lecture 11 & 12

Question 1

what are the ways to assure ourselves that there is a true association between exposure and outcome

Answer 1

• confounding or effect modification (interaction)
• check for bias
• check for statistical significance

Question 2

define confounding variable

Answer 2

A 3rd variable that distorts an observed relationship between the exposure and the outcome

Question 3

what are ways that confounders have impact

Answer 3

• intensity/magnitude/strength

* direction

Question 4

tests for confounding steps

Answer 4

1. calculate crude outcome measure of association (OR/RR) between exposure and outcome
2. re-calculate outcome measure of association (OR/RR) between exposure and outcome while statistically controlling the effects of the confounders
3. compare the crude vs. Adjusted

Question 5

what percentage differences decides if there is a confounding variable present

Answer 5

20%

Question 6

what is the purpose for controlling for confounders

Answer 6

To get a more precise estimate of the true association between the exposure and the disease/outcome

Question 7

ways to control confounding in the study design stage

Answer 7

Randomization (blocked or stratified)
restriction
matching

Question 8

ways to control confounding in the analysis of data stage

Answer 8

stratification

multivariate statistical analysis

Question 9

what is the strength of randomization in controlling for confounding

Answer 9

with sufficient sample size (N), randomization will likely be successful in serving its purpose

Question 10

what are the weaknesses of controlling for confounding

Answer 10

• sample size may not be large enough to control for all known and unknown confounders
• randomization process doesn’t guarantee successful, equal allocation between all intervention groups for all known and unknown confounders
• practical only for interventional studies

Question 11

what are the strengths of restriction of controlling for confounding

Answer 11

• straight forward, convenient and inexpensive

* does not negatively impact internal validity

Question 12

what are the weaknesses of controlling for confounding

Answer 12

• sufficiently narrow restriction criteria may negatively impact ability to enroll subjects
• if restriction criteria is not sufficiently narrow it will allow the introduction of residual confounding effects
• eliminates researchers ability to evaluate varying levels of the factor being excluded
• can negatively impact external validity (generalizability)

Question 13

what are the strengths of matching

Answer 13

*intuitive, some feel it gives greater analytic efficiency

Question 14

what are the weaknesses of matching when controlling for confounding

Answer 14

• difficult to accomplish, very time consuming and expensive
• doesn’t control for any confounders other than those matched on (over-matching possible=will mask findings)

Question 15

what are the strengths of stratification for controlling for confounding

Answer 15

intuitive (to some), straight-forward and enhance understanding of the data

Question 16

what is the strength of stratification when controlling for confounding

Answer 16

impractical for simultaneous control of multiple confounders, especially those with multiple strata within each variable being controlled

Question 17

what is the strength of using multivariate analysis when controlling for confounding

Answer 17

• can simultaneously control for multiple confounding variables
• in logistic regression, beta coefficients can be directly converted to ORs

Question 18

what are the weaknesses of using multivariate analysis

Answer 18

process can cause some individuals to not clearly understand the data

Question 19

what is effect modification

Answer 19

a 3rd variable, that when present, modifies the magnitude of effect of an association by varying it within different levels of a 3rd variable

Question 20

what are the steps for testing for effect modification

Answer 20

1. Crude outcome measure of association between exposure and outcome (OR/RR)
2. Calculate crude outcome measure of association (OR/RR) between exposure and outcome for each strata (layers) of the effect-modifying variable
3. Compare the stratum-specific measures of associations (for each strata of the 3rd variable) between the exposure and outcome (OR/RR)