ep test 2c
Terms
undefined, object
copy deck
- Case-control study design
- Retrospective, observational design Determine group (case/control) membership on the basis of whether an individual meets a case definition (i.e., has the outcome of interest)
- Case-control study goal
- Goal is to examine or test the relationship between specific determinant(s) or exposures and case status
- Case-control (ex)
- Research question – is aspirin use, especially in children, associated with development of Reye’s syndrome? Clinical trial not ethical Cohort study possible, but inefficient because of rare outcome CaCo is best choice
- Case-control (ex2)
- Research question – is there a relationship between high-dose oral contraceptives and breast cancer? Again, CaCo best option For most cancer studies (excluding treatment of), CaCo employed due to ethical reasons and long latency period of CA
- Selection of CaCo Cases
- Source population should be well-defined Strict case definition or diagnostic criteria Inclusion / exclusion criteria If possible, cases should be representative of all cases Convenience sample vs. population-based Incident vs. prevalent (Inclusion/exclusion criteria will ensure there are clear associations) Sampling cases from convenience sample (e.g., hospital) may not be representative of all cases Incident cases preferred If exposure more common in incident assoc. w/ development of disease If exposure more common in prevalent assoc. w/ duration and/or development Survivorship bias is not an issue with incident cases, but is a problem with prevalent cases Restriction to improve validity
- Selection of Controls
- Controls represent the population of non-diseased persons (or those w/out outcome) who would have been included as cases had they developed the disease Must be chosen from source population that produced cases (i.e., same baseline risk of developing the outcome) Source population can be defined before cases appear by a geographical area or some other identifiable entity (e.g., hospital) Start with cases and then attempt to identify hypothetical cohort that gave rise to them Selection MUST NOT depend on exposure Single or multiple control groups Number of controls : cases can vary from 1:1 to 4:1 Controls can be matched by group characteristics (frequency-matched) or paired (individually matched) Match to control for confounders MUST be a strong confounder
- Matching
- Can control only a limited number of confounders Potentially expensive in $$ and time Matched factor cannot be tested as to its effect on the outcome Useful strategy for environmental / genetic factors Useful when number of cases is small and multiple controls can be matched
- Strengths of CaCo Design
- (Generally) less expensive and time consuming than cohort design Efficient for studying rare diseases (or outcomes) (don’t have to wait) Efficient for studying diseases with long latency periods (e.g., chronic diseases) Don’t have to follow subj’s for long time for study Possible to study many different exposures with respect to outcome of interest
- Weaknesses of CaCo Design
- Temporal relationship difficult to define (causal inference less clear) Risk or rate of outcome cannot be estimated directly Insufficient for studying rare exposures Particularly susceptible to both selection and information biases
- Primary CaCo Design Concerns
- Selection Bias cases and controls are not selected from the same source population Information Bias bias in the measurement of exposure resulting in misclassification recall bias detection bias
- Analysis & Interpretation
- Assessing bias selection, recall and non-response Assessing potential confounders Data-derived hypotheses more suspect than a priori hypotheses Calculation of Odds Ratio and its Confidence Interval Can also compare means between cases and controls