ABSTRACT
Computer-based systems for medical diagnosis have several potential uses, including consultation, tutoring, and evaluation. Each of these uses has profound implications for the design of an underlying expert system’s human-machine interaction capability. An awareness of these issues has led us to study the person– person interaction process that takes place in teaching rounds. We believe one of the most significant observations to emerge from our analyses is the presence of multiple levels of diagnostic activity. We propose a representation and process model that defines a general framework for diagnosis, based on the synthesis of a coherent interpretation of multiple (and often inconsistent) phenomena. A discussion of the implications of this model for the development of multiple levels of diagnostic activity within human-machine interaction concludes this chapter.
