台灣醫療機構的資訊系統已經結合臨床決策支援系統與電子處方開立功能，能適時提供重要用藥提示以避免不適當處方，保障病人用藥安全；然而，儘管建立預防疏失的提示機制愈來愈受重視，提示的內容、設計與執行過程，仍欠缺整體性的了解與評估。本文目的是分析2007 年1 月1 日至2010 年12月31 日間的藥師服務通報紀錄，呈現門、急診重複處方經歷三階段：醫師端開方提示(9/2007)、藥師即時線上(on-line)審核(2/2008)及藥師端旗子(2/2009)提示機制的通報結果、常見重複用藥的型態、重複處方發生的原因與執行上的挑戰。依據藥師通報對醫師所做處方修改建議的紀錄，接受藥師建議的通報件數由1/2007-6/2007之34件(通報率為0.01件/1000發藥筆數)逐漸增加至1/2010-12/2010之644件(0.19件/1000發藥筆數)。重複處方中最常見的藥品類別依序是降血壓等類(11.3%)、治療胃酸分泌異常(8.1%)、鎮靜安眠藥物(7.2%)。常見專科別依序是神經內科系(13.0%)與胸腔內科系(12.9%)、心臟內科(9.1%)、新陳代謝科(7.3%)與兒童內外科(7.2%)。處方重複發生原因，以不同專科在相同或不同日開立相同藥物為主(41.6%)。從醫師未修改處方的理由中發現，病人因素(27.6%)是重複處方無法避免的狀況。除了提示內容的完整性與專一性之外，為避免醫療人員產生提示疲乏，提示閥值或條件的設定、提示的技巧，需要依據重複處方發生原因、並考量使用者實際操作的便利性，方能確保臨床決策支援系統的接受度與成效。另外，醫師與藥師端的雙方提示，也是非常重要降低重複處方發生的機制。

The use of computerized prescribing and clinical decision support to reduce medication error is acommon element of medication safety policy. Although the integrated system has shown beneficial effects on thequality of professional performance, efficiency and safety in patient care in western countries. At present, thepotential benefit in Taiwan health care system remains unclear. This paper is to analyze the reasons, sources anddrug classifications of one type of inappropriate medication uses- duplication of prescriptions in ambulatory caresetting through spontaneous reports from pharmacists and to explore the potential strategies to improveefficiency of prescription alerts implemented to avoid duplicate prescriptions. Pharmacist’s spontaneous reportswere based on the recommendation that is made to suggest physicians to avoid duplication in prescriptions. Thecomputerized alerts system has been modified and improved over time in a large medical center; the first stageof alerts system only focusing on physician in January 2007, following by making medication history accessibleto pharmacists (February, 2008) and then flagging duplicate orders to pharmacists (February, 2009 ). Of total1,671 duplicate prescribing orders were reported and analyzed during January 2007 and December 2010, theduplicate medication reporting rate increased from 0.01 order/1000 prescription dispended orders (34/3,426,688)in 1/2007-6/2007 to 0.19 order/1000 prescription dispended orders (644/3,476,480) in 7/2010-12/2010. The mostcommon drug class among duplicate prescriptions was anti-hypertensive agents (11.3%), drugs used for aciddisorders (8.1%) and psycholeptics (7.2%). The most duplication of prescriptions came from neurological(13.0%) and respiratory (12.9%) specialists, cardiovascular (9.1%), endocrine (7.3%) and pediatric specialists(7.2%). The reasons of duplication mainly caused from different physicians prescribed the same drug at differenttime of clinic visits (41.6%). From reported data, we realized some duplicate prescriptions are due topatient-related concerns (27.6%) and may not be avoidable, such as clinic visit earlier than scheduled date, andrequiring additional drug amount for traveling. As yet there is little evidence that decision support is effective inchanging patient outcome, the progress was made in this study on the development and adoption of alerts systembased on our understandings of the causes of duplication and involvement of physicians, and to ensureimproving preventable inappropriate prescriptions patient safety in the future. Lastly, the two-ways alert(physician and pharmacist) was a critical double-checking mechanism to avoid a duplicate prescription.