| Summary: | 碩士 === 輔仁大學 === 統計資訊學系應用統計碩士班 === 103 === It has been noticed that the number of patients having bladder cancer is increasing in Taiwan in recent years. Patients with bladder cancer can survive longer than those with other cancers, but higher recurrence rates for these patients are also observed. Therefore, utilization of the medical resources of patients with bladder cancer (UMRPBC) should have some different characteristics from that of other cancer patients. In this thesis, using the National Health Insurance Research Database of Taiwan the related factors and forecasting models for UMRPBC were investigated. Patients of age above 20 years with ICD-9-CM code 188 and being hospitalized at least one day were included in the study. The results show as below. First, area is the determinant factor of duration hospitalization. And hospital ownership and area are the determinant factors of the total cost of hospitalization. Second, in the Cross-sectional model of duration hospitalization, the preferred approach is Johnson & Wichern method, but in the total cost of hospitalization, the preferred one is Hosmer & Lemeshow method. Last, in longitudinal model of duration hospitalization, the best prediction model is SARIMA(1,0,2)*(2,0,1)12. And the best prediction model for the total cost of hospitalization is SARIMA(2,1,1)*(2,1,2)12.
|
|---|
