Using the decision tree to evaluate chronic disease refill prescriptions pharmacy strategy by National Health Insurance Research Databases

• Main Authors: Huang,Shiun-Yi, 黃薰儀
• Other Authors: Shen,Ching-Cheng
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/6xjh78

碩士 === 萬能科技大學 === 資訊管理研究所 === 105 === In order to reduce the frequency of medical treatment of chronic patients and non-essential health insurance payments, encourage people to receive medicine from the nearest pharmacies, implement the pharmaceutical hospitals, the province's medical institutions are promoting chronic disease prescriptions (refill prescription) evaluation calculation all around the island. This study was based on the 2000 National Health Insurance Database of the National Health Center, from 2004 to 2013. All the continuous prescriptions cases, including the hospital and the pharmacy were analyzed by using the technology of data explore decision tree. Sixteen input factors, they are hospital right, hospital grade, health insurance class, hospital location, medical treatment, disease type, daily cases, sex, age, job category etc., were employed to establish decision tree model which predicts the factors to select patients’ medical refill locations. The results show that "refill prescription" has increased year by year, accounting for about 39% of chronic diseases. The total number of prescriptions: take 2 times and 3 times over the calendar year are 43.6% and 56.4%, respectively. The 3 times prescription has a little increase trend, which "hospital" level had higher average ratio, and the medical center grabs the highest rate of 2 times prescription. The prescriptions proportion of the hospital or the pharmacy was 59.8% and 40.2%, respectively, and the prescription of hospital came mostly from back to the hospital, meanwhile the rate of to the pharmacy was also increased gradually. It is worth noting that from the center hospital changed to the pharmacy increased apparently, Tainan medical center got the highest ratio of all. The decision tree analysis shows that patients return to the hospital or to the pharmacy was affected by the plan very much. Earlier than 2010 (second wave), most hospital level prescription were back to the hospital, the situation was changed by the factor of ''city'' beyond 2010. In addition, the study also found that the total number of prescriptions and re-swap sites have inter-influenced, the 2 times prescriptions prefer going back to hospital while 3 times prescriptions going back to the pharmacy.