A social robot as a means to motivate and support diabetic children in keeping a diary

Abstract

This study is part of the European FP7 ALIZ-E project. The goal of this project is to develop the theory and practice behind an embodied social robot that is capable of maintaining affective interactions with young users over an extended and possibly discontinuous period of time. The target audience of this research consists of young patients suffering from diabetes mellitus. In previous work on ALIZ-E, several activities have been developed that the user can carry out with the robot, such as playing a quiz, math game or imitation game. Research has shown that monitoring one’s diabetes by keeping a diary is beneficial to the child’s health on the short and long term. In this master thesis, we develop a new activity that allows a diabetic child to keep an online diabetes diary together with the robot. We evaluate this new activity by investigating the following research question:

How can a robot contribute to the activity of keeping a web-based diary together with diabetic children?

We present a comprehensive robot dialogue model that has two main foci: 1) supporting and motivating the child in the task of keeping a diary and 2) forming a bond between the child and the robot through personalized interaction. We tested the dialogue model by performing a pilot experiment in a real-world setting using a within-subjects design (N=6). The subjects were diabetic children between the ages of 9 to 12. Over the course of eleven days, they kept an online diary in which they logged their blood glucose values, carbohydrate intake, insulin dosages, exercise, emotional well-being and their daily activities. Every other day the children remotely connected to the robot, which they believed to be autonomous, and performed this activity together. The robot provided explanations where necessary; asked the child relevant questions and shared personal information to establish a social bond. We measured adherence by comparing the frequency of diary entries and the diary content between the two conditions. Engagement was measured by observing the children’s participation and attention during the interactions. We used a questionnaire and observations made during the experiment as indicators for the development of a bond between the robot and the child.

The results show that low diary adherence is a common problem among diabetic children. We showed that interacting with the robot while keeping a diary helped to improve adherence. Children had significantly more completed diary entries with the robot present than while the robot was absent. Overall, both the diary and the robot were rated quite positively. Children felt the robot supported them well in performing the task, and they thought the robot was friendly, trustworthy and human-like. The children openly shared information with the robot about their daily lives, and were very inquisitive about the robot’s life as well. This was also explicit in their diaries, where we found a significantly larger amount of information shared in the diaries on days with the robot. The children eventually came to see the robot as a friend.

We need to be careful in saying the results we found are solely the result of our intervention. But we do think that using a robot can greatly enhance the pleasure of the activity and therefore the motivation of the child. Keeping a diary with the support from a robot can help the children take the next step in taking responsibility for their own self-management.