Assessing Physicians Peer Review System by Using Data Mining Techniques

• Main Authors: Tsu-Yi Tso, 左祖儀
• Other Authors: Ching-Wen Chien
• Format: Others
• Language: zh-TW
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/90704271469930080393

碩士 === 國立陽明大學 === 醫務管理研究所 === 91 === National Health Insurance (NHI) has been implemented since 1995. Nowadays, the most serious problem the NHI has is its financial crisis. To select the abnormal situations and control the medical expenditure, the NHI urgently needs to establish a good peer review model. By data mining technology, the NHI can analyze the databases built by collecting past files physicians have provided and then pick up the unknown problem patterns. Through analyzing and classifying the data files, the NHI will be able to recognize the characteristics and attributes of different physician groups and apply its discovery to the peer review model. The purpose of this study includes (1) establishing a quality and quantity control method to improve the present peer review; (2) using data mining technology to discover the characteristics and backgrounds of physicians that were in the different claims of medical services; (3) explaining how to find out the anomalous claims of medical services by using certain related indicators. The study was based on the NHI’s claimed and reclaimed reimbursement database and doctors’ personal information from 2000 to 2001. The conclusion of the study indicates that: (1) the different claims of medical services of physicians, quality and quantity control methods, are separated into 17 groups by the decision tree analysis method; (2) the situations of claims of medical services have no relationship with the hospital ownerships or medical specialists; (3) each physicians’ probability of claim are unequal. The study suggests that the NHI and hospitals can employ quality and quantity control methods and abnormal sampling audit indicators to quickly find out the physicians who anomalously claim medical services.