Reference ranges, standardization and decision levels

Abstract

BACKGROUND-AIMResults of laboratory investigations (especially if values fall in"grey-zone") have limited meaning on medical decisions, without reference intervals (RIs). The majority of medical laboratories adopt intervals (reccomended by manufacturers or by scientific literature) but take ensurance that they are proper for clinical use and compatible with its own population. Almost none of laboratories spend time and resources to establish reference intervals themselves due to dificulties in selection of reference group. We aimed to use a posteriori technique to calculate reference intervals from collected data using statistical techniques (robust method after eliminating outliers). METHODSData collection was exhaustive: all patients investigated in Central Medical Laboratory of Tirgu Mures, Romania, Emergency Hospital between 2010 and 2013 were included in the study. The data collected are values for biochemical (from COBAS 6000 and ARCHITECT platforms) and haematological parameters (from Sysmex and Cell Dyne analysers). The methodology involved data processing using sophisticated filters to eliminate outliers and complex statistical algorithms to calculate percentiles 2.5 and 97.5 and histogram plotting with cumulative percentages to derive the final reference intervals (indirect Hoffmann method).. RESULTSRobust linear regressions (biomedical parameter results vs cumulative percent) were obtained in order to find the best fitted linear segment of the curves, which contained reference intervals. 60-85 % of collected data could be used to obtain RIs. The obtained ranges were evaluated in comparison to the ranges indicated by reagent manufacturers and between different manufacturers. The reproducibility and acuracy of obtained RIs were in acceptable limits. A multicenter comparative evaluation of these intervals should be indicated to confirm the results. CONCLUSIONRobust "a posteriori" studies can be used to establish and to check transferred RIs. Indirect visual method -Hoffmann provides accurate/ reproducible RIs. Detection and elimination of outliers is not necessary (the variation of RIs values are insignificant). Supplimentary verifications are needed to check appropriately intervention in medical decisions. BACKGROUND-AIMAll the medical laboratories in Nepal are using the reference range of kit supplied in the reagents due to lack of local reference range. Establishment of reference range locally is important as analyzing the specimens. Reference interval is essential to make decision about the patient's condition. Reference range is affected by age, sex, diet, ethnicity, environment and genetics. This is the reason for each laboratory should have its own reference range for each test.This study was designed to establishes reference interval for outline biochemical analytes for the Nepalese adult population , a part of multicenter reference interval project being conducted around the world by the IFCC,C-RIDL. METHODSFrom six different regions, fasting venous blood was obtained from apparently healthy volu...