| 摘要: | 懷孕期間藥物使用安全性與用藥錯誤是個重要議題。根據內政部生育年齡統計，懷孕女性年齡逐漸上升，高齡產婦比例增加，可能伴隨慢性病風險及長期藥物使用。這項研究利用2016年~2019年健康保險資料庫的門診就醫紀錄，篩選懷孕婦女後建立懷孕處方模組，進行疾病與藥物的關聯性分析。後續使用2016年~2020年臺北醫學大學臨床研究資料庫三家附屬醫院的懷孕婦女處方做驗證，並將處方設計成問卷後，邀請三位專家判斷疾病與藥物是否適當，並與健保資料庫的懷孕模組進行評估。挑選200張處方中，120張為模組判斷為適當，80張為不適當。統計結果，扣除13筆意見不一致的情況後，共有187筆處方被用於進行敏感度、特異度、陽性預測值、陰性預測值和F1 Score的分析。這200張處方包含了462種不同的疾病與藥物組合，其中23筆因意見不一致和3筆缺乏記錄而被排除，最終分析了436筆組合。在這436筆組合中，模組預測303筆為適當，133筆為不適當，而專家的判斷則是399筆適當，37筆不適當。這些結果被用來評估模組的準確性，其中以閾值Q>1.5的敏感度、特異度的表現最好。也針對不同案例進行討論，並詢問不同專家意見，供後續研究進行調整，使這套模組未來能更加完善，應用在臨床端。以期提升臨床用藥評估的準確性和安全性，從而降低用藥錯誤的風險，並致力於達到更安全的用藥環境。
The study focuses on medication safety and prescription accuracy during pregnancy. With an increasing trend of older maternal age, there's a rising concern for chronic diseases and long-term medication use in pregnant women. Utilizing health insurance data from 2016 to 2019, the study developed a pregnancy prescription model to analyze the relationship between diseases and medications in pregnant women. This model was validated with prescription data from 2016 to 2020 from three hospitals affiliated with Taipei Medical University. Among 200 selected prescriptions, the model deemed 120 appropriate and 80 inappropriate. After expert consultation and excluding inconsistent opinions, 187 prescriptions were analyzed for sensitivity, specificity, positive and negative predictive values, and F1 Score. Out of 462 disease-medication combinations, 436 were analyzed. The model predicted 303 combinations as appropriate and 133 as inappropriate, contrasting with experts' assessment of 399 appropriate and 37 inappropriate combinations. The best performance in terms of sensitivity and specificity was observed at a threshold value of Q > 1.5. The results will inform future refinements of the model, aiming to enhance clinical medication assessment accuracy and safety, thus reducing medication errors and striving to achieve a safer medication environment. |
