EPID 301 Chapter Notes - Chapter 9: Standardized Mortality Ratio, Sampling Probability, Null Hypothesis
Document Summary
Results in a failure to accurately assess important (useful, actionable) variables from a variety of variables at play in a population. The problem of confounding : intermixing of effects. Descriptive epidemiologic comparisons of different populations are complicated by the fact that potentially interesting differences might merely reflect underlying differences in non-modifiable determinants. Confounding: the intermixing of more and less interesting effects when populations are compared. Standardization -> persistent difference (effects of modifiable risk factors) -> descriptive studies to analytical studies. Recall: descriptive studies= generate hypotheses about etiology vs. analytical studies= address specific hypotheses about disease etiology. Risk marker or risk indicator = a variable associated with increased risk in a way that is not due to a cause-effect relationship. To identify opportunities for prevention, you want to know whether the difference in mortality is due to something more than differences in age or sex. Strategies to adjust for confounding in descriptive studies. Simplest type of adjustment to un-confound variables.