Quiz Flashcards
Confounding definition
confusion, or mixing, of effects; the effect of the exposure of interest is distorted because the effect of an extraneous factor is mistaken for or mixed with the actual exposure effect
Confounding occurs when…
an extraneous variable, either partially or completely accounts for an apparent association between study variable and outcome
primary prevention (association)
causal association between the risk factor and disease must exist
secondary prevention (association)
association may either be causal or statistical: the association may be confounded
three ways of thinking about confounding
Classical approach, collapsibility approach, and counterfactual approach
classical approach to confounding criteria for confounding
1) must be a risk factor for the disease independent form the study exposure
2) must be associated with the exposure under study in the source population
3) must not be a result of the exposure- not an intermediate variable in the causal pathway between the exposure and the disease
exceptions to the criteria for confounding under the classical approach
1)random statistical association due to sampling variability
2) marker of unmeasured confounder
3) confounder is an intermediate variable, but question is whether other causal pathways exist
Collapsibility approach to confounding criteria
the effect measure is homogenous across the strata defined by the confounder
collapsibility is the equality of stratum-specific measures of effect with the crude, unstratified measure
counterfactual model approach to confounding
theoretical approach for defining the ideal comparison group; considers what the risk of the outcome would have been in the same exposed individuals if exposure had been absent
restriction (to control for confounding in design stage)
confounding cannot occur if the distribution of the potential confounding factors do not vary across exposure or disease categories
matching (to control for confounding in design stage)
selecting subjects according to the value of suspected confounder to ensure equal distribution among study groups
matching in cohort studies
matching unexposed to exposed without regard to disease status
matching in case-control status
matching non-diseases to diseased subjects
control confounding in analysis stage
stratification; multivariable analysis; propensity scores; marginal structural models; direct acyclic graphs
stratification
evaluating the association between exposure and disease within homogenous categories of the confounding variable
traditional multivariable analysis
can handle large numbers of confounders at once; based on statistical regression models; selection of potential confounders based on a prior knowledge, evaluation of data, and best judgment of the investigator, NOT on statistical significance
conditional confounding
a presumed confounder may be cofounded by other variables
residual confounding
adjustment does not completely remove the confounding effect due to a given variable or set of variables
collinearity (confounding)
excessive correlation between confounder and exposure of interest
biological interaction
co-participation in a causal mechanism of two or more component causes
statistical interaction
effect modification: effect measure modification; heterogeneity of effects; subgroup effects and statistical interaction
interaction- heterogeneity definition
the effect of risk factor (A) on the risk of an outcome (Y) varies according to the levels of a third variable (Z)
interaction - statistical interaction/joint effects definition
the measure of association when both risk factors are present is different than what would be expected form combining their individual effects
interaction (general knowledge)
reciprocity of interaction; goal is to identify and report
additive interaction
absolute difference or attributable risk model (Cohort study ONLY)
multiplicative interaction
relative difference or ratio model
quantitative interaction
when the association between factor A and outcome Y exists and is of the same direction in each stratum formed by Z, but the strength of the association varies across strata
Qualitative interaction
the effects of A on the outcome Y are in opposite directions according to eh presence of the third variable Z or when there is an association in one strata formed by Z but not in the other
cumulative incidence
proportion of the population at risk that will develop an outcome in a given period of time (time is not part of calculation)
incidence rate
rate at which new cases occur per unit of time (time is part of calculation)
Attributable risk
how much of the disease in people who are exposed is due to the exposure
AR=CIe-CInoe)
population attributable risk
how much of the disease in the population is due to the exposure
PAR=CItoalpop-CInoE
cohort study requirements
1) cohort members must meet criteria for being at risk of the disease
2) subjects initially free of disease under study
3) subjects are classified by the exposure status
4) subjects are followed prospectively over time for incidence of disease
5) incidence of outcome is compared by exposure status
6) usual measured of association are ratios of incidence proportion or rates or SMRs, SIRs
select comparison groups-cohort
unexposed, internal comparison, external comparisons
counterfactual example definition
risk of disease in the exposed if they had never been exposed
etiologically relevant time window (cohort)
period of time during which an exposure is capable of causing disease
lag time
how soon after exposure would you expect to see the disease
information bias in cohort study
misclassification of exposure at baseline; changes in exposure status over time; ascertainment of outcomes during follow-up
immortal time bias
assignment of exposure during a time when the participant is considered ‘immortal’
among exposed, duration of exposure increases during follow-up