醫學中心作者之學術研究評比與排名

Assessment and Ranking in Research Achievement among Taiwanese Medical Centers and Authors

醫學中心六大任務的評鑑囑咐：受評醫院「應著眼於該任務之表現，且儘量陳述醫院在該指標比其他醫院特殊之處」。對於醫院著作的成果比較，尚無一個醫院間比較的依據。本研究蒐集Pubmed資料庫檢索內國內19家醫學中心2003至2007年5年間上萬篇的SCI論文，搭配其影響係數，評比與排名國內醫學中心的學術水準。利用試題反應理論的Rasch模式，檢測資料評量的單維性。以殘差之因素負荷量，分析醫院間在件數與影響係數上的特徵異同。採用羅倫茲曲線及吉尼係數，檢測5年來醫院間著作的表現是否漸趨平均。排名論文第一作者、期刊與其類別，並估計各「作者的著作影響係數」。研究發現：(1)評量方法與標的之不同，會有不同的檢定成果；(2)SCI件數與其影響係數在19家醫學中心間漸趨均衡分佈；(3)作者的著作影響係數，平均0.26（標準差0.11，最低0.14，最高1.34）。被刊登的SCI期刊最多的是台灣之：J Formos Med Assoc。

The Six Missions of hospital accreditation for medical centers in Taiwan are urged to present data as much as possible in comparison with other counterparts. No any platform and sufficient data can be obtained to compare paper publication among hospitals. First, we downloaded 11,600 research papers derived from Taiwanese medical centers in past five years via Pubmed journal database. Then, Impact factors of journals were collected to evaluate and rank the academic abilities of hospitals and authors. The Rasch model was applied to examine the data unidimensionality. Standardized residuals of factor loading were used to assess the similarity among hospitals. Lorenz curve and Gini coefficient were introduced in this study for investigating the equality of paper publication in past years among medical centers. In addition, the impact coefficients of individual authors were estimated. The results reveal that 1) different criterion used for evaluation will produce different outcomes, 2) the equality of paper publication in past years among Taiwanese medical centers gradually exhibited even, 3) the author impact coefficients averaged at 0.26 (5D=0.11, lowest=0.14, highest=1.34). J Formos Med Assoc, was demonstrated to be the most popular journal for authors in Taiwanese medical centers.