Procedure for studying commutability validated by biological variation

• Published in: Clinica chimica acta Vol. 268; no. 1; pp. 73 - 83
• Main Authors: Ricós, Carmen, Juvany, Roser, Jiménez, Carlos Victor, Perich, Carmen, Minchinela, Joana, Hernández, Amparo, Simón, Margarita, Alvarez, Virtudes
• Format: Journal Article
• Language: English
• Published: Netherlands Elsevier B.V 10.12.1997
• Subjects: Chromatography, High Pressure Liquid; Clinical Chemistry Tests - standards; Commutability; Creatinine - blood; Humans; Kidney Diseases - blood; Quality assurance; Reference Standards; Reproducibility of Results; Transferability; Commutability; Quality assurance; Transferability
• ISSN: 0009-8981; 1873-3492
• DOI: 10.1016/S0009-8981(97)00172-1

In the field of laboratory medicine, the two quantitative approaches designed to identify the stabilized materials that produce results that are commutable with results from patients' samples were found to differ. In commutability evaluations, the responses of each material and each method studied are specific, thus, it is vital to standardise the procedure used for determining this characteristic. We incorporated statistical components from the two described methods that seemed to be consistent, and added a new element based on biological variation, to validate the criterion of acceptability that determines whether or not a material is commutable. The three methods for studying commutability (using the confidence interval [α=0.05], the ±2s yx formula, and the limit based on biological variation as acceptability criteria) were applied to creatinine results from 31 stabilised materials and serum samples analysed with seven instruments, when compared against a reference method for creatinine analysis. Over the wide range of concentrations studied, the confidence interval limit and the biological variation limit coincided in the identification of commutable materials, whereas the ±2s yx was excessively permissive at normal and low concentration levels. We therefore recommend the use of Passing-Bablok regression with its confidence interval (α=0.05) in studies concerning commutability. Using this method, commutability is simple to calculate with available software and, as validated by biological variation, results are reliable.