Lecture 6: Association & Causality Flashcards Preview

Epidemiology Exam 1 > Lecture 6: Association & Causality > Flashcards

Flashcards in Lecture 6: Association & Causality Deck (23)
Loading flashcards...

define cause

a pre-cursor event required for the occurrence of the disease


a study may yield associations between exposure and disease, but this does not mean ……?

that the exposure is the cause of the disease


3 types of associations

artifactual associations
non-causal associations
causal associations


artifactual associations

associations that are false/wrong

#'s may show an association but it's actually not related at all

can come from bias or confounding


non-causal associations

can occur in 2 ways
1. the disease causes the exposure
2. disease and exposure are both related to a confounding variable


types of causal relationships

1. sufficient cause
2. necessary cause
3. component cause


a set of minimal conditions that inevitably will produce disease 100% of the time

sufficient cause


type of cause which precedes a disease, and if present, the disease will always occur

sufficient cause


type of cause that must be present for the disease to occur, yet you may have this cause and never get the disease

necessary cause


a factor that if present, increases the probability of a particular disease. most common example?

component cause
ex. age


2 interactions in causal research

1. synergism
2. parallelism


define 'synergism'

interaction of at least 2 component-causes, such that the combined effect of the components is greater than the effect of just one cause being present


define 'parallelism'

interaction of at least 2 component-causes, such that the measure of effect is greater if either one is present.

but they are not occurring at the same time. must have 2 variables to compare and their effects on RR separately


_______ causes work in concert to collectively become ______ causes.

multiple component causes together become sufficient causes


how can we decide if the RR's of risk factors contain enough of a relationship to be called a cause?

use Hill's Guidelines to create causal inferences


Hill's criteria/Guidlines

1. strength
2. consistency
3. temporality
4. biologic gradient
5. plausibility


explain the strength guideline

refers to the size of the measure of association (RR)

the greater an association value, the more convincing it is to show a causality


explain consistency guideline

the repeated observation of an association, across different studies, populations, or circumstances

multiple studies show that same result


consistency may obscure _______

the truth!

observational studies might show an association but is possible for it to be wrong after doing randomized blind studies


explain temporality guideline

reflects that the cause precedes the outcome
proximal or distant cause in time-line

a cause happens just before the outcome so you assume there is an association. but this is not always the truth


explain biological gradient guideline

the presence of a gradient of risk associated with the exposure

more of exposure = greater probability of an outcome
ex. 10 packs a day = greater chance of lung cancer than 1 cigarette a day


explain plausibility guideline

biological feasibility of the association

can the cause be explained or understood physiologically


issue with plausibility

decisions of plausibility are based upon known beliefs but our current beliefs may be wrong. we don't know everything or understand everything that happens