Case-Control Studies Flashcards
(17 cards)
What is a case-control study?
Observational study.
Group of individuals with a condition (outcome) are identified (cases), along with a group of individuals who do not have the condition (controls) and their exposures are identified and compared.
Notable: SIDS study in Tasmania
What is the definition of a control in a case-control study?
A group of people without an outcome of interest, matched to the cases on certain characteristics.
How are data collected in a case-control study?
- Interview with case/proxy and control
- Record review
- Linkage
What is the most challenging issue in the design of a case-control study?
Selection of controls.
Ideally, how should controls be selected in a case-control study?
- Same source population
- Matched on key variables e.g. age, sex, SES
What are some possible sources of controls in case-control studies?
- Population-based controls e.g. electoral role
- GP registers
- Neighbourhood - overmatching possible
- Hospital - efficient but biased as unhealthy population, more risk factors
- Friends / family - overmatching
What are key sources of bias in a case-control study?
- Recall bias - cases more likely to recall»_space; overestimation of association
- Interviewer bias - if aware of case/control, questions may be asked in different way
- Selection bias - if selectors have knowldege of exposure status
- Ascertainment bias - likelihood of detecting disease varies depending on exposure e.g. if exposed more likely to investigate for disease»_space; overestimation
- Confounding - systematic differences between groups
What is a nested case-control study?
Cases identified from within a cohort study and then matched to controls within same cohort study – can match more effectively
What is the case-case method?
Compare cases of different diseases (e.g., Salmonella and Campylobacteria) to explore exposures and associations – minimises recall bias
What is the case-crossover method?
Cases are own controls; transient exposures/acute events
How are case-control studies analysed?
Odds ratio (OR).
The odds of exposure in cases vs. controls.
Cannot calculate other measures of association as there is no clear denominator / population at risk
What are the advantages of case-control studies?
- Good for rare outcomes, long-latency
- Exposures: common; multiple
- Complexity: simple
- Cost: cheap
- Time: quick
- Ethics: sound
- Can do nested case-control within cohort - improves matching
What are disadvantages of case-control studies?
- Main: selecting/matching controls
- Exposure: not suitable for rare exposures
- Outcomes: N/A
- Bias: recall and selection bias
- Confounders: yes
- Temporality: difficult to establish
- Causality: associations only
- Analysis: cannot determine incidence / RR
What are the key points about case-control studies?
- Control selection
- Exposure measurement: recall bias
- Don’t interpret as incidence / risk
What is an acceptable ratio for cases:controls?
1:1 but generally, more controls the better.
However beyond 1:4, the additional study power gained is limited.
What tool can be used to appraise case-control studies?
ROBINS-E (Cochrane)
How are the design and methods of case-control studies appraised?
- Clearly defined research question (PICO)
- Pre-published protocol
- Reporting guideline: STROBE
- Sample size calculation / power
- Study population base - time period, location, age-restriction etc.
- Selection of cases / controls
- Response rates (cases / controls)
- Misclassification
- Measurement of exposure
- Measurement and adjustment for confounder
- Analysis - appropriate; as per protocol
- Interpretation - potential bias, limitations
- Funding, conflicts of interest
