IMPROVEMENT OF UNCERTAINTY BY MCMC FOR BLOOD CHEMICAL ANALYSIS

Yasuo Iwaki, Tadao Inmuta

Abstract

The purpose of this research is improving so that the result of blood chemical analysis (BCA) can be assured. It can be useful for clinical diagnosis. It is also included. ISO-GUM was created as guidance for making an analysis result into an assurance performance. In supplement 1 of ISO-GUM (ISO-Guide to the express of uncertainty in measurement) in detail, it recommended using Markov Chain Monte Carlo (MCMC) for "law of propagation of the uncertainty (LPU)". The main features of ISO-GUM changed the mode of expression of an analysis result into uncertainty from error, and processing of data analysis was changed into Exploratory Data Analysis (EDA) from conventional ANalysis Of VAriance (ANOVA). The reason for change is for obtaining an exact result, even if a measurement result is abnormal distribution. MCMC was taken in also in the field of BCA. In the research, it experimented for the quality assurance (QA) of calibration curve use quantitative analysis. An accuracy of calibration curve is importance situation as intermediate accuracy in a whole measurement system. The important factor is a set up reference value and a confidence interval for assurance of an analysis result. Since the satisfy conclusion was obtained, it is reported.

Keywords

MCMC, ISO-GUM, uncertainty

Download

IMEKO-WC-2009-TC7-041.pdf

DOI

-

IMEKO TC

TC7 - Measurement Science

Back to the proceedings