| Summary: | 碩士 === 國立中正大學 === 資訊管理所 === 96 === The Follow-up Visit Model for Patients of Diabetes Mellitus
－A study by using Neural Network Approach
Purpose:
The purpose of the paper is to detect the factors of significance and weight of Diabetes Mellitus patients’ follow-up visit model and find out the order of these important factors. To construct the adequate follow-up visit model in accord with the planning of Bureau of National Health Insurance effectively saves medical resource along with promoting the performance of management and improves the service quality for Diabetes Mellitus patients.
Method:
From Jan.1, 2006 to Dec.31, 2006, the usable 2007 samples of Diabetes Mellitus patients for one year is derived from the database of a regional teaching hospital in Chia-yi. The collected data were analyzed by comparison Neural Network analysis with Logistic regression to find out the better statistical method to predict the actual value to verify the assumptions made in the study.
Conclusions:
After comparing the accuracy of predicted probabilities, the value of neural network approach is 83.1589% better than logistic regression 76.7%. Between one year, the follow-up visit record and medical amount of September(the ninth month) will be the extremely significant factor to impact the follow-up visit record and medical amount of next 3 months(October to December). In order to promote better medical service, the hospitals need to enhance the regular follow-up visit of Diabetes Mellitus patients to avoid from the impact on quality of life from the disease. The future research suggested that further statistical analysis should be practiced to find out the related factors of Diabetes Mellitus patients for adequate medical care utilization and treatment to help Diabetes Mellitus patients better their living quality and satisfaction.
Keyword: Follow-up Visit Model; Neural Network Approach; Diabetes Mellitus
|
|---|
