Epidemiology 2200B - Observational Studies Detailed Notes
Epidemiology 2200B - Introduction to Epidemiology
Page 1
- Course Title: Epidemiology 2200B
- Course Type: Introduction to Epidemiology
- Instructor: Dr. MK Campbell
- Date: February 11, 2025
- Institution: Western
Page 2
Topics to be covered today (Feb. 11):
- Case Reports and Case Series
- Ecological Study (can be descriptive or analytic)
- Cross-sectional studies (can be descriptive or analytic)
- Case-control studies
Next class (Feb 25):
- Cohort studies
Mid-Term Examination: Will be discussed at the end of the class.
Page 3
- Hierarchy of Evidence:
- Common figure often referenced - weight of evidence based on study design and conduct.
- Important to think of it as a hierarchy of bias risk that needs to be controlled.
Page 4
1. Case Reports and Case Series
- Case Reports:
- Describe “new” illnesses observed, often in a single practice.
- No defined population.
- Suggest common features among patients.
- Alerts health professionals about similar cases.
Page 5
Hypothesis Generation from Case Reports
- Example:
- MMWR (1981): Five cases of Pneumocystis pneumonia among healthy young men in LA with lab-confirmed cytomegalovirus and candidal infections.
- Led to early investigations proposing drug-related causes.
- This built foundation for recognizing AIDS in 1982 with subsequent CDC reports clarifying HIV as the causative agent by 1984.
Page 6
Case Reports and Case Series: Uses & Limitations
Uses:
- Suggest common features and generate hypotheses regarding etiology.
Limitations:
- Lack of comparison group.
- External validity challenges.
- Potential confounding variables affecting associations.
Page 7
2. Ecological Studies
- Focuses on groups/populations rather than individuals.
- Example: exposure measured at population level, e.g., BCG vaccine coverage and COVID-19 mortality.
Page 8
Group-level Exposure & Outcome
- Research shows countries with high BCG vaccination rates have lower COVID-19 mortality rates.
Page 9
Ecological Fallacy
- Misinterpretation occurs when assuming ecological associations apply at the individual level.
- Important to align target inferences with the level of analysis to avoid this fallacy.
Page 10
Rationale for Ecological Studies
- Low cost and convenient.
- Hypothesis generation.
- Useful when variation within groups is minimal.
Page 11
3. Cross-Sectional Studies
- Provide a “snapshot” in time; can be descriptive or analytic.
- Collect data on multiple characteristics at a single point in time.
- Example: link between physical activity and Coronary Heart Disease (CHD).
Page 12
Cross-Sectional Study Design
- Population or sample selection (often random).
- Data on exposure and outcomes collected simultaneously.
Page 13
Data Presentation Methodology
- Can present findings in a 2 by 2 table to estimate exposure and disease prevalence among groups.
Page 14
Example Analysis
- Analyzing activity vs. CHD prevalence; cannot assume etiology due to cross-sectional nature.
Page 15
Cautions in Cross-Sectional Studies
- Inability to infer causality due to simultaneous assessment of exposure and disease; confusion about the direction of associations.
Page 16
4. Case-Control Studies
- In-depth analysis of design, biases, matching, specifics on when to use.
Page 17
Case-Control Design Overview
- Sample based on disease presence; look back for prior exposure.
Page 18
Cases vs. Controls
- Cases: individuals with the health outcome; requires clear disease definition.
- Controls: representative subjects reflective of exposure history.
Page 19
Case-Control Design Table Example
- Select cases and controls; measure past exposure for the study outcomes.
Page 20
Classic Example: Smoking and Lung Cancer Study (Doll and Hill, 1952)
- First case-control study focusing on lung cancer and smoking associations.
Page 21
Results and Conclusions
- Findings suggest high smoking prevalence among lung cancer patients but may not infer incidence due to design limitations.
Page 22
Dose-Response Relationship
- Critical to investigate exposure levels and their effects in any epidemiological study.
Page 23
Presentation of Non-Exposed Group Information
- Analysis of the control groups in terms of exposure characteristics.
Page 24
Information Bias in Case-Control Studies
- Differing exposure data acquisition can lead to biases and inaccuracies in study results.
Page 25
Selection Bias Explanation
- Occurs when the exposure history for participants does not represent the general population.
Page 26
Case Source Selection Issues
- Differing participation rates and characteristics can introduce bias; hospitals and disease registries count.
Page 27
Prevalent vs. Incident Cases
- Distinction crucial for identifying disease risk factors and participant timing considerations.
Page 28
Controls Representation
- Controls should reflect the typical exposure experience of the population.
Page 29
Unique Potential Biases from Various Control Sources
- Hospitalized vs. non-hospitalized control characteristics can alter exposure assessments significantly.
Page 30
Information Bias Types
- Quality of collected data from interviews vs medical records; risk of recall bias.
Page 31
Minimizing Recall Inaccuracies
- Supplement recall data, use standardized interviews, and manage subject knowledge.
Page 32
Matching in Case-Control Studies
- Importance of matching to control for confounding variables affecting outcomes.
Page 33
Group Matching Challenges
- Ensuring proportional representation can be difficult; need for similar characteristic distributions.
Page 34
Individual Matching Approach
- Matching controls during case selection can potentially balance demographic disparities.
Page 35
Challenges of Matching and Limitations
- Limitations in factors matched on and their impact on study outcomes.
Page 36
Multiple Controls Usage
- Use for enhanced sample size and reliability in findings; comparisons with diverse control types.
Page 37
Case-Control Study Design Considerations
- Addressing when case-control studies are most warranted, especially for rare conditions.
Page 38
Case-Crossover Design
- Use for acute events where individual subjects serve as their controls for previous exposure comparison.
Page 39
Case-Crossover Design Methodology
- Reliant on accurate recall over different periods pre-incident.
Page 40
Case-Crossover Design Example Visualization
- Comparative timelines displayed in controlled study formats.
Page 41
Example: Epidemic of Blindness Exploration
- Early case reports initiated subsequent studies on new conditions.
Page 42
Natural History of RLF Study Methodology
- Tracking premature infant health over time to correlate risk factors.
Page 43
Ecological Study on RLF Treatments
- Treatment protocols showed associations with rates of RLF in hospital settings.
Page 44
Case-Control Study Insights on RLF
- Data comparison showed nursery stays linked to RLF outcomes while still questioning etiology.
Page 45
Experimental Approach to RLF
- 1953-54 trials to assess oxygen levels furnished crucial evidence on RLF incidence reduction.
Page 46
Post-Trial Outcome Implications
- Recommendations shifted clinical practices dramatically curtailing RLF incidence.
Page 47
Continuing Relevance of ROP
- Discussions surrounding ROP emphasize the need for careful treatment protocols for premature infants.
Page 48
Questions and Comments Session
- Open floor for student discussions and inquiries.
Page 49
Mid-Term Examination Details
- Review of exam materials and protocols for upcoming tests.
Page 50
Exam Format Preview
- Upcoming exam structure, mix of questions, and content coverage outlined for students.