Assessment of fitness for surgical procedures and the variability of anaesthetists' judgments.

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.