Case Control Studies Flashcards
What is the difference between observational and intervention analytic studies?
Does the exposure increase or decrease the risk of the outcome?
Observational: watch and measure exposure status and what the outcome is
- Cross-sectional, ecological, cohort, case-control studies
Intervention: assign an exposure
- Randomised controlled trials studies
What are case-control studies used for that cohort studies can’t be used for?
- Designed for rare/slow to develop outcomes
- Can efficiently examine acute or transient exposures
Case control vs. cohort studies
Cohort: ascertain exposure status, then find out outcome(s)
- measure exposure status
- follow over time
- compare development of outcome (incidence)
Case-control: ascertain outcome status, then find out exposure(s)
- Identify people with outcome
- Find people without outcome
- Compare exposure likelihood beforehand (odds)
What is the process used for case-control studies?
Looks at odd ox exposure (cases)/Odds of exposure (controls)
1. Identify source population
2. Identify people with outcome (cases)
3. Sample people without outcome for same population (controls)
4. Measure exposure prior to outcome in cases and controls
5. Compare odds of exposure to calculate measure of association (odds ratio)
Why use odds and how can you calculate them and the odds ratio?
We can’t use incidence in this case (because it will be 100 and 1 for each group). Prevalence won’t give as good results as odds.
- Odds are not a measure of occurrence!!!
- Odds ratio IS a measure of occurrence (tells us how many times as likely cases are to have the exposure compared to controls
Odds of exposure in cases (a/c) / Odds of exposure in controls (b/d)
= Odds ratio
- Null value is 1
How do we report/interpret the odds ratio?
The same way we interpret the relative risk (for this course).
Ie. ‘outcome’ were ‘value’ times as likely to have ‘exposure’ than ‘comparison’.
Compare RR to OR
Relative risk (RR):
People with ‘exposure’ are x times as likely to develop the ‘outcome’ as people with ‘comparison’
Odds Ratio (OR):
People with ‘outcome’ are x times as likely to have had the ‘exposure’ than people without the ‘outcome’
What is the major problem with case-control studies and how do we rectify this?
Problem: Controls down have outcome. So when do you measure exposure for controls?
We use index dates, which is when you detect a case you find a control (or many controls) in the population and measure the exposure at the same time as the case (so basically just recording the control group as having the exposure at the same time as the person who developed a case). This date is the index date.
Describe case selection
Usually try to identify incident cases
- sometimes just recruit people with the outcome (prevalent cases)
Case-control studies defined by outcome
- so only one per study
- really important clearly defined and readily identifiable
Need to represent the odds of exposure in people without the outcome in source population.
- these people must be capable of becoming a case
Often select multiple controls per case for statistical power
Must be very careful where you select the controls from as things such as hospital controls can sometimes not give very good/accurate controls and this affects the data
Describe some issues with exposure measurement
Exposure measurement must be comparable:
- Dead cases vs. alive controls
- Interviewers may act differently for cases and controls (because they know the hypothesis)
- Cases trying to work out what made them sick
- Outcome may affect recall ability (eg. if you got an injury due to lack of work safety and then when interviewed you get asked about safety, you will likely be a lot more safety conscious now compared to someone who hasn’t been injured)
What are the strengths and limitations of case-control studies?
Strengths:
- Rare outcomes, transient exposures
- Multiple exposures
- Temporal sequencing
- Often comparatively quick and inexpensive
Limitations:
- Usually can only study one outcome
- Difficult to select appropriate control group
- Can be susceptible to selection and recall bias
