Module 5: Critical Thinking: Questioning Assumptions And Evaluating The Evidence

Question 1

Characteristics of a confounder

Answer 1

1. Variable must be associated with the exposure of interest
2. Variable must be an independent risk factor for the outcome of interest
3. Variable must not be on the causal pathways between exposure of interest and outcome of interest

Question 2

Possible effects that a confounder may have on the measure of association

Answer 2

1. Move towards the null
   Eg:
   True RR = 1.4
   Study RR = 1.1
2. Move away from the null
   Eg:
   True RR = 1.1
   Study RR = 1.4
3. Reverse (Simpsons paradox)
   Eg:
   True RR = 0.8
   Study RR = 1.2

Question 3

Ways to reduce confounding at the design phase

Answer 3

1. Restriction
2. Randomisation
3. Matching

Question 4

Ways to adjust for confounding during the analysis phase

Answer 4

1. Stratification
2. Standardisation
3. Regression models/multi variable analysis

Question 5

Outline the 3 steps for recall bias

Answer 5

Step 1: Who does the bias effect?
Step 2: How does the bias effect them?
Step 3: What effect does this have on the measure of association?

Question 6

Why is it important to have an adequate sample size in a study?

Answer 6

An adequate sample size increases the probability of finding an association if one truly exists