Premium
Relationship of quality goals and measurement performance to the selection of quality control procedures for multi‐channel haematology analysers
Author(s) -
Westgard J. O.,
Cembrowski G. S.
Publication year - 1990
Publication title -
european journal of haematology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.904
H-Index - 84
eISSN - 1600-0609
pISSN - 0902-4441
DOI - 10.1111/j.1600-0609.1990.tb01521.x
Subject(s) - quality (philosophy) , sample (material) , computer science , selection (genetic algorithm) , term (time) , stability (learning theory) , statistics , reliability engineering , algorithm , mathematics , artificial intelligence , machine learning , engineering , philosophy , chemistry , physics , epistemology , chromatography , quantum mechanics
An approach is described for selecting QC procedures based on goals for analytical quality, the performance characteristics of the measurement procedure (imprecision, bias, frequency of errors), and the performance characteristics of the control procedure (probabilities for error detection and false rejection). Performance characteristics of stable sample QC procedures and patient data QC procedures (retained patient specimens, Bull's single‐rule algorithm, Bull's multi‐rule algorithm) are compared to determine when to apply these different QC procedures to multichannel hematology ! analysers. Precise, stable methods may be controlled using Bull's single‐rule algorithm; less precise, less stable methods require Bull's multi‐rule algorithm, retained patient specimens, or stable sample QC procedures; imprecise and unstable methods are best controlled using stable sample QC procedures. “Multi‐stage” designs may employ stable samples for “startup” testing, retained patient specimens for short term monitoring, and Bull's single‐rule algorithm for long term monitoring.