4. Confounding

Question 1

What is confounding

Answer 1

Confounding is a distortion (inaccuracy) in the estimated measure of association that occurs when the primary exposure of interest is mixed up with some other factor that is associated with the outcome.

Question 2

3 types of confounding

Answer 2

Question 3

Residual confounding

Answer 3

Residual confounding occurs when adjustment is not sufficiently fine to take into account the full variability of the outcome under study,

or when a confounding variable remains unaccounted for.

• There were additional confounding factors that were not considered.
• Control of confounding was not tight enough. or classification was too simple in data collection phase.
• There were many errors in the classification of subjects with respect to confounding variables.

Question 4

Mantel-Haenszel Method for Rate Ratios

Answer 4

weighted value of all stratums

Question 5

Mantel-Haenszel Method for Odds Ratio

Answer 5

Question 6

Control Strategies for Confounding

Answer 6

•Study Design:

1. Randomization
2. Restriction
3. Matching

•Analysis:

1. Stratification
2. Adjustment based on stratification\

• Standardization
• Mantel-Haenszel methods

3. Regression modeling

Question 7

things to consider when assessing confounding

Answer 7

1. Likelihood that confounding occurred
2. Magnitude of the confounding
3. Direction of the confounding shifts the risk estimate