Preventing isolation of people with dementia: Effectiveness and challenges of communication robots

Abstract

Participants: R Yamazaki-Skov (Denmark), X Liu (China), J Yang (Japan) R Makino (Japan), M Furukawa (Japan). ISSUE Social isolation and loneliness among older adults is a growing public health and public policy concern, made even more apparent by the COVID-19 pandemic. Social isolation and loneliness shorten older adults’ lives, and damage their mental and physical health and quality of life. (WHO, 2021). A review conducted in 2015 indicated that social isolation and loneliness were associated with a 29% and 26% increased likelihood of mortality. (Holt-Lunstad et. al). Especially, people with dementia (PWD) intend to avoid talking to people around them due to their declining cognitive function. Preventing social isolation and loneliness in people with dementia can not only increase their own wellbeing but also reduce social care costs. A key concept in dementia care is to help PWD maintain communication and keep connections with their families, caregivers, and environment. However, with the increasing number of older adults living alone and the shortage of caregivers, this has become difficult. Therefore, several communication robots have been developed to cope with this problem, and recently it has been reported that they were effective by using the reminiscence methods. Reminiscence methods are known to be effective in maintaining the communication skills of PWD and preventing the occurrence of behavioral and phycological symptoms of dementia (BPSD), such as wandering, shouting, and problematic behavior. This symposium aims to discuss how to develop easily accessible communication robots that can prevent isolation and loneliness for people with disabilities. The goal of this symposium is to clarify what problems existed on this process and what was attempted to solve them. CONTENT 1. Yamazaki-Skov (Denmark) will present an overview of research on intervention for PWD using communication robots. With/without the reminiscence method, the WoZ model was mostly applied for communication robot, and artificial intelligence (AI) robots showed potential for conversations with PWD. 2. Liu (Chaina) will present the effectiveness of the reminiscence method with using modified Pepper, a commercially available robot, into a WoZ model. The WoZ model showed almost the same effectiveness as the reminiscence method performed by a human coordinator.3. Yang (Japan) will report the results of an interactive experiment by implementing an AI model into RoBoHoN, a commercially available communication tool. Specifically, the changes they have made to RoBoHoN so far are as follows. ①Changed the voice recognition installed in RoBoHoN to Google STT ②Implemented GPT-4 model (AI model) in RoBoHoN ③On top of ②, we will present about the development of an AI model using RoBoHoN, which generates repetitive utterances automatically. 4. Makino (Japan) will present a project to analyze social workers' conversational behavior with the aim of finding a distinctive conversational model that can contribute AI to maintain conversation with PWD. The findings of the interaction analysis will be useful in considering what communication role should be assigned in the robot's conversation with PWD. 5. Furukawa (Japan) will analyze a case in which a communication robot was used by PWD who lives alone, and will report on the problems found in its use. CONCLUSION The presentations will show the effectiveness of communication robots preventing isolation and loneliness of PWD. Furthermore, in the development of AI communication robots, we will propose a multidisciplinary team approach that includes analysis of chat-based conversation corpora, behavioral analysis, and social system analysis.