Confounder & Effect Modifiers

Question 1

What are the three characteristics of a confounder?

Answer 1

It is a variable that is:

1. Associated with the outcome
2. Associated with the exposure
3. Not on the causal pathway

Question 2

How do you control for confounding?

Answer 2

• Matching
• Restricting
• Randomisation
• Stratification

Question 3

What is matching?

Answer 3

Finding confounders in the design stage and evenly distributing them in each group.

Question 4

What is restricting?

Answer 4

Restricting entry into the study i.e. the inclusion criteria is limited to individuals that fall outside of confounding factors.

Question 5

What is randomisation?

Answer 5

Where the study population is randomly distributed into their group in the design stage. The idea is that this will control for known and unknown confounders.

Question 6

What is stratification?

Answer 6

This is a statistical method that divides the study population into strata (subgroups) according to their level of potential confounding factor. This is done in the analysis stage and works best if there is not a lot of strata.

Question 7

What is an effect modifier?

Answer 7

Not an independent risk factor for the outcome but changes the association between study and outcome factors.

Question 8

Is confounding present when the rate is < 10% ?

Answer 8

No.

Question 9

Is confounding present when the rate is > 10% ?

Answer 9

Yes.

Question 10

You are interested in studying whether eating pizza once a week for 10 years is associated with the development of lung cancer in a region that just implemented stop smoking advertisements on TV. Stop smoking advertisements is not an independent risk factor for lung cancer. The measure of association between pizza eating and lung cancer is different in regions with and without stop smoking advertisements. Stop smoking advertisements is most likely to be what sort of a variable in this scenario?
Select one:
a. Outcome factor
b. Effectmodifier 
c. Confounder
d. Study factor

Answer 10

b. Effectmodifier

Question 11

You are interested in studying whether eating pizza once a week for 10 years is associated with the development of lung cancer in a region that just implemented stop smoking advertisements on TV. Smoking is a known independent risk factor for lung cancer. Smoking is most likely to be what sort of a variable in this scenario?
Select one:
a. Outcome factor
b. Study factor
c. Effect modifier 
d. Confounder

Answer 11

d. Confounder

Question 12

What does an adjusted rate tell us?

Answer 12

• Whether confounding is present, by comparing the adjusted rate with the crude rate
• How much of the study factor is associated with the outcome factor

Question 13

What is a crude rate also known as?

Answer 13

Unadjusted rate.