Confounding

Question 1

Define confounding

Answer 1

A fundamental methodological concern in
epidemiology, as well as other branches of science such as economics, psychology, and sociology that rely largely on observational data.
the term to describe the distortion of the estimate of The effect of an exposure of interest because it is mixed with the effect of an
extraneous factor, an independent variable that varies systematically with
the hypothetical “causal” variable understudy.

Question 2

Define confounding factor

Answer 2

A characteristic that can cause or prevent the disease and is
also associated with the exposure of interest.

Question 3

What are the 3 criteria of a confounder in the a priori approach?

Answer 3

a) a confounder must be causally or noncausally associated with the exposure in the source population (study base)
being studied;
b) a confounder must be a causal risk factor (or a
surrogate measure of a cause) for the disease in the
unexposed cohort; and
c) a confounder must not be an intermediate cause (in other words, a confounder must not be an intermediate step in the causal pathway between the exposure and the disease)

Question 4

What is the definition of confounding according to the mixing of effects approach?

Answer 4

Confusion, or mixing, of
effects; the effect of the exposure is mixed
together with the effect of another variable,
leading to bias”

Question 5

What is the definition of a confounder according to the data-based approach?

Answer 5

a) the effect measure is homogeneous across the strata defined by the confounder and
b) the crude and common stratum-specific (adjusted)
effect measures are unequal (this is called “lack of
collapsibility”)
• Usually evaluated using 2x2 tables, and simple
stratified analyses to compare crude effects with
adjusted effects

Question 6

How are confounders evaluated according to the data-based approaches?

Answer 6

Stratify the 2x2 table by confounders. Calculate the OR for each stratum. If stratum-specific ORs are similar, calculate the adjusted OR (using M-H test of homogeneity).

Question 7

Using the data-based approach, when calculating the stratified OR, how does comparison with the crude OR determine confounding?

Answer 7

If crude RR/OR is about equal to the adjusted RR/OR, confounding is unlikely. If the crude OR is not equal to the OR, with a change in estimate is >10%, confounding is likely.

Question 8

What is the formula for relative risk (RR)?

Answer 8

Risk in exposed/risk in unexposed [a/(a+b)/(c/(c+d))]

Question 9

Formula for OR

Answer 9

ad/bc

Question 10

DAG definition of confounder

Answer 10

Standard definition: “C” is a confounder if it is 1) associated with the exposure; 2) associated with the outcome among the
unexposed; 3) is not a downstream consequence of exposure
or outcome.

Question 11

T or F: If there are common causes of exposure and
outcome, there is no way to infer a causal effect in
the study. Assuming that there is no unmeasured
confounding.

Answer 11

False, confounder can be blocked

Question 12

T or F: Selection bias results from conditioning on a

common cause of exposure and outcome

Answer 12

False, this is a confounder, selection bias results from common effect of exposure and outcome

Question 13

T or F: If there is no confounding, there will always be no

selection bias

Answer 13

False, 2 different concepts

Question 14

T or F: Selection bias can only occur in a study where the measure of interest is cumulative incidence.

Answer 14

false, not dependent on association measurements

Question 15

What are 3 methods of control at the design stage?

Answer 15

Randomization, Restriction, Matching

Question 16

What is one conventional approach and one newer approach to control confounding at the analysis stage?

Answer 16

Conventional approaches
• Stratified analyses
• Multivariate analyses
Newer approaches
• Graphical approaches using DAGs
• Propensity scores
• Instrumental variables
• Marginal structural models

Question 17

Name 2/4 disadvantages of restriction to limit confounding?

Answer 17

1. Will limit number of eligible subjects
2. Will limit ability to generalize the study findings
3. Residual confounding may persist if restriction
   categories not sufficiently narrow (e.g. “decade of
   age” might be too broad)
4. Not possible to evaluate the relationship of interest at different levels of the confounder

Question 18

Incidence rate difference formula

Answer 18

Risk in exposed - risk in unexposed = a/a+b - c/c+d