Collecting and Interpreting Temporal Data in an Expert System

The collection and interpretation of temporal data in the cardiovascular domain is particularly important because the temporal characteristics of symptoms can provide important clues to the appropriate diagnosis. It is challenging because findings may be single events, continuous states, episodic events, or a progression described by more than one finding. This paper relates the nature of the problem and our experience over the past year as physicians have entered cases into the Heart Disease Program.