The most common biases in interpretation of clinical studies in nephrology

Abstract

Randomized clinical trials are very rare in nephrology, therefore an observational study are often used to guide clinical practice. Unfortunately, they are very prone to biases and should be interpreted very cautiously. The most common biases in nephrology
literature are reviewed, providing examples. Besides of biases type I (disease mongering) and type II, publication bias is often met. Additional four types of less obvious biases are described. These are: lead time bias, survivor bias, immortal time bias and index event bias. Finally, cognitive bias and its possible interference with clinical decisions, patient’s notes and observational studies is mentioned.