Wayne Diane B, Barsuk Jeffrey H, Cohen Elaine, McGaghie William C
Department of Medicine, Northwestern University Feinberg School of Medicine, Galter 3-150, 201 East Huron Street, Chicago, IL 60611, USA.
Acad Med. 2007 Oct;82(10 Suppl):S105-8. doi: 10.1097/ACM.0b013e318141f464.
Clinical skills examinations (CSEs) are used frequently in medical education. Standard setting for CSEs may employ different methods with or without prior performance data.
An expert panel provided item-based (Angoff) and group-based (Hofstee) judgments about two central venous catheter insertion performance checklists on three occasions. Judges did not receive baseline performance data on the first occasion but did on occasions two and three. Judges' ratings were used to calculate a minimum passing standard (MPS) for the CSE. Interrater reliabilities and test-retest reliability (stability) were calculated. Passing standards are compared using performance data from a pilot study.
Both methods produced reliable and stable data. Baseline data influenced the judges' decisions. Use of the Angoff method alone yielded lenient MPSs, whereas the Hofstee method alone yielded stringent MPSs.
Standard setting is a critical component of CSEs. Baseline data influence judges' decisions. Averaging Angoff and Hofstee outcomes produced the optimal MPS.
临床技能考试（CSEs）在医学教育中频繁使用。CSEs的标准设定可能采用不同方法，有无先前的表现数据均可。
一个专家小组三次就两份中心静脉导管插入操作检查表给出基于项目（安格夫法）和基于组（霍夫斯泰法）的判断。评判员在第一次未收到基线表现数据，但在第二次和第三次收到了。评判员的评分用于计算CSE的最低及格标准（MPS）。计算了评分者间信度和重测信度（稳定性）。使用一项试点研究的表现数据比较及格标准。
两种方法均产生了可靠且稳定的数据。基线数据影响了评判员的决策。单独使用安格夫法得出的MPS较宽松，而单独使用霍夫斯泰法得出的MPS较严格。
标准设定是CSEs的关键组成部分。基线数据影响评判员的决策。将安格夫法和霍夫斯泰法的结果平均得出了最佳MPS。
