The document discusses the concept of confounding in epidemiological studies. There are three key points:
1) Confounding occurs when a third variable influences both the exposure and outcome, creating a spurious association between the two. Smoking is a confounding variable for the relationship between coffee drinking and pancreatic cancer risk.
2) There are three criteria for confounding: the factor must be associated with both the exposure and outcome, be unevenly distributed among exposure groups, and not be on the causal pathway between exposure and outcome.
3) Methods to control for confounding include matching, stratification, randomization, statistical adjustment like regression, and ensuring a large effect size overwhelms confounding. Confound
2. Introduction
• Confounder (also confounding variable, confounding factor,
or lurking variable) is a variable that influences both the dependent
variable and independent variable, causing a spurious association.
• Confounding is a causal concept, and as such, cannot be
described in terms of correlations or associations.
• When a non-casual association is observed between a given
exposure and outcome is as a result of the influence of a third
variable, it is termed confounding, with the third variable termed a
confounding variable.
• A confounding variable is causally associated with the outcome
of interest, and non-causally or causally associated with the
exposure, but is not an intermediate variable in the causal pathway
between exposure and outcome (Szklo & Nieto, 2007).
3. There are three conditions that must
be present for confounding to occur
• The confounding factor must be associated with
both the risk factor of interest and the outcome.
• The confounding factor must be distributed
unequally among the groups being compared.
• A confounder cannot be an intermediary step in the
causal pathway from the exposure of interest to the
outcome of interest.
6. Confounding example
Demonstrates that coffee (exposure) was associated with an increased risk of
developing pancreatic cancer (disease) with the dark arrow. A third factor, smoking,
which is a confounder is actually correlated with an increased risk of developing
pancreatic cancer (light arrow). Coffee was spuriously correlated with increasing the
risk of developing pancreatic cancer and once the confounding variable, smoking is
taken into account the correlation between coffee and pancreatic cancer disappears.
It should be noted that both smoking and coffee are also both correlated for
confounding to occur (two-way arrow).
7. Controlling of confounding
• Case-control studies : Assign confounders to both
groups, cases and controls, equally.
• Cohort studies: A degree of matching is also possible
and it is often done by only admitting certain age
groups or a certain sex into the study population,
creating a cohort of people who share similar
characteristics and thus all cohorts are comparable in
regard to the possible confounding variable.
• Double blinding: Conceals from the trial population and
the observers the experiment group membership of the
participants.
8. Controlling of confounding
• Randomized controlled trial: The study population
is divided randomly in order to mitigate the
chances of self-selection by participants or bias by
the study designers.
• Stratification: for small number of confounders.
9. Types
• Operational confounding: Occur in both experimental and
non-experimental research designs. This type of
confounding occurs when a measure designed to assess a
particular construct inadvertently measures something else
as well.
• Procedural confounding: Occur in a laboratory experiment
or a quasi-experiment. This type of confound occurs when
the researcher mistakenly allows another variable to change
along with the manipulated independent variable.
• Person confounding: Occurs when two or more groups of
units are analyzed together (e.g., workers from different
occupations), despite varying according to one or more
other (observed or unobserved) characteristics (e.g.,
gender).
10. Effects of Confounding
• May cause an overestimate of the true association
(positive confounding) or an underestimate of the
association (negative confounding).
11. Preventive steps
• Randomization is the best way to reduce the risk of
confounding.
• Stratification and statistical adjustment can reduce
the risk of confounding.
• Use of propensity scores, in which potential
confounders are used to build a statistical model
that assigns to each person a number called their
propensity score: the people with high scores are
more likely to have certain confounders, and those
with low scores are less likely.
• A very large effect size can outweigh the combined
effects of plausible confounders.
12. Statistical methods
• Small number of potential confounders:
Stratification
• Larger number of potential confounders :
multivariate analysis, logistic or linear regression
13. Determining Whether a Variable is a Confounder
• Perform formal tests of hypothesis to assess whether the variable
is associated with the risk factor and with the outcome.
• Other investigators do not conduct statistical tests but instead
inspect the data, and, if there is a practically important or
clinically meaningful relationship between the variable and the
risk factor and between the variable and the outcome (regardless
of whether that relationship reaches statistical significance), the
variable is said to be a confounder.
• Still other investigators determine whether there is confounding
by estimating the measure of association before and after
adjusting for a potential confounding variable.
• A change in the estimated measure of association of 10% or
more would be evidence that confounding was present, but if the
measure of association changes by <10%, there is likely to be
little, if any, confounding by that variable.
• Using the example above, the adjusted risk ratio would be about
1.43 and the crude risk ratio (before adjustment) = 1.78. So the %
difference = (1.78-1.43)/1.43 = 24.5% difference. Therefore, there
was confounding by age. This is discussed further in the section
on multiple linear regression later in this module.
24. Effect modifier
• Presence or absence of an effect modifier changes the
association of an exposure with the outcome of interest.
• The situation where the magnitude of the effect of an
exposure variable on an outcome variable
differs depending on a third variable.
• occurs when an exposure has a different effect among
different subgroups. Effect modification is associated
with the outcome but not the exposure.
• For example, imagine you are testing out a new
treatment that has come onto the market, Drug X. If
Drug X works in females but does not work in males, this
is an example of effect modification.
• It provides important information. The magnitude of the
effect of an exposure on an outcome will vary according
to the presence of a third factor.
25. • Confounding factors are a “nuisance” and can
account for all or part of an apparent association
between an exposure and a disease.
• Confounding factors simply need to be eliminated
to prevent distortion of results.