Spoken Input for Medical Diagnostic Systems.
Lawrence M. Fagan, Wu A, Nathan A. Harris, Ams Poon, Smadar Shiffman · PubMed Central · 1990
One of the major limitations of decision support systems has been the flexibility of the user interface. One possible mode of interaction that has not been heavily explored for decision support systems is the use of computer recognition of spoken language. We are exploring possible designs for spoken interfaces to medical diagnostic systems such as QMR. The goal of the project is to allow for a flexible mapping between the findings (signs, symptoms, laboratory results) that a particular physician has in mind with the specific terminology expected by the diagnostic system. We are using a combination of graphical prompting and spoken phrases to identify the specific findings that can be used in the differential diagnostic process. We have implemented one version of this approach using Hypercard and Voice Navigator (an isolated word speech-input system) on the Macintosh.