Abstract
Since measurement values contain measurement errors, it is important to control the accuracy of measurement in clinical chemistry inspection. The Japan Medical Association and the Japanese Association of Medical Technologists conducted a survey to help all institutions improve the quality of their accuracy management. In the survey, the samples of the 'same' sera were distributed to participating institutions. However, since sera contain many unstable ingredients, the quality of a few of the samples had changed and their concentrations had deteriorated. At the insitutions that received those deteriorated samples, the quality of the samples was inaccurately evaluated. As a result, those who conducted the survey need to identify which samples had deteriorated and distribute normal samples to the affected institutions.
In this paper, we made a model that any institution can use to measure the values of the samples they receive, and we proposed a method to detect deteriorated samples by statistical analysis of measurement values in the survey.
The proposed method makes it possible to detect deteriorated samples and obtain correct data to achieve the purpose of the survey.
