MCC-UE 1040 Lecture Notes - Lecture 16: Asthma, Confounding
Document Summary
Confounding: a type of bias that distorts the observed relationship because it is mixed with the effect of an extraneous factor. It is associated w/the outcome - a known risk factor for the outcome. It is not an intermediate step in the pathway b/w exposure & outcome. Outside the context of study, there are many reasons people take one medication over another, and these reasons are often related to outcome(s) of interest. Confounding: risk factors for the disease are associated w/use of drug. If randomization is successful, there is no confounding at baseline restriction/filter. Assess association b/w exposure & outcome w/in each stratum of the potential confounder. If the observed association b/w exposure & disease is due to confounding: Crude measures of association will differ from stratum- specific (adjusted) measures. Expect no difference in measure of association b/w strata. Education: