The role of humanoid social robot in nursing home for cognitive impaired seniors : what are the effects and benefits?

Abstract

The world’s population is ageing quickly, and the number of seniors living with dementia is rising. Seniors’ psychosocial well-being is a top priority for their family members and caregivers in nursing homes. Assistive technology is a promising panacea, potentially relieving the distress of caregivers facing many challenges. This study sought to examine the role, effects, and benefits of providing humanoid social robots in nursing homes for cognitively impaired seniors. A mixed-methods randomised controlled trial (RCT) was conducted over 32 weeks in 7 nursing homes run by Sheng Kung Hui, which is a non-governmental organisation in Hong Kong. One hundred and three seniors with dementia (mean age of 87.2) were randomised into intervention and control groups. Seniors in the control group received usual care, and seniors in the intervention group were provided with a Kabochan companion robot as their individual 24-hour companion over the course of two 8-week periods in each B of the ABAB design. Local therapists and frontline workers were trained to support the deployment of the social robot intervention. Data collection was conducted at baseline (T0), at the end of A1 (T1), at the end of the Kabochan intervention at the end of B1 (T2), at the end of the Kabochan intervention withdrawal at the end of A2 (T3), and at the end of the Kabochan intervention at B2 (T4). Seniors were assessed for their cognitive status and activities of daily living (ADL) and using the Senior Technology Acceptance Model questionnaire (STAM), Quality of Life in Alzheimer’s Disease (QoL-AD), Neuropsychiatric Index (NPI), Geriatric Depression Scale (GDS), Modified Barthel Index (MBI), and Montreal Cognitive Assessment (MoCA). The interaction patterns between Kabochan and seniors were observed weekly during B1 and B2. Additional interaction information was gathered from motion sensors that were implemented in the Kabochan robots for 28 seniors during B2. Pre- and post-intervention focus groups were conducted with stakeholders and seniors for more insight into the Kabochan intervention. The study had a low dropout rate of 8.7%. NPI-Q caregiver distress and NPI-Q severity significantly improved for the intervention group, compared with that of the control group, from T0 to T4. Similarly, there was a significant difference between the two groups for the MoCA from T0 to T4 and for the QoL-AD and the GDS-15 from T0 to T2. No significance was detected for the MBI, which assesses the ADL of seniors. From the STAM scores, we found that perceived usefulness and attitude towards technology were significant at T0 and T4. From the observation data on interaction patterns, we found that seniors with dementia liked cuddling Kabochan the most, which was followed by general interaction with the robot. The seniors with a higher interaction level with Kabochan had significant positive outcomes in terms of NPI-Q caregiver distress, gerontechnology anxiety, attitude towards gerontechnology (AT), and perceived usefulness. The findings from the motion sensor data showed that the seniors were most active with Kabochan at 11:00 and 16:00. From the focus group studies, we learnt about the best practices and benefits of engaging with a humanoid robot, laying a strong foundation for seniors with dementia to develop a deeper relationship with their 24-hour companions. In a person-centred approach, it is more suitable for seniors with dementia to accept social robots individually rather than as a group. The results set a path for future research on the introduction of humanoid social robots as companions with further localisation of language and artificial intelligence to induce enhanced interactions with seniors with dementia. We need to continue gathering high-fidelity research-based evidence to build new knowledge that can be effectively replicated to accelerate outreach and benefit more seniors with dementia with cost-effective support.