bias

Bias is an estimate of systematic measurement error: the difference between the average of repeated measurements and an appropriate reference value. It can be expressed in concentration units or as a percentage of that reference value. A positive bias indicates higher average results; a negative bias indicates lower average results. Random variation can obscure bias when too few observations are available.

The strength of the estimate depends on the reference used. A peer-group mean may reveal disagreement with comparable laboratories but does not necessarily establish closeness to a higher-order reference. Consider material commutability, concentration, calibration traceability, and the uncertainty of assigned values. Method-comparison differences may vary across the measuring interval, so one overall percentage can hide clinically important concentration-dependent behavior.

Assessors expect laboratories to recognize systematic differences that could affect clinical interpretation. Document how bias was estimated, what comparison was used, and the acceptance criteria applied. Unexpected bias should trigger investigation of calibration, reagents, and procedure performance, with assessment of patient impact where relevant, rather than an unsupported adjustment to reported results.