studies?

5

Answer 21

Several aetiological factors for a single disease.

Is optimal for the study of rare disease.

Well suited for disease with long latency periods.

Relatively quick and inexpensive.

Yields estimates of relative risk.

Question 22

What are the disadvantages of case-control studies?

4

Answer 22

Yields only relative risk (not absolute risk).

Potential for selection bias.

Retrospective collection of information.

Can only look at one disease.