Cases and controls in a case-control study – Exposed and nonexposed in a cohort study • Types of differential bias – Selection – Information (aka Observation)
Selection Bias Mechanism of Selection Bias in the Classic Example 38 Target/Source Population Cases Coffee Controls Coffee Cases No Coffee Controls No Coffee Study Population RR true = 1 RR observed = 2.7 Example : Selected in a greater proportion of controls who were not drinking coffee than in the source population. Selection Bias Occurred: RR true < RR observed Distortion of Association A Classic Example of Selection Bias Cases : Patients with histologic diagnosis of pancreatic cancer in any of 11 large hospitals in Boston and Rhode Island between October 1974 and August 1979. Controls : Other patients under the care of the same physician as the cases with pancreatic cancer. Patients with diseases known to be associated with smoking or alcohol consumption were excluded. Coffee and cancer of the pancreas MacMahon et al. N Eng J Med 1981 ; 304:630-3
Information Bias (aka Observation Bias)
Module 3: Confounding & Effect Modification (10/7 - Dr. Platz) • Identified and learned how to eliminate or reduce confounding in the design and analysis of observational epidemiologic studies. A type of bias • Identified and learned how to highlight the presence of effect modification. Not bias, possibly biology
Confounder – Classic Definition • A confounding variable or confounder is a 3rd factor that is (1) A risk factor for the disease, (2) Associated with the exposure, and (3) Not a factor in the causal pathway from exposure to disease (i.e., not a mediator). • If any of these criteria is not satisfied, then the 3rd factor is not a confounder. B Disease Exposure Spurious association B is a confounder
Empirical Way to Identify a Confounder a) Calculate crude RR D+ D- E+ E- RR crude b) Stratify by potential confounder and calculate stratum-specific RRs C+ C- D+ D- D+ D- E+ E+ E- E- RR 1 RR 2 Using the actual data from your study: Compare crude and stratum- specific RRs • If they differ by more than 10- 20%, then confounding is likely present.
Ex: Assess Presence of Confounding • Is smoking a confounder of the alcohol-MI association? - Yes, the stratum-specific ORs differ from the crude OR • Is alcohol drinking associated with MI? - No, the un-confounded, stratum- specific ORs=1.0 Adapted from Schlesselman 1982 MI+ MI- Alcohol+ 71 52 Alcohol- 29 48 OR crude = 2.26 Smoking+ Smoking- MI+ MI- MI+ MI- Alcohol+ 8 16 Alcohol+ 63 36 Alcohol- 22 44 Alcohol- 7 4 OR 1 = 1.0 OR 2 = 1.0
Methods to Minimize Confounding
Effect Modification • In contrast to confounding, effect modification is not an error in the design or analysis of the study. • Effect modification is present when the association between an exposure and outcome differs between categories of a third factor.
You've reached the end of your free preview.
Want to read all 99 pages?
- Fall '15
- Epidemiology, Review, Exam