Comparability study of analytical results between a group of clinical laboratories

OBJECTIVE: To describe the study of the comparability of the measurements levels of biological tests processed in biochemistry in Catlab's 4 laboratories. MATERIAL AND METHODS: Quality requirements, coefficients of variation and total error (CV% and TE %) were established. Controls were verified with the precision requirements (CV%) in each test and each individual laboratory analyser. Fresh serum samples were used for the comparability study. The differences were analysed using a Microsoft Access® application that produces modified Bland-Altman plots. RESULTS: The comparison of 32 biological parameters that are performed in more than one laboratory and/or analyser generated 306 Bland-Altman graphs. Of these, 101 (33.1%) fell within the accepted range of values based on biological variability, and 205 (66.9%) required revision. Data were re-analysed based on consensus minimum specifications for analytical quality (consensus of the Asociacion Espanola de Farmaceuticos Analistas (AEFA), the Sociedad Espanola de Bioquimica Clinica y Patologia Molecular (SEQC), the Asociacion Espanola de Biopatologia Medica (AEBM) and the Sociedad Espanola de Hematologia y Hemoterapia (SEHH), October 2013). With the new specifications, 170 comparisons (56%) fitted the requirements and 136 (44%) required additional review. Taking into account the number of points that exceeded the requirement, random errors, range of results in which discrepancies were detected, and range of clinical decision, it was shown that the 44% that required review were acceptable, and the 32 tests were comparable in all laboratories and analysers. CONCLUSIONS: The analysis of the results showed that the consensus requirements of the 4 scientific societies were met. However, each laboratory should aim to meet stricter criteria for total error.