In Australia the usual evaluation of whether a student will perform adequately as a doctor is a subjective evaluation of his clinical performance, usually at the completion of five or six years at medical school. The evaluation is performed on an inadequate and uncontrolled patient sample and appears to be subject to many errors. Recent work investigating and attempting to define "clinical competence" has produced alternative ways of approaching the problem of evaluation. At the same time, new methods of testing competence have been developed and are being integrated into various certifying examinations. This report reviews the current situation in evaluating and defining clinical competence, and describes attempts to develop a model suitable for producing a certifying examination for Australia. It is hoped that the model may have a wider application for graduate trainees and practicing physicians. Appended are outlines of: categories of clinical performance; elements of the physician patient relationship; criteria for classifying physician actions; a definition of clinical competence; a statement of behavioral objectives for evaluation; terminal learning objectives of the Department of Medicine of the University of Manitoba; a "problem blueprint" for examination purposes; and test selection in the assessment of clinical competence. (MSE)