Causal Inference Flashcards
a cause consists of BOTH ___ agents and __ conditions
a cause consists of BOTH active agents and static conditions
why do we care about causality?
if we find a cause, we can intervene to reduce or prevent disease
goal of epidemiology
to determine factors that cause distribution of disease
2 steps to review evidence
- is there an association btwn exposure and risk of disease
- is it likely to be causal
investigating correlation between dietary fat intake and breast cancer: what 2 studies should be done?
- ecologic
- retrospective cohort
benefits of observational studies
cheaper
fewer ethical issues
faster (sometimes)
can test multiple hypotheses and associations
benefits of experimental studies
variables of interest are more controlled
other extraneous variables are controlled
sequence of studies in humans (5)
- clinical observations
- available data
- case-control
- cohort
- randomized trials
meaning of spurious
association that is false
due to confounders/selection bias
causality is based on strenght of __
causality is based on strenght of evidence
is one study enough to develop a causal relationship?
sometimes, but not usually
when looking for a causal relationship, we must consider
generalizability
direct causal pathway
a fator directly causes a disease without an intermediate
indirect causal pathway
factor causes a disease but only through an intermediate
4 types of causal relationships
- necessary AND sufficient
- necessary BUT NOT sufficient
- sufficient BUT NOT necessary
- NEITHER sufficient NOR necessary
necessary AND sufficient
factor A –> disease
pathogenic TB exposure –> TB
necessary BUT NOT sufficient
factor A + factor B + factor C –> disease
toxin + diet –> disease
sufficient BUT NOT necessary
factor A or factor B or factor C –> disease
NEITHER sufficient NOR necessary
factor A + B or factor C +D or factor E + F –> disease
diet, exercise
when alpha = 0.05, type I errors happen
5% of the time
specificity of association
specific exposure associated with only one disease indicates causality
Henle-Koch
absence of specificity doesn’t mean __
no causal relationship
conclusion 1: categorize evidence by quality
1) clinical trials
- randomized, blinded
- randomized
- non-randomized
2) cohort/case-control
3) other (case series)
conclusion 2: major guideliens for evaluating causality (4)
- temporal relationship
- biologic plausability
- consistency
- alternate explanations explored
3 other considerations for causality
- dose repsonse
- strength of association
- cessation of effects
