unit 8 additional questions

unit 8 additional questions - Click to edit Master subtitle...

Info iconThis preview shows pages 1–9. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Click to edit Master subtitle style Unit 8 Fall 2010 Questions n Confounding Definition and identification of confounding Determining if a variable is a potential confounder Ways to control for confounding and when: randomization, restriction, matching (design), stratification, multivariate analysis, direct and indirect standardization Mantel-Haenszel OR/RR OR for matched studies and McNemars test Definition and mechanics of rate adjustment: direct and indirect n Effect modification Definition and identification of effect modification Understanding the difference between effect modification and confounding Evaluation of effect-modification through stratified analysis Study Guide Confounding is: n Mixing of effects of extraneous factors with the effect of int erest. n Change in an effect measure upon stratification or adjustment for extraneous factors n Var iables that can cause or prevent the outcome of interest, are not int er mediat e var iables, and are associat ed with the exposure(s) under investigation How do we identify whether confounding exists? n Know the literature well. Understand what other risk factors are related to your outcome Understand how other risk factors are related to your exposure Figure out biologically plausible DAG. n Know your data well. Perform univariate, bivariate analyses to characterize distributions of variables in each comparison group (or cohorts); ways that single variables are related to Which DAG shows E PCF D E D PCF PCF E D YES/ NO YES/ NO YES/ NO I s obesity associated with n Note: Hyper tension may be a PCF. n Draw the DAG and stat e the role of hyper t ension. Multiple sexual partners and HIV Multiple partners (E) HI V infection (D) ? Older age (PCF) Randomization - design PCF E D Randomization equalizes the E and by PCF. Therefore, the PCF-E association is eliminated, thereby eliminating potential confounding. Even if PCF is related to D, confounding by PCF will not exist if randomization is done correctly....
View Full Document

Page1 / 30

unit 8 additional questions - Click to edit Master subtitle...

This preview shows document pages 1 - 9. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online