Clinical chemistry uses automated measurement techniques and medical knowledge in the interest of patients and healthy subjects. Automation has reduced repeatability and day-to-day variation considerably. Bias has been reduced to a lesser extent by reference measurement systems. It is vital to minimize clinically important bias, in particular bias within conglomerates of laboratories that measure samples from the same patients. Small and variable bias components will over time show random error properties and conventional random-error based methods for calculating measurement uncertainty can then be applied. The present overview of bias presents the general principles of error and uncertainty concepts, terminology and analysis, and suggests methods to minimize bias and measurement uncertainty in the interest of healthcare.