目的：以資料包絡分析法評估出醫院臨床科別經營績效的優劣，及逐年改善情形。 方法：以資料包絡分析法(Data Envelopment Analysis, DEA)爲方法，評估某區域教學醫院臨床各科的經營績效。決策單位包括内科組7個臨床科，外科組12個臨床科，利用93及94年度兩個投入變項（醫師人數、護理人數）與3個產出變項（門急診人數、住院人日、醫療收入）的資料進行分析，以DEA中CCR模式(Charnes Cooper Rhodes mode)評估該醫院各年度及跨年度各單位的相對效率值，並作内科組與外科組相對效率的比較，之後以Malmquist指數做跨期效率分析，探討技術移動及追趕程度。 結果：内科組跨年度平均效率值爲96.25％，外科組跨年度平均效率值為84.03％，該院的内科經營績效優於外科。94年度全部臨床科別的平均效率值爲92.29％，其經營績效亦比93年度的87.28％優異。依照Malmquist指數計算，有14個科都逐年改善。 結論：以DEA投入導向的CCR模式，分辨出各臨床科別經營績效的好壞，以及Malmquist指數看出逐年進步與退步的趨勢，將有助於決策者對投入資源的正確分配，以達整體規模的最適境界。

Objectives: To evaluate the operative performance and efficiency of clinical sections and its improvement year by year using Data Envelopment Analysis (DEA) Methods: We evaluated the operative performance of each clinical section in a regional teaching hospital using the DEA method. The Decision Making Units (DMU) includes 7 sections of the internal medicine group, and 12 sections of the surgical group. The study was conducted using data in years 2004 to 2005. The two input variables were the number of attending doctors and nurses, and the three output variables were OPD and emergency case number, admission day medical income. The aforementioned data was analyzed with the CCR model of the DEA method to obtain the relative efficiency scores of various sections in the hospital, after which a comparison was made between the relative efficiency score of sections in the internal medicine group and those belonging to the surgical group. Finally, we discuss the shift in technology and catching up in efficiency by calculating cross-year efficiency score according to the Malmquist index. Results: The average efficiency score of the internal medicine group in the cross-year was 96.25%, and the average efficiency score of the surgical group in the cross-year was 84.03%. This result indicates that the operative performance of the internal medicine group in this hospital is higher than that of the surgical group. The average efficiency score of all the clinical sections in the year 2005 was 92.29%, which also indicates a higher operative performance than the 87.28% in 2004. From calculations according to the Malmquist index, we can see that there are 14 sections whose performance is increasing year by year. Conclusions: By using the CCR model of DEA, we can distinguish the performance of each clinical section; by suing the Malmquist index, we can follow trend of improvement or recession year. There results can help the decision-maker of the hospital use resources allocation correctly to achieve the optimal level of the institution.