Case-Control

Question 1

What is a Case-Control Study?

Answer 1

• •Design Characteristics

–Select participants as cases and controls

–Data on exposure collected just after group allocation

Question 2

Inclusion of Incident vs. Prevalent cases in case-controls

Answer 2

• Prevalent cases more easily attainable
• With prevalent cases, potential for factors being related to survival or duration of disease, not causal
• Survival prior to disease could be an issue for incident cases too
• Interpretation of exposure preceding disease

Question 3

Selection of controls in case-control

Answer 3

• Want controls as similar to cases on all factors other than the exposure of interest
• Minimize biases: selection bias, confounding, information bias
• Using these principals: study base principal, deconfounding, comparable accuracy
• Study efficiency and cost also considerations

Question 4

Issues in Ascertainment of Cases in case-control

Answer 4

• Well‑defined diagnostic criteria, representative of all cases
• Sources (hospital, general population)
• Incident or prevalent cases

Question 5

Issues in Selection of Controls in case-control study

Answer 5

• Theoretical considerations (representative of the reference population; matched or comparable for certain variables)
• Practical considerations (feasibility and cost)
• Sources of controls
  
  • population of defined area
  • hospital patients
  • neighbors
  • siblings, spouses or other relatives
  • friends or associates of cases

Question 6

Handling Potential Confounding Factors in case-control study

Answer 6

• In the process of selecting controls
  
  • Matching
    
    • stratification (group or frequency matching)
    • pairing ‑‑ individual matching
  • In the data analysis
    
    • Stratification
    • Adjustment

Question 7

pros of case-control study

Answer 7

• –Cheap, quick and easy

–Multiple Exposures

Question 8

cons of case-control study

Answer 8

• Rare exposures

Limited number of outcomes

Question 9

Biases in case-control study

Answer 9

• –Temporal bias

–Recall bias

–Selection bias
* interviewer bias

Question 10

Measure of association in case-control

Answer 10

odds ratio = (a/c) / (b/d)