Font Size: a A A

The Nested Logit Model And Its' Application In The Analysis Of Health Service Utility

Posted on:2007-06-14Degree:MasterType:Thesis
Country:ChinaCandidate:R BaiFull Text:PDF
GTID:2144360212465988Subject:Epidemiology and Health Statistics
Abstract/Summary:PDF Full Text Request
Objective The aim of the study is to research and compare the application of multinomial logit model and nested logit model among discrete choice model in health seeking behavior of rural and urban residents in Zhejiang province, to select the most appropriate model, to discuss the factors of affecting health seeking behavior of rural and urban residents in Zhejiang province, and then to provide references for studying and making relevant health policies and adjusting health system.Method Research data is acquired from database of questionnaire about family health in Zhejiang province in health service survey in 2003. The research object is composed of informants whose age is above 15 and who loose their forms in the two weeks before the survey begin. The research fits two different health seeking model using the two discrete choice models above, compare the economics hypothesis, evaluation of statistics about model fitting and the application of the two models. The logistic regression model is used to analyze the factors of affecting whether to seek health care when someone is ill. The soft SAS 9.0 and STATA 8.0 is used in the research.Result The multinomial logit model of health seeking behavior of rural and urban residents can not satisfy the independence from irrelevant alternatives hypothesis. The MLR of the multinomial logit model and the nested logit model about rural residents is 603.97, the Adjusted Estrella values are 0.61 and 0.90, and the accuracy rate of prediction are 60% and 70%. The MLR of the multinomial logit model and the nested logit model about urban residents is 138.54, the Adjusted Estrella values are 0.48 and 0.91, and the accuracy rate of prediction are 65% and 75%.Conclusion The nested logit model is more appropriate than the multinomial logit to analyze the health seeking behavior of rural and urban residents in Zhejiang province. The factors affecting the decision of rural residents of whether to seek health care when ill are job, health insurance, per capita living space, the kind of washroom, workout in spare time, to gain knowledge actively and degree of severity of disease. The factor affecting the decision of urban residents of whether to seek health care when ill is the kind of job and the degree of severity of disease. Health insurance, the degree of severity of disease, acute disease or chronic disease, degree of richness of household, the time cost of seeking health care and the quality of health providers are the leading factors to affect the urban or rural residents'selection of health providers in Zhejiang province.
Keywords/Search Tags:Nested Logit Model, Multinomial Logit Model, Health Service Utility, Affecting Factors
PDF Full Text Request
Related items