Wilson M E, WIlliams N B, Baskett P J, Bennett J A, Skene A M
Br Med J. 1980 Feb 23;280(6213):509-12. doi: 10.1136/bmj.280.6213.509.
Ten anaesthetists were asked to make judgments on fitness for elective operation on data derived from 200 patients. The extent of their agreement was measured using a kappa statistic, and clusters of anaesthetists who agreed well with each other were identified. Using an alternative technique, the "true" fitness category of each patient was estimated using a maximum likelihood method which estimated the error involved in making judgments on limited amounts of information. It was possible to compare the performance of each anaesthetist against the consensus and to measure deviation on an "optimism--pessimism" continuum. A simple questionnaire predicted fitness for operation by all 10 anaesthetists in 96% of cases.
十位麻醉师被要求根据200名患者的数据对择期手术的适宜性做出判断。使用kappa统计量来衡量他们的一致程度，并确定彼此意见高度一致的麻醉师群体。使用另一种技术，通过最大似然法估计每位患者的“真实”适宜性类别，该方法估计了基于有限信息量做出判断时所涉及的误差。可以将每位麻醉师的表现与共识进行比较，并在“乐观——悲观”连续体上衡量偏差。一份简单的问卷在96%的情况下预测了所有10位麻醉师对手术适宜性的判断。
