Differential Diagnosis Listing as Relevance Feedback: An Essential User Interface for Clinical Decision Support Systems

Takashi Okumura, Tomoko Kajiyama, Noboru Sonehara

Abstract: The user interface of clinical decision support systems (CDSSs) has not yet been studied well. In a previous study, we investigated the presentation of diagnostic output and proposed the hierarchical representation of differential diagnosis lists that provides an effective way of interfacing with a CDSS. In this research, we propose the incorporation of relevance feedback to improve the user interface. Such feedback can be used in three ways: to sharpen the search query, to initiate a dialogue with the user with regard to the presence or absence of particular diagnostic information, and to improve the accuracy of the diagnostic algorithm. This research qualitatively evaluates our approach and demonstrates that relevance feedback has desirable characteristics as a user interface element for clinical decision support.