Observational Studies Summary
Case Reports and Case Series
- Simplest study designs.
- Objective: detailed description of cases.
- Key for hypothesis generation.
- Limitations:
- No comparison group.
- Limited generalizability due to biased case selection.
Ecologic Studies
- Examine group characteristics to determine associations.
- Ecologic fallacy: Ascribing group characteristics to individuals.
- Grouped data available (e.g., country-level).
- Can suggest promising research avenues for etiologic relationships.
- Do not demonstrate conclusive associations.
- Limited variability can make associations more valid.
Cross-Sectional Studies
- Exposure and outcome assessed simultaneously.
- Provide a snapshot of the population at a single point in time.
- Identify prevalent cases.
- Limitations:
- Temporal bias (reverse causality).
- Survival/selection bias (prevalence-incidence bias).
- Hypothesis generation.
- Two analysis approaches:
- Compare disease prevalence in exposed vs. unexposed.
- Compare exposure prevalence in diseased vs. non-diseased.
- Useful for evaluating trends in disease prevalence over time and validating self-reported data.
Case-Control Studies
- Identify cases (with disease) and controls (without disease).
- Assess past exposure in both groups.
- If an association exists, exposure history should be higher in cases.
- Cannot estimate disease prevalence directly.
- Selection and Information biases are potential issues.
Potential Biases in Case-Control Studies
Selection Bias
- Sources of Cases:
- Hospital patients: risk factors may be specific to the hospital.
- Community registries: can be valuable sources.
- Eligibility criteria should be specified in writing prior to the study.
- Incident vs. Prevalent Cases:
- Incident cases are preferred.
- Prevalent cases may be related to survival rather than disease development.
- Selection of Controls:
- Controls should be representative of the population from which cases are selected.
- Matching the cases and controls can help reduce this type of bias.
- Limitations in Recall: affects all subjects, leading to exposure misclassification.
- Recall Bias: Cases may over-recall exposures compared to controls.
* Known as the "rumination bias"
Other issues in Case-Control Studies
Matching
- Selecting controls similar to cases in certain characteristics (age, race, sex, etc.)
- Types:
- Group matching (frequency matching).
- Individual matching (matched pairs).
- Problems:
- Difficult to find appropriate controls with too many characteristics.
- Cannot study matched characteristics.
- Overmatching occurs with unplanned matching on variables related to the exposure.
Use of Multiple Controls
- Same Type:
- Increase statistical power.
- Gain noticeable increase in power up to a ratio of about 1 case to 4 controls
- Different Types:
- Explore alternate hypotheses and take into account potential biases.
When is a Case-Control Study Warranted?
- Useful as a first step when searching for a cause of an adverse health outcome.
- Valuable when the disease being investigated is rare.
Case-Crossover Design
- Used for acute outcomes with transient exposures.
- Each case serves as its own control.
- Compares exposure level during an at-risk period to a control period.
- Inherently matches variables that do not change over time.