Identifying Potential Biases in the Nephrology Literature.

Abstract

Observational studies are common in the nephrology literature, particularly given the lack of large randomized trials. While these studies have identified important associations, potential biases, if unrecognized, can result in misinterpreted conclusions. In this review, we present an example of four potentially important biases (lead time bias, survivor bias, immortal time bias, and index event bias) that can result in inaccurate estimates of association between risk factors or treatments and outcomes. Recognition of these potential biases can help improve study design and interpretation.