Background: Behavioural and psychological symptoms of dementia (BPSD) are complex neuropsychiatric symptoms that contribute to caregiver strain, increased rates of institutionalisation and reduced quality of life. Virtual reality (VR) has gained interest as a non-pharmacological approach to potentially reduce BPSD severity.

Objective: This review sought to synthesise evidence on the effectiveness of VR in reducing BPSD severity, while exploring its acceptability, safety, and optimal dosage in dementia care.

Methods: MEDLINE, EMBASE, CINAHL and SCOPUS were searched for randomised and quasi-experimental trials assessing VR's effect on BPSD. JBI critical appraisal checklists were used to assess methodological quality. Findings were presented narratively, with meta-analysis performed on a subset of BPSD symptoms where data were available.

Results: Of the ten included studies, four found no significant change in overall BPSD. Mixed findings were observed for individual BPSD symptoms. Meta-analysis showed a significant reduction in depressive symptoms (mean diff -0.38, P= .026) and no reduction in agitation (mean diff 1.87, P = .2). Two studies reported reduced aggression and mixed findings were found for anxiety. Reduced apathy was observed in one study following each VR session and during the session in another. VR was generally well-accepted with few side effects reported.

Conclusion: VR appears to be an acceptable non-pharmacological intervention for BPSD reduction. However, the limited available studies, methodological variations and quality issues suggest the need for future larger-scale research to confirm its efficacy and effectiveness.