Survival and Discriminant Analysis of Hospitals in Taiwan

• Main Authors: TUNG,CHEN-CHEN, 董振楨
• Other Authors: 張永源
• Format: Others
• Language: zh-TW
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/k5ywmm

碩士 === 高雄醫學大學 === 公共衛生學研究所碩士在職專班 === 91 === There are 1295 hospitals of the samples in all. The earliest founded hospital established in 1911, is the National Taiwan University Hospital. The last hospital established in February 2003 is the Chang-you Hospital. Excluding out the 83 hospitals which established in2003 and the 213 hospitals which there were not data of the equipment and amount of service, the rest amount of hospitals is 999.The censored interval was from January 1, 1995 to December 31, 2002; it is 8 years. There are 508 censored hospitals and 491 uncensored hospitals on the ending observation. The purpose of this paper is going for survival and discriminant analysis of the hospitals in Taiwan. To find the affect factors that work upon the hospitals, the variables are present in the function of the survival and the proportional hazard regression model and the classification of the practice or non-practice condition and rank of the evaluation of the hospitals. There is statistical significance of the change of the hospitals ecological distribution to be checked under these rigorous markets and hostile environment. The analysis is simple, concrete and objective; and not subjective in the variables. Compute a function to evaluate the states of the operation of the hospitals for the private and public usage. Proceed to the next step to diagnose the fault of the practice of the hospitals. Then make up with tactics for the correction. The material and methods which were used in the theses are Kaplan-Meier procedures and Cox proportional hazard regression analysis and discriminant analysis, which are able to compute survival probability, hazard rate, classification of the groups of the rank of the evaluated hospitals and classification of the groups of the bankruptcy hospitals or not. The variables are classified to six domains which are the equipment(inputs),the amount of service (outputs), the financial affairs, the culture of the organization, the economy of the market, and government policies and laws and regulations. Each domain has its variable: the equipment including the number of beds, the number of the doctors, nurses and coworkers; the amount of the service including the number of outpatients, the emergency, the operations, hospital day; the financial affair including margin of sale profit, the turnover rate of asset and multiply equity; the culture of organization including the spirit of the enterprise, the thought of operation, the morale of the organization; the economy of market including the efficiency, the quality, the competition, the severity of disease, the contribution of the doctors and prosperity; the government policies, laws and regulations including the change of them. Almost all the data of the study are taken from DOC and NHI in Taiwan. For the principles of the simplicity, objectivity and specifity,the variables are almost taken from the secondary data. Although some evaluations are abstract, they are replaced with the relative data, For example, the prosperity being taken place of the rate of unemployment. The rate of the leaving from the jobs is on behalf of the spirit of the enterprise, the morale of the organization. The number of the hospital days is on behalf of the competition. About the financial domain, the severity of the disease and the contribution of the doctors are replaced with the efficiency. The survival interval is longer, the ratio of the closing the hospitals is higher, as well as hazard rate is. Under the affect of the Accreditation of DOC, the medical center has the highest survival rate and longest survival interval of the 5-rank-hospital ,by the order, the next is the regional hospitals, no-pass-accreditation hospital, district hospital and the last is the district teaching hospital; but the above phenomenon has been corrected since the accreditation of JACOH. The hazard model of the hospitals of existence is: h(t)＝[ho（t）] e^[（-0.953）×unemployed rate＋（-0.002）×total number of beds＋（0.090）×T_COV_*accreditation of DOC]. To classified the existence of the hospitals, the standardized discriminant function is: Score＝（-0.346）×accreditation of DOC＋（0.894）×unemployed rate. The correct classified rate is 92.3%.. To classified the 5 ranks hospital, The standardized discriminant functions has four. The variables have the number of beds, the number of the staffs, doctors, registered nurses, license nurse, inpatient number, outpatient numbers, the number of the emergency patients. The correct classified rate is 91.6%.. The study specially discusses the affected factors which work on the discriminant analysis.