Developing a virtual coach for chronic patients: A user study on the impact of similarity, familiarity and realism

Abstract

Healthcare costs are increasing dramatically due to disproportional consumption of healthcare resources by chronic patients. Automated forms of health coaching can contribute to improved patient self-management while reducing costs due to increased scalability and availability of the use of human health coaches. Embodied Conversational Agents (ECAs) seem to be good candidates to function as automated coaches, as they introduce a social component to human-computer interactions which makes them particularly suitable to influence a user’s attitude or behavior. To date, there is limited knowledge on the impact of appearance-related characteristics of an ECA as a virtual coach among a chronically ill elderly population. The primary aim of this study is to investigate the impact of three appearance cues on user acceptance: (i) similarity; (ii) familiarity; and (iii) realism. Findings demonstrate that patients (a) preferred the realistic-looking ECA over the more stylized one; (b) showed no preference for the familiar over the unfamiliar ECAs (but did evaluate the unfamiliar ECAs as more positive than the familiar one); and (c) evaluated an ECA as virtual coach for self-management support as useful.