A report is made of a computerized medical consultation system that prompts the physician user for symtoms related to a particular class of ailment. The paper focuses on an assessment of human factors issues in the use of the system. Two distinct patterns were noted during videotaped consultations: minimal computer use during the consultation, and conversational use of the terminal. Possible problems with response time, length of consultation sessions, coverage of symptom topics, errors in categorization of symptoms, and sequence of topic discussion are evaluated. A significant conclusion was that computer-aided consultations followed a similar pattern to that of precomputer consultations.