Generating monologue and dialogue to present personalised medical information to patients

Abstract

Medical information is notoriously difficult to convey to patients because the content is complex, emotionally sensitive, and hard to explain without recourse to technical terms. We describe a pilot system for communicating the contents of electronic health records (EHRs) to patients. It generates two alternative presentations, which we have compared in a preliminary evaluation study: the first takes the form of a monologue, which elaborates the information taken from the patient's EHR by adding explanations of some concepts and procedures; the second takes the form of a scripted dialogue, in which the content is recast as a series of questions, answers and statements assigned to two characters in the dialogue, a senior and a junior nurse. Our discourse planning method designs these presentations in tandem, first producing a monologue plan which is then elaborated into a dialogue plan.

Adaptive Generation of Multilingual Questions and Answers from Web Content

“…There are already relevant works that combine NL generation and UM techniques for generating personalized dialogues in a particular domain, such as that described in [14], in the medical domain. NL and UM are also used for the adaptive presentation of dynamic hypertext in the information-systems Ilex and PEBA-II (described in [15]), that dynamically generate pages, according to the user profiles, from canned text with various types of annotations and items from a knowledge base.…”

Cross-cultural assessment of automatically generated multimodal referring expressions in a virtual world

“…We employed a Fully Generated Scripted Dialogue (FGSD) approach (Andre´et al, 2000;Williams et al, 2007) to evaluate the output of the Van der Sluis and Krahmer algorithm. With FGSD entire dialogues are produced by one generator.…”

Evaluating an Algorithm for the Generation of Multimodal Referring Expressions in a Virtual World: A Pilot Study

“…This allows for complete semantic and pragmatic transparency, which is important for a content determination task like the generation of referring expressions. In recent years, an alternative paradigm for computational work on agents has emerged ( [1], [23]), with which entire dialogues are produced by one generator. Initially, scripted dialogues made heavy use of canned text, but recently this approach has been integrated with Natural Language Generation techniques, resulting in the Fully Generated Scripted Dialogue (FGSD) ( [18]; [14]).…”