Method for quality control of laboratory tests using histograms of daily patient data.

Abstract

A method for controlling the quality of laboratory tests is proposed. Histograms of patients' daily results which fall within reference ranges of healthy individuals are used for estimating accuracy and precision of measurements. For the determination of accuracy, three methods are evaluated; computing an average of patients' results; determining the location of the peak of the histogram; approximating the histogram by an Erland distribution and determining the peak of the distribution. For precision control, standard deviations are calculated from patient data. We applied these methods to serum aspartate aminotransferase (AST or SGOT) and total cholesterol of patients in a general hospital. Averages, peaks of approximated Erland distribution, and standard deviations were found to be useful to daily quality control in laboratories of large hospitals.