The reliability of calculated laboratory results

Abstract

AbstractIn clinical laboratories, patient results can be obtained in two ways: (i) by direct determination of requested tests using various chemical methods, (ii) by calculation of unknown test results, based on relationships between measured tests. The reliability of measured tests can be checked by various quality control rules. However, no test is performed to check the reliability of calculated data. In this study we develop a method using Taylor series expansion and an alternative equation to obtain the standard deviation of calculated laboratory tests and discuss the reliability of calculated data. To obtain reliable test results by calculation instead of being measured by chemical methods, the standard deviation of each measured component of the equation must be thoroughly analyzed and then the standard deviation of the equation must be determined. We conclude that the analytical coefficient of variation of any measured component must be lower so as to obtain an acceptable analytical coefficient of variation for calculated tests. Otherwise we should measure the concentration of requested tests by chemical methods instead of calculation by equation using specified components.