A Study on The Predictive Model of Fatty Liver

• Main Authors: Kuo-Liang Chai, 柴國樑
• Other Authors: Chiao-Pin Bao Ph.D
• Format: Others
• Language: en_US
• Published: 2001
• Online Access: http://ndltd.ncl.edu.tw/handle/61234030899992360029

碩士 === 義守大學 === 管理科學研究所 === 89 === Fatty liver is the most common disease among liver ailment. The most important reasons are obesity, hyperlipidemia, diabetes mellitus, long-term drinking, taking drugs, viral hepatitis, etc., which are very related to modern people’s daily life. Under observation, the seriousness of fatty liver is not that important, but when we discuss the causes thoroughly, they represent a warning, which also is an important health index. If it is not thought highly of, then the chances of getting cardiovascular diseases or metabolic diseases are high. Unnecessary examination and drugs are not only bad for the cure of fatty liver, but a huge and insurmountable waste of medical resources. After looking through correlated documents, we found the correlated researches on predicting this kind of disease are rare. The purpose of this research is to use simple clinical basic datum to search for the key index of predicting fatty liver, and set up a model of predicting fatty liver. This research take the employees of southern branch of some state-owned enterprise unit as the sample. After the design of the research of proof document is done, measurement is carried out and datum are collected while the employees are taking physical examination. Total of 200 sets of observations fit into the conditions of this research. Numbers of having fatty liver and not having fatty liver are 100 for each. After processing and analyzing the datum by multivariate analysis, quantification theory type II, we get important discoveries by adding up the results as follows: 1.Gender, age, amount of triglyceride in blood, WHR, and body mass index are five variables which are related to having fatty liver or not. The order from the most related to the lesser one is body mass index, triglyceride, WHR, gender, and age. 2.Gender, age, amount of triglyceeride in blood, WHR, and body mass index are not highly related among each other. therefore, we can utilize these five variables to set up a mode of predicting fatty liver. 3.Take these five variables to analyze by quantification theory type II method and analyze the scores of samples. We get 92.5% for the percentage of hits of distinction, which is very high. 4.Add the scores of the suitable category of the clinical data of the testees and the scores of the five variables. If the sum is positive, there is 92.5% of chance of having fatty liver. However, if the sum is negative, there is 92.5% of chances of not having fatty liver. We use this mode to predict fatty liver. The result of this research shows the predicting mode has a clinical practical value. Because this research is only a small scope initial research of disease predicting mode, samples are minor, we can’t get a better precise score of categories in the predicting mode. Researchers suggest to take the predicting mode to process a large scope investigation. As a consequence, an extinguish chronic disease predicting mode is set up for clinical and prevention medical science to use. I t not only is good for personal health, disease prevention and clinical diagnose, but reduce the burden of medication finances of personal, family, society and country.