ACCEPTABILITY OF AN EMBODIED CONVERSATIONAL AGENT FOR TYPE 2 DIABETES SELF-MANAGEMENT EDUCATION AND SUPPORT VIA A SMARTPHONE APP: A MIXED-METHODS STUDY.

Session Name
HUMAN FACTOR IN THE USE OF DIABETES TECHNOLOGY
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:32 - 09:33
Presenter
  • Shaira Baptista, Australia
Authors
  • Shaira Baptista, Australia
  • Greg Wadley, Australia
  • Dominique Bird, Australia
  • Mark Harris, Australia
  • Michaela Riddell, Australia
  • Anthony Russell, Australia
  • Paul Scuffham, Australia
  • Emily Williams, United Kingdom
  • Brian Oldenburg, Australia
  • Jane Speight, Australia

Abstract

Background and Aims

Embodied conversational agents (ECAs) have the potential to deliver self-management support to people with type 2 diabetes (T2D). We evaluated the acceptability of an ECA, ‘Laura’ (see figure) for delivering self-management support in the ‘My Diabetes Coach (MDC)’ smartphone app.

embodied conversational agent ‘laura’.jpg

Methods

In a randomised controlled trial, participants allocated to use the MDC app completed a 6-month post-baseline survey assessing attitudes to and interactions with Laura. In-depth qualitative interviews further explored users’ experiences of Laura. Using an explanatory mixed-methods approach, we analysed survey responses with descriptive statistics and integrated these findings with a thematic analysis of the interview data.

Results

Of the 67 survey respondents, 33 (50%) were women, aged 57±9 years. Most (77%) endorsed positive descriptions of Laura (e.g. competent) and their reactions, following interactions with her (e.g. motivated). Fewer respondents (35%) described Laura negatively (e.g. boring) and a minority (12%) reported negative reactions to interactions with Laura (e.g. frustration). Interview participants (N=19) included 8 (42%) women, aged 60±8 years. Four themes emerged regarding Laura’s acceptability: 1) Laura’s perceived role as a “friendly coach”; 2) Laura’s “value add” in providing motivational and emotional support; 3) Laura’s “human-like” nature being preferred to a non-human character; 4) Dissonance in Laura’s speech and body language is frustrating.

Conclusions

Overall, these findings suggest that an ECA is an acceptable means for delivering T2D self-management support. Improving acceptability will require a better understanding of the role that users expect an ECA to play in self-management and perfecting the ECA’s ‘natural’ communication.

Hide