Lecture 12a - Review Session on Confounding, Effect...

Info iconThis preview shows pages 1–10. Sign up to view the full content.

View Full Document Right Arrow Icon
Review Session on Confounding, Effect Modification and Bias Define effect modification and confounding Approach to assessing study for presence of effect modification and confounding Stratified analysis to assess effect modification Stratified analysis to assess confounding Review of homework from last week Selection biases Measurement biases External validity
Background image of page 1

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

View Full DocumentRight Arrow Icon
Effect Modification Occurs when the direction or magnitude of an association between the study exposure and disease varies at different levels of a third factor (the effect modifier) Put another way: If the association is of different strengths in different strata formed on the basis of some other variable (example: sex) than you have observed an interaction between that other variable and your exposure in producing disease
Background image of page 2
Karapetis, NEJM 2008
Background image of page 3

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

View Full DocumentRight Arrow Icon
Test for interaction (effect modification) P < 0.001 Test for interaction (effect modification) P < 0.001 HR = .98 HR = .55 Overall Survival Karapetis, NEJM 2008
Background image of page 4
Example of effect modification in a case-control study Lung cancer Controls Exposed to asbestos 110 50 Not exposed to asbestos 190 550 Exposure Odds Ratio: (110/190) / (50/550) = 6.4 Disease Odds Ratio: (110/50) / (190/550) = 6.4 Crude data: all data in one table, as in a simple analysis
Background image of page 5

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

View Full DocumentRight Arrow Icon
Example of effect modification in a case-control study Lung cancer Controls Exposed to asbestos 20 10 Not exposed to asbestos 180 190 Odds ratio Non-Tobacco users = 2.1 Data stratified by tobacco use, an effect modifier Lung cancer Controls Exposed to asbestos 90 40 Not exposed to asbestos 10 360 Non-Tobacco Users Tobacco Users Odds ratio Tobacco users = 81 In this example, the crude odds ratio of 6.4 is not relevant because it applies neither to the effect of asbestos on lung cancer among tobacco users nor among nonusers – and each of us is either one or the other
Background image of page 6
Detecting Effect Modification 1. Observe stratum-specific odds ratios to see if they differ in magnitude or direction 2. Consider whether effect modification makes biologic sense 3. Perform a test of heterogeneity of the odds ratios 4. If effect modification exists, stratum- specific results should be reported
Background image of page 7

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

View Full DocumentRight Arrow Icon
Confounding An association between an exposure and outcome is distorted by an extraneous third variable (referred to as the confounding factor) Put another way: Because we are not conducting an experiment, our exposed and unexposed (comparison) groups may differ on some other risk factor for disease that would interfere with our ability to isolate the exposure effect we are interested in
Background image of page 8
Example of confounding in a case control study: Is being male associated with a higher risk of malaria infection? Malaria
Background image of page 9

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

View Full DocumentRight Arrow Icon
Image of page 10
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 30

Lecture 12a - Review Session on Confounding, Effect...

This preview shows document pages 1 - 10. Sign up to view the full document.

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