EPID 301 Chapter Notes - Chapter 15: Relative Risk, Dependent And Independent Variables, Odds Ratio
11 Jul 2018
School
Department
Course
Professor
Document Summary
Chapter 15: stratified analysis and regression modelling in analytical studies. Analysis-stage techniques- stratified analysis and regression modelling add confidence to conclusions about exposure-disease associations. Randomization: ensure an equal distribution of all confounders, whether measured or unmeasured, between the groups in a study. An unbalanced distribution of confounding variable may occur by chance during randomization. Restriction: eliminates extraneous variables by restricting who can participate in a study. Matching: eliminates the effect of extraneous variables by balancing their presence in the groups a study compares. Recall: confounding is an intermixing of the effects of an exposure with those of an extraneous variable = un-mix these effects. If sex is a confounder, conducting the analysis separately in men and women will un-mix the effect of an exposure from the effect of sex. Confounding occurs when an appropriate adjustment for confounding leads to a change in the estimated effect. When an estimate is stratified, the stratified estimates are type of adjusted estimate.
