unit 8 pt 1 - Unit 8: Confounding and effect 8: Confounding...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon
Unit 8: Confounding and effect (measure) and effect (measure) modification Part 1 Dr. Borenstein Unit 8: Confounding and effect Unit 8: Confounding and effect (measure) modification ± Confounding ² Definition and identification of confounding ² Determining if a variable is a potential confounder ² Ways to control for confounding and when: randomization, restriction, matching (design), stratification, multivariate analysis, direct and indirect standardization ² Mantel-Haenszel OR/RR ² OR for matched studies and McNemar’s test ² Definition and mechanics of rate adjustment: direct and indirect ± Effect modification ² Definition and identification of effect modification ² Understanding the difference between effect modification and confounding ² Evaluation of effect-modification through stratified analysis ² Understanding when to report effect-modification and when to report confounding Confounding ± Effect of an extraneous (third) variable that wholly or partially accounts for the apparent effect (association) between the study exposure and the disease, or that masks an underlying association (“reverse confounding”) Confounding Alcohol consumption (E) heart disease (D) + + Smoking (PCF) ± Cases smoke more than controls (smoking is an independent RF for CHD independent of alcohol consumption (E) ± Smoking is associated with alcohol (ETOH) consumption but Smoking is associated with alcohol (ETOH) consumption but is not a consequence of it (ETOH is not in the causal pathway) ± PROBLEM: elevated OR ETOH-CHD may be due to lack of controlling (adjusting) for smoking controlling (adjusting) for smoking. Important Important! A PCF must satisfy three conditions ± A PCF must satisfy three conditions: 1. It is associated with study exposure 2. It is a risk factor for the disease/outcome of interest independently of exposure of interest 3. It is not an intermediate step in the causal pathway between the exposure and outcome Situations in which PCF IS a confounder fo Situations in which PCF IS a confounder for a E-D association. [ non-causal causal] What is the causal pathway [Directed E acyclic graph (DAG)]? PCF D
Background image of page 1

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

View Full DocumentRight Arrow Icon
Situations in which PCF is NOT a confounder for a E-D association. [ non-causal causal] E PCF D E D PCF PCF is associated with E, not D PCF is a result of exposure, is intermediate in the causal 1. 2. pathway, and occurs AFTER E in time (not a determinant of exposure) PCF ED PCF is a consequence of D and cannot confound E-D association 3. E PCF D PCF is a consequence of E 4. Example of an intermediate variable in a directed acyclic graph (DAG) – used to map out causal webs Diet: Salt intake Hypertension Stroke ± In the subset of people who are sensitive to salt, high salt intake is associated with having a stroke.
Background image of page 2
Image of page 3
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 7

unit 8 pt 1 - Unit 8: Confounding and effect 8: Confounding...

This preview shows document pages 1 - 3. Sign up to view the full document.

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