A Medical Profession Review Research for Oversea Emergency Reimbursement Using Logistic Regression and Two-stage Integrated Models: Based on The Bureau of National Health Insurance - Taipei Devision

• Main Authors: Yu-Ju Yen, 嚴玉茹
• Other Authors: Chien-Ho Wu
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/88911895822467874362

碩士 === 輔仁大學 === 應用統計學研究所 === 99 === The overall political and economic environment have changed a lot in recent years, people has more opportunity to go abroad more frequently for business, tourism, visiting relatives and other reasons, therefore, the application cases for NHI(National Health Insurance) overseas emergency reimbursement medical expenses are also increasing. Facing the growing complexity and high volume of application cases, how to reduce the error of administrative review by subjective judgments and increase the accuracy and how to establish a good pattern of medical profession review and data segment to pay on time are getting crucial. Therefore, this study, based on NHI data, analyze the source of application cases, medical institutions location and the distribution of personal characteristics, growing payment of medical expenses, and make use of the decision tree to mining the relationship between medical profession review and the independent variable group, Screening the key variables and affecting important group of independent variables to strengthen the management level, in the meantime, the study also build a reimbursement medical expenses predictive model by using logistic regression one-stage and two-stage integrated decision tree to improve the accuracy of payment jobs.The conclusion of the study indicates that Mainland China has the most overseas emergency reimbursement case and its ratio is 77.0% in which Taiwanese capital hospitals in China ratio is 32.0% and major place of medical treatment is Shanghai and Jiangsu that ratio is 49.7%. The research findings indicate that the important variables of decision tree are the number of medical treatment, medical illness, payment conditions. The two models, logistic regression one-stage and two-stage integrated decision tree, have the same forcasting accuracy and prediction capability.