在日益重視病人安全的議題下，為保障用藥安全，各大醫療院所無不建置電腦自動篩選與監測系統，提供醫師開立處方時某種程度上的輔助，且無庸置疑的是這些篩檢警示對於用藥安全之確保，有其一定程度之提醒意義。藥物交互作用警示系統即是目前最熱門也具爭議的話題之一，故本研究針對衛生署藥品交互作用資料庫篩選出本院民國九十三年一月至九十四年十二月具一級交互作用之處方共8,160筆，刪除重複患者之處方後共415筆進行調閱病歷的審查，用以評估此系統之適當性。審查結果顯示：最常出現交互作用之藥物配對組合為aspirin-warfarin、amiodarone-warfarin、digoxin-furosemide及amiodarone- digoxin。最常出現交互作用之藥品為warfarin、amiodarone、digoxin、aspirin和furosemide，而這些藥物均為長期使用之處方，且這些配對組合針對特定疾病的治療效果業已經證實，因此，不考慮醫師處方行為之改變，例如：調整劑量、頻次或是監測相關副作用及臨床數據等，單單利用本系統全有全無篩選出來的交互作用筆數作為教學醫院評鑑的醫療品質參考指標是值得進一步討論的。

To ensure medication safety, a computer-based automatic screening or monitoring system provides a physician with certain assistance. However, too many alerts or warnings will cramp the prescribing behavior and paralyze the alertness of the physician; moreover, they might create some adverse effects in reliability of communication between physicians and patients. Drug-drug interaction alerting system may be one of the hottest but considerably controversial issues in recent years. Our study focused on outpatient prescriptions with potential drug-drug interactions during the period from January 2004 to December 2005, which were labeled as level one in significance rating and were screened from the database system of drug-drug interaction in the Department of Health, Executive Yuan, Taiwan. After deleting the duplicated ones, we reviewed the 415 out of the 8,160 medical charts of these cases and evaluated the appropriateness of the drug-drug interaction database system. Our study showed that the most common drug pairs with potential drug-drug interaction in our hospital were aspirin-warfarin, amiodarone-warfarin, digoxin-furosemide, and amiodarone-digoxin. In addition, medications which were most frequently involved in drug-drug interactions were warfarin, amiodarone, digoxin, aspirin, and furosemide. Moreover, these medications are always for long-term use and the therapeutic effects of those drug pairs with drug drug interaction have been demonstrated successfully in the management of specific diseases. In summary, the numbers of screening outcome by the drug- drug interaction database, using all or none technique, may not practically or adequately reflect the prescription behaviors of physicians or be referred to a reference indicator of medical healthcare quality in teaching hospital accreditation.