| Summary: | 碩士 === 國立雲林科技大學 === 工業工程與管理研究所碩士班 === 95 === The end stage renal disease（ESRD）patients require dialysis to survive. The numbers of patients used Hemodialysis is up to 92.4% in 2007 in Taiwan. However hemodialysis often has many complication in which the Dialysis-induced Hypotension（DIH）is the most serious one, especially for diabetic. The lighter symptom of DIH is vertigo and uncomfortable but the serious symptom will result in arrhythmia or even death. In order to allay the seriousness of the symptom during hemodialysis to differ danger patients, the aim of the study is to construct a prognosis model for DIH by regression analysis and backpropagation neural network（BPN）and to find the important factors affecting dialysis diabetic and non-diabetic dialysis patient. Furthermore, the study uses genetic algorithms to find the optimal parameter combination. In the end, the better model is BPN and it indicated that controlling the quantity of dehydration, sodium deepness, temperatures of dialysis liquid, kinds of dialysis liquid, initial velocity of blood stream for dialysis diabetic. The above factors plus controlling the type of artificial kidney affect non-diabetic dialysis patients. This result can reduce the degrees of dropping blood pressure during hemodialysis and provide physicians with diagnosis information and nurses with preventive caution.
|
|---|
