Background: Chronic pain is a common problem in people with dementia living in residential aged care facilities (RACFs); however, it is often under-recognised and untreated due to the cognitive impairment of residents, and lack of knowledge and communication among care providers. Untreated pain is reported to be associated with behavioural and psychological symptoms of dementia (BPSD), such as agitation, depression and anxiety. Moreover, pain is also related to poor sleep and reduced physical activity, leading to a lower quality of life in people with dementia. Social robots, such as the robotic seal PARO, are promising psychosocial interventions to improve mood and manage behaviours in people with dementia, and the results from feasibility studies also indicate the potential effect of PARO to reduce pain in people with dementia. However, no randomised controlled trials (RCT) have been conducted to explore the effect of PARO on chronic pain in people with dementia living in RACFs. Objectives: This PhD study had four main aims: (1) to obtain the feasibility of a pilot RCT of a PARO intervention on observational pain levels in people with dementia living in RACFs; (2) to explore the effect of the PARO intervention on pain, agitation, depression, anxiety and quantified dosage of medication use in people with dementia living in RACFs; (3) to explore the effect of the PARO intervention on physiological responses (i.e., sleep and motor activity, salivary cortisol) in people with dementia; and (4) to explore experiences and perceptions of people with dementia towards the PARO intervention. Methods: A pilot RCT followed by semi-structured interviews was conducted with residents aged ≥ 65 years with dementia and chronic pain living in RACFs in Southeast Queensland, Australia. Participants were randomised into either a PARO intervention group (individual, non-facilitated, 30-minute sessions, 5 days per week for 6 weeks) or a usual care group (music, church, singing, story listening, etc.) using a computer-generated random number list. Feasibility of this pilot study included recruitment, eligibility, attrition, protocol adherence, data collection and safety issues. The primary outcome was the effect size of the change in the mean difference of researcher-observed pain level measured by the Pain Assessment in Advanced Dementia Scale (PAINAD) between the two groups. Secondary outcomes included staff-rated pain, agitation, depression, anxiety, quantified dosage of regular and pro re nata (PRN) medications, sleep and motor activity (i.e., step counts, physical activity and energy expenditure) measured by actigraphy, as well as stress level measured by salivary cortisol. Quantitative analyses followed the intention-to-treat approach. Differences in outcomes between two groups were examined using the generalised estimating equation model. Covariate-adjusted mean differences with 95% confidence intervals accounted for the potential confounding factors of age, gender, cognitive status and medications at baseline. Cohen’s d for effect size was calculated, and statistical significance was set at p < .05. Participants who were capable of verbal communication and comprehension were interviewed about their experiences and perceptions of the PARO intervention. Qualitative data were analysed using inductive thematic analysis. Results: Feasibility of the pilot RCT was established. Forty-three participants with dementia and chronic pain living in three RACFs participated in the study. The recruitment rate for the eligible residents was 60.6% (43/71) after a strict screening process. The attrition rate of the participants was low, with three out of 43 participants (7.0%) dropping out during the intervention process. The average attended sessions among participants was 23.3 (±7.3) and the average duration of each session was 16.36 (±8.13) minutes. A total of 2,470 out of 2,520 (98.02%) observational pain assessments were completed. No adverse events were reported during the study. Compared to participants in the usual care group, participants in the PARO group had a significantly lowered level of observed pain (-0.514, 95% confidence interval [CI] -0.774 to -0.254, p < .001, Cohen’s d = -0.765) after receiving the PARO intervention and used fewer prescribed PRN medications (-1.175, 95% CI -2.205 to -0.145, p = .025, Cohen’s d = 0.690) during the 6-week PARO intervention, which were adjusted for age, sex, Mini-Mental Status Examination (MMSE) and medication at baseline. No significant results have been found on staff-rated pain, agitation, depression, anxiety and regularly scheduled medications after the 6-week PARO intervention. At the end of the 6-week PARO intervention, compared to the usual care group, participants in the PARO group showed a greater increase in the duration (hours) of daytime wakefulness (1.91, 95% CI: 0.09 to 3.73, p = .042, Cohen’s d = 0.655) and a greater reduction in the duration (hours) of daytime sleep (-1.35, 95% CI: -2.65 to -0.05, p = .040, Cohen’s d = 0.664). After one session of PARO intervention, the increase in the duration (hours) of night sleep was significantly higher in the PARO group (1.81, 95% CI: 0.22 to 3.84, p = .030, Cohen’s d = 0.570). No significant results were found for motor activity. Several difficulties in the collection and analysis of salivary cortisol were encountered, including cognitive impairment of participants and inadequate saliva volume for assay. Four themes emerged from qualitative interviews of 11 participants with mild to moderate dementia: (1) perceptions of PARO, (2) therapeutic effects of PARO, (3) limitations of PARO, and (4) program improvement. Residents with dementia expressed positive attitudes towards the use of PARO and acknowledged the therapeutic benefits of PARO on mood improvement and relaxation for pain relief, but also mentioned the limitations of its weight, voice and characteristics. Residents’ responses to the PARO intervention could fluctuate during the intervention process. Conclusions: Results from this study indicate that the PARO intervention may be incorporated into daily practice as a psychosocial intervention to manage pain in people with dementia. It also provides subsequent benefits on sleep and promotes positive mood for participants. However, the effect of the PARO intervention on BPSD (i.e., agitation, depression and anxiety), motor activity and salivary cortisol level needs further exploration. In addition, considering the challenges of obtaining valid saliva samples, salivary cortisol may not be a feasible biomarker of physiological stress in people with dementia and chronic pain. Larger randomised controlled trials with longer time frames are needed to evaluate the use of PARO for people with dementia living in long-term care settings. Moreover, individual preferences need to be considered before the use of PARO intervention. This thesis consists of nine chapters. Notably, Chapter 2, the literature review, presents two published systematic reviews, and Chapters 5 - 8 present four published/submitted research articles reporting the effects of PARO from the pilot RCT and findings from interviews. This dissertation does not include a traditional discussion chapter as each research article details the major findings of the study and a discussion of the results. The final chapter gives a summary of the main findings, methodological limitations as well as implications for future research, nursing practice and healthcare education.