Thompson W G, Lipkin M, Gilbert D A, Guzzo R A, Roberson L
Department of Medicine, New York University School of Medicine.
J Gen Intern Med. 1990 May-Jun;5(3):214-7. doi: 10.1007/BF02600537.
The American Board of Internal Medicine suggests use of a standard form to rate residents on nine dimensions (such as clinical judgment and overall clinical competence) on a scale of 1 to 9. The authors examined the psychometric evidence for reliability and validity of 1,039 ratings of 85 residents by 135 attendings in a single internal medicine residency program. Of these ratings, 95.6% were from 6 to 9. Factor analysis revealed that high correlations among the nine dimensions (r ranged from 0.72 to 0.92) resulted from a single global factor accounting for 86% of the variance. The study also examined whether the form reliably distinguishes among residents scoring between 6 and 9. Agreement among attendings rating the same individual was weak (average reliability = 0.64, by the method of James). The rating method fails to discriminate dimensions of clinical care and has low reliability for distinguishing among competent residents.
美国内科医学委员会建议使用一种标准表格,对住院医师在九个维度(如临床判断力和总体临床能力)上进行评分,评分范围为1至9分。作者在一个内科住院医师培训项目中,检查了135名主治医师对85名住院医师的1039份评分的信度和效度的心理测量学证据。在这些评分中,95.6%为6至9分。因子分析显示,九个维度之间的高相关性(r值范围为0.72至0.92)是由一个单一的全局因子导致的,该因子解释了86%的方差。该研究还考察了该表格是否能可靠地区分评分在6至9分之间的住院医师。主治医师对同一住院医师的评分之间的一致性较弱(采用詹姆斯方法,平均信度=0.64)。该评分方法未能区分临床护理的维度,且在区分合格住院医师方面的信度较低。