Bias & Confounding in Clinical Studies: USMLE Cheat Sheet
Bias is any systematic error that leads to an incorrect estimate. Unlike random error, bias cannot be fixed by increasing sample size.
Types of Bias
- Selection Bias: Berkson's bias, healthy worker effect, loss to follow-up
- Information Bias: Recall bias, observer bias, Hawthorne effect
- Screening Bias: Lead-time bias, length-time bias
Confounding
A confounder must be associated with both the exposure AND the outcome, and must NOT be on the causal pathway. Control with randomization, matching, stratification, or multivariable regression.