30. DEALING WITH CONFOUNDING

Question 1

1. What 3-step procedure do we need to follow to deal with confounding?

Answer 1

STEP 1:
- identify the potential confounders

STEP 2:
- adjust for the potential confounders

STEP 3:
- compare the Crude and Adjusted Estimates

Question 2

2. What is the aim of the Statistical Adjustment Process?

Answer 2

• it aims at eliminating or reducing the confounding
  effect
• of the potential confounders
• in any exposure-outcome association

IN OTHER WORDS:
- it aims at removing any effects of the Potential
Confounder on the outcome

Question 3

3. What is the best result of using Statistical Adjustment?

Answer 3

• the complete elimination of the Confounding Effect

IF WE CANNOT ELIMINATE IT:
- we at least aim to reduce it

Question 4

4. What effect does Statistical Adjustment Process have on the association?

Answer 4

• it provides a more clean and accurate estimate of the
  Exposure-Outcome Association

Question 5

5. What is another term for “Adjusting”?

Answer 5

• “controlling”

Question 6

6. What happens after we adjust our estimates for a Potential Confounder?

Answer 6

THESE ESTIMATES ARE NOW SAID TO BE:
- “adjusted for that confounder”
- “controlled for that confounder”

EXAMPLES:
- “age-adjusted confounder”
- “smoking-adjusted confounder”

Question 7

7. What does Statistical Adjustment aim to do to the Potential Confounder?

Answer 7

• it aims to keep the Potential Confounder constant
• this no longer makes it a variable
• this is because everyone is now equally exposed to the
  Confounder

Question 8

8. What does Statistical Adjustment re-calculate?

Answer 8

• it re-calculates the estimates for the
  Exposure-Outcome Association
• this gives us more accurate estimates for the
  Exposure-Outcome Associations

IT CAN RE-CALCULATE THE:
- Odds Ratio
- Mean Difference
- Regression Coefficient
- etc.

Question 9

9. What is one way of adjusting the given confounder?

Answer 9

• we can Stratify the analysis
• we base this on the categories of the Confounder

Question 10

10. What does the process of Stratification involve?

Answer 10

• we perform the analysis for the Exposure-Outcome
  association separately
• within the categories of the Confounder

IN OTHER WORDS:
- we divide the exposure and the outcome by the levels
of the Confounder

Question 11

11. What is the result of Stratification?

Answer 11

• we now have 2 (or more) different, stratified estimates
  for the exposure-outcome association

THERE IS AN ESTIMATE:
- for each category of the Confounder variable

Question 12

12. What happens once we have our stratified estimates for the Exposure-Outcome Association?

Answer 12

• these estimates need to be combined
• this gives us a Combined (Adjusted) Estimate

Question 13

13. What is this “Adjusted” Estimate?

Answer 13

• this estimate is clear of any Confounding effects
• these effects come from the specific Confounder

THIS IS BECAUSE:
- the analysis was performed separately in its categories
- the results were then combined

Question 14

14. Define Effect Modification.

Answer 14

THIS IS WHEN THE EXPOSURE-OUTCOME ASSOCIATION IS:
- substantially different in the categories of the
confounder
- the estimates cannot be combined
- they need to be presented separately