碩士 / 淡江大學 / 數學學系 / 92 / Up to now, there is no rapid and accurate way to diagnose diabetes. It will take up to 3 examination procedures to verify whether he/she does suffer from the diabetes. To save the precious medical resources and to increase the treatment effect due to early diagnosis, we use logistic regression to establish a prediction model, which is also according with the“Clinical Prediction Rules”proposed by Wasson et al（1985）in the New England Journal of Medicine. A total of 1033 patients who were suspected to be diabetic and accomplished three days’ physical examination in some medical center from Jan. 1990 to Mar. 1992 and available from the medical records would be collected. Among them, 2/3 of the patients (672 patients) were randomly selected to serve as the“Training set”for establishing a prediction model. The other 1/3 patients served as the “Testing set”for verifying the accuracy of the established prediction model. We also established a grading system with a “best” cut-off point based on the prediction model and to satisfy the characteristic of easy to be used for prediction. The accuracy of the proposed prediction method was evaluated by the sensitivity, specificity and correct prediction rate. The proposed prediction model will provide a rapid and accurate method for doctors to diagnosis diabetes and to determine whether a further examination is needed or not.
| Identifer | oai:union.ndltd.org:TW/092TKU00479010 |
| Date | January 2004 |
| Creators | Yi-Tai Wan, 萬義泰 |
| Contributors | Yue-Cune Chang, 張玉坤 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | zh-TW |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 29 |
Page generated in 0.019 seconds