| Summary: | 碩士 === 國立勤益科技大學 === 工業工程與管理系 === 103 === As the nation’s living standard has been improved, patients’ satisfaction and quality of medical services has become key factors of increasing competitiveness by hospitals. This article uses the industry of hemodialysis for example, which is an industry with fierce competition where hemodialysis patients have been increasing, so have dedicated doctors, and the preventive healthcare also requires the reduction of dialyzer clotting cases. How to upgrade the medical quality of hemodialysis has become an important issue for hospitals and institutions. When performing hemodialysis treatment, the medical persons should constantly observe the patients’ vital signs and be more attentive to whether dialyzer clotting happens. What require attention are particularly the inadequate dosage of anticoagulant, air remaining in artificial kidney, too small blood flow, high volume of blood transfusion-resulted hematocrit, and too long interruption of hemodialysis pumps, among other things, which stop blood flow and all significantly contribute to dialyzer clotting. These, in serious cases, will put the hemodialysis process to a halt and require immediate replacement of artificial kidneys and related consumables, which not only costs nursing human power but also leads to severe anemia in patients or even endangers their lives.
Given the medical requirement for hemodialysis quality and dialyzer clotting rates, this study began by being comprehensive of the adjustment parameters for hemodialysis and the problem of optimization of the factor parameters for the requirement of decreasing clotting rates. Considering that inappropriate design of control parameters tends to increase clotting rates and may affect the patients’ safety and the quality of hemodialysis, we designed the most appropriate control parameters. This study used Taguchi method to identify the key factors and conducted OMEGA transformation, before testing the optimal parameters for dialyzer clotting by neural algorithm. Cross effects between the factors were investigated in the principle of economy to further arrive at the basis for reference in optimization. That would effectively reduce medical costs and expenses, enhance patient safety and increase hemodialysis quality.
By Taguchi method and OMEGA transformation, a better combination of factor levels was obtained which consisted of 500 units of anticoagulant, blood flow at 300 ml/min, dehydration at 4 kg, and patient’s own vessels. We predicted a decrease of flaw rate to 13.17% under the optimal combination of factor levels, and a validation experiment by simulating the process of hemodialysis demonstrated the decrease in flaw rate to 16%, which really did reduce the flaw rate of hemodialysis process and was approximate to our forecast with 13.17%, a good reproducibility. Last, we used Back-Propagation Neural Network as the experimental network model in this study and conducted the experiment with Matlab of R2012a version. With four factors of the Taguchi method as the input parameters and the experiment data on hemodialysis, it was revealed that MSE was 0.00030457 with a convergence achieved.
|
|---|
