Confounding an Effect Modification

Question 1

at what percentage is a factor considered to be a confounder and what is the percentage comparing?

Answer 1

15% and it’s comparing the crude and adjusted (average) OR/RR

Question 2

Describe the relationship that confounders have with the exposure and outcome of interest

Answer 2

1. (Independently) associated with the exposure
2. (Independently) associated with the outcome
3. not directly in the causal-pathway that links exposure to outcome

Question 3

how do you assess for Confounding?

Answer 3

1. calculate the crude OR/RR (aka unadjusted association)
2. calculate outcome OR/RR for each individual strata
3. Compare the crude vs. adjusted measures of association between exposure and outcome

Question 4

at what percentage is a factor considered to be causing effect modification and what is the percentage comparing?

Answer 4

15% and its comparing the lowest and highest strata layers of the 3rd variable

Question 5

How do you assess for effect modification?

Answer 5

1. Calculate crude OR/RR between exposure and outcome
2. Calculate strata-specific OR/RR between exposure and outcome for EACH strata
3. Compare strata specific OR/RR between each other, while also referencing the adjusted OR/RR

Question 6

Describe Confounding and how it impacts measures of association

Answer 6

a 3rd variable (related to the study subjects) that distorts an association between the exposure and the outcome

It impacts the Magnitude and direction

Question 7

Describe Effect Modification and how it impacts measures of association

Answer 7

a 3rd variable that modifies the magnitude of effect of a TRUE association by varying it within different strata

Impacts by inaccurately representing the group, and instead representing a combination of strata as one (incorrectly)

Question 8

What is residual confounding, and where might it arise in a study?

Answer 8

Residual confounding occurs when attempting to use the restriction method to prevent confounding in a study, yet failing to create sufficiently narrow restricting criteria. This allows other confounding variables to be included in the subjects, and impact the measures of association in the study.

Question 9

What are the 2 main stages of a study where researchers can attempt to control for confounding?

Answer 9

Study Design Stage

Analysis of Data Stage

Question 10

what are the specific means one can use to control for confounding in the Study Design Stage?

Answer 10

Randomization: “hopefully” selects subjects for each group that represent equal numbers of subjects with and without the confounding variable

Restriction: does not allow subjects into the study that are affected by the suspected confounder

Matching: “matching” a confounder and non confounder subject among each study group in order to reach an equal distribution of the confounding variable in the study

Question 11

what are the specific means one can use to control for confounding in the Analysis of Data Stage?

Answer 11

Stratification (with weighting): evaluating the association between exposure and outcome within the different strata in the confounding variable

Multivariate statistical analysis: mathematically factoring out the effects of the confounding variable.

Question 12

List the Strengths and Weaknesses of the Randomization process

Answer 12

Strengths: sufficient sample sizes (the larger the better) it is likely to be successful

Weaknesses: does not guarantee success due to potentially unequal allocation of known and unknown confounders between groups

Question 13

List the Strengths and Weaknesses of the Restriction process

Answer 13

Strengths: simple, convenient, and inexpensive. Also does not negatively impact the internal validity of the study

Weaknesses: may reduce the sample size too much, and if the restriction criteria is softened it may allow residual confounders into the study

Question 14

List the Strengths and Weaknesses of the Matching process

Answer 14

Strengths: intuitive and provides analytic efficiency

Weaknesses: difficult to accomplish and doesn’t control for unidentified confounders. Can lead to “over-matching” which blunts the findings of the study.

Question 15

List the Strengths and Weaknesses of the Stratification process

Answer 15

Strengths: intuitive, and enhances the understanding of the data

Weaknesses: impractical for simultaneous control of multiple confounders, especially confounders with multiple strata within each variable.

Question 16

List the Strengths and Weaknesses of the Multivariate

Statistical Analysis process

Answer 16

Strengths: can simultaneously control for multiple confounding variables. In statistical regressions, OR’s can be obtained and interpreted

Weaknesses: is time consuming and can be difficult to understand/write about clearly

Question 17

Give a simplified definition of “confounding variable”

Answer 17

an alternative explanation for the association