Journal of The Japanese Society for Quality Control
Online ISSN : 2432-1044
Print ISSN : 0386-8230
Current issue
Showing 1-12 articles out of 12 articles from the selected issue
Invited Article
  • Hideki NOMOTO
    Type: Invited Article
    2020 Volume 50 Issue 3 Pages 192-197
    Published: July 15, 2020
    Released: October 30, 2020
    JOURNALS RESTRICTED ACCESS
     In this paper, methodology of developing highly reliable system using FRAM (Functional Resonance Analysis Methodology) is described. Increase of IoT or AI applications for the safety critical systems such as self-driving car is changing the system engineering. The main factor of the change is that the system itself is beginning to have capability to adopt to environment change by IoT or machine learning. In this situation, the need for the new way of system engineering is getting attention, especially for the system which has great flexibility to cope with the changing environment. FRAM is the methodology to model how each function is connected and interacting with each other to cope with the changes.
    Download PDF (1251K)
  • Masaomi KIMURA
    Type: Invited Article
    2020 Volume 50 Issue 3 Pages 198-203
    Published: July 15, 2020
    Released: October 30, 2020
    JOURNALS RESTRICTED ACCESS
     Confusing package designs or drug names can cause severe medical accidents or incidents. We applied a data mining technique to incident case data and found drug mixing-up cases caused by confusing designs or names. In this article, we reviewed studies on indices to measure similarities of label designs of injection ampoule container, PTP package designs for tablets/capsules, look-alikeness and sound-alikeness of drug names.
    Download PDF (993K)
Report of Special Interest Group
Report of Research Group
Research Papers
Contributed Paper
  • Ryoko SHIMONO, Rie AKINAGA, Satoko TSURU
    Type: Contributed Paper
    2020 Volume 50 Issue 3 Pages 234-244
    Published: July 15, 2020
    Released: October 30, 2020
    JOURNALS RESTRICTED ACCESS
    In order to manage competence of personnel who contribute to quality assurance,a practical method for competence evaluation which can specifically indicate knowledge and skills is required.Though such a method for competence evaluation has been developed in a clinical laboratory test,effectiveness of this method has not yet analyzed.This method indicates that competence items are factors that affect the variation of test results using cause-and-effect diagrams.The purpose of this study is to verify the effectiveness of the method of developing competence evaluation items through description of cause-and-effect diagrams in clinical laboratory tests.We applied and compared the aforementioned method with that of developing competence evaluation items through description of the operation process.To compare the results,we developed two original indicators : one expressing the potential competence to be achieved and the other expressing the difference in competence levels among clinical technologists.The findings verified the effectiveness of the method that used cause-and-effect diagrams for enhancing competence levels.It is implied that compared to the method that used description of the operation process,the method of building competence evaluation items using cause-and-effect diagrams is more useful for evaluating the required knowledge for quality assurance of clinical tests.
    Download PDF (1510K)
Quality Report
  • Kenji SHIMIZU, Takeshi NAKAJO
    Type: Quality Report
    2020 Volume 50 Issue 3 Pages 245-251
    Published: July 15, 2020
    Released: October 30, 2020
    JOURNALS RESTRICTED ACCESS
     Behavior observation is effective in understanding latent needs that customers themselves are not aware of. However, in behavior observation, it is necessary to find a need while processing a large amount of information at a time, and its success depends on the skill of the observer. In order to deal with this problem, Sekiya et al. have proposed the limited observation method that makes it easier to find latent needs by narrowing down the behaviors to be observed in advance. In this study, the behaviors to be observed were narrowed down by various methods, and the behaviors were actually observed. Based on the obtained data, the relationships between how to narrow down the behaviors to be observed and the needs obtained were investigated. As a result, it showed the followings: 1)Under the condition that the total observation time is the same, when the number of behaviors to be observed decrease, the number and concreteness of the newly obtained required qualities related to the focused quality increase; 2) When narrowing down the behaviors to be observed, the behaviors related to the required quality to be obtained should be selected as a set; 3) The relationships between how to narrow down the behaviors to be observed and the obtained required qualities are the same even if the observers are different.
    Download PDF (1161K)
  • Tetsuo MATSUMOTO, Toshiyasu SATO
    Type: Quality Report
    2020 Volume 50 Issue 3 Pages 252-260
    Published: July 15, 2020
    Released: October 30, 2020
    JOURNALS RESTRICTED ACCESS
     Preferable analysis method has not been established as an analysis for temporal data or for a covariate which influences the response.
     ANCOVA is a possible candidate, and also there are three alternative candidates. First is regression analysis which uses quantification method of the first type. Second is ANOVA which uses differences between before-treatment and after-treatment. And the last is usual ANOVA.
     Power of test of treatment effects as well as the bias of estimation in treatment effects for the four methods are compared, and also were investigated by typical examples and numerical simulations.
     If it is not to be particular about the theory hard, the use of ANCOVA or usual ANOVA is proposed as follows, under consideration for the size of the correlation in the response and the covariate.
     When there is the correlation between the response and covariate, ANCOVA is appropriate and usual ANOVA is suitable only when the correlation is small, because it can’t adjust the correlation.
     When the correlation is relatively high and the treatment effects are about equal to the size of the standard deviation, power of test must meet the practical demands enough.
    Download PDF (1483K)
feedback
Top