Abstract. Full text article available at: https://doi.org/10.1002/gps.5460
To conduct a systematic review and network meta‐analysis examining the effects of non‐pharmacological interventions on neuropsychiatric symptoms (NPS) in a community‐dwelling population with preclinical and mild dementia to identify the design characteristics of more effective interventions. A comprehensive search was conducted in 10 electronic databases. Two reviewers independently appraised the methodological quality of studies using the Risk of Bias 2.0 tool. A pairwise meta‐analysis was performed to estimate the standardized mean differences (SMD) with 95% confidence intervals. Network meta‐analysis was then used to estimate the relative effects and rankings of different interventions. Twenty‐one studies involving 1773 participants were included. Seven studies focused on preclinical dementia, and 14 studies focused on mild dementia. The majority of studies reported a single domain of NPS (depression, anxiety, apathy or agitation) rather than overall NPS. Data on depression were pooled. Pairwise analysis and network meta‐analysis indicated that multimodal interventions (SMD = −0.47, p = 0.01) were superior to psycho‐behavioral educative interventions (SMD = −0.04; p = 0.65), cognitive training (SMD = −0.27, p = 0.10), and art‐based interventions for improving depression. The more crucial design characteristics included those that emphasized skill transferal into daily life, psycho‐behavioral content to encourage a positive outlook and self‐identity, and disease‐specific educational content to improve symptom management. This review indicated that multimodal interventions with cognitive, psycho‐behavioral and educative components were the most effective approaches for improving depression in patients with preclinical and mild dementia. More comprehensive evaluations using standardized and robust measures of NPS in patients with preclinical dementia are warranted.