目的：本研究旨在探討南部某醫學中心護理人員給藥錯誤情形及其影響之相關因素。 方法：本研究爲回溯性相關研究。研究樣本爲2003年8月至2006年7月護理人員每月主動通報之給藥錯誤異常事件共317人次，其中有229人發生給藥錯誤。 結果：研究結果顯示護理人員給藥錯誤率爲8.78%，平均錯誤次數爲1.38±0.77次，最高爲5次。以年齡層在20-24歲、三個月內試用期及NO層級的新進人員最多。給藥錯誤率呈現不穩定狀態與人力不足及新進人員大幅增加可能有關。錯誤原因以執行「三讀五對」過程中出錯比率最高，其次爲因藥物外型或藥名相似、多種劑量等辨識不足。在給藥辨識項目以藥物的錯誤最多，而藥物類型以針劑類抗生素的錯誤最多，在白班錯誤比率高於其他班別。給錯藥對病人的傷害，經嚴重度分級評估，受到中度以上傷害需持續治療者有5.36%。護理人員的年資、科別、進階層級及班別與錯誤次數有顯著差異。外科病房及N進階層級人員是影響給藥錯誤之預測因子。 結論：護理人員給藥時應確實遵循「三讀五對」原則。護理主管應加號新進人員給藥標準作業執行及對藥物認識，針對錯誤頻率較高的單位及容易出錯人員列爲優先輔導對象。醫院應致力系統方面等問題之改善，以降低護理人員給藥錯誤風險。

Objective: This study aimed to investigate the status of and factors related to the occurrence of nursing medication errors (NMEs) in a medical center in southern Taiwan. Methods: A total of 317 adverse events reported voluntarily by 229 nurses from August 2003 to July 2006 at this medical center were analyzed. Descriptive and correlation analysis was performed to identify factors associated with the occurrence of NMEs. Results: The NME occurrence rate was 8.78%. The mean±SD frequency of NMEs for the reporting nurses was 1.38±0.77 times, with a maximum frequency of 5 times. Nurses who were aged 20 to 24 years, who were still completing the 3-month probation period, and who were at the N0 stage of the clinical ladder had higher NME rates than the other nurses. The NME rate reported to the hospital varied from month to month, in association with nurse workforce insufficiency and surges of recruitment of new staff during some periods. The most frequent cause of NMEs was lack of compliance with the ”three-reads, five-rights” process, followed by confusion of drugs with similar names and appearances. Among drug types, antibiotics, especially those given in injection form, were most commonly involved in errors. Nurses who worked the day shift had a higher NME rate than those who worked other shifts. Regarding the consequences of medication errors, 5.36% of patients suffered from at least a moderate degree of injury for which they needed treatment. The nurses' number of years of experience, position on the clinical ladder, units, and shifts were significantly correlated with NME frequency. Working in the surgical ward and being on the N clinical ladder were significant independent variables predicting the frequency of NMEs. Conclusion: Based on our results, we suggest that hospital nurses strictly follow the ”three-reads, five-rights” process when administering medicine to patients. Improving new nurses’ drug-identification ability and compliance with standard operating procedure is an important issue for the clinical nursing administration.