Objective: To assess the efficacy, comparative effectiveness, and adverse effects of nonpharmacologic interventions for agitation and aggression in individuals with dementia.
Data sources: Ovid MEDLINE®, Ovid Embase®, and the Cochrane Central Register of Controlled Trials bibliographic databases; hand searches of references of relevant studies.
Review methods: Two investigators screened abstracts and full-text articles of identified references for eligibility. Eligible studies included randomized controlled trials evaluating nonpharmacologic interventions to manage agitation/aggression in individuals with dementia in nursing home, assisted living, or community settings. We analyzed outcomes of agitation/aggression, general behavior, patient quality of life, admission to long-term care, and staff and caregiver outcomes related to patient behavior and care burden. We assessed risk of bias, extracted data, and evaluated strength of evidence for each comparison and outcome. We analyzed pooled estimates to assess efficacy and comparative effectiveness. We conducted a qualitative analysis when data could not be pooled.
Results: We identified 126 unique randomized controlled trials as of July 2015. Patient-level interventions involving music, aromatherapy with lavender, and bright light were similar to usual treatment or attention control at managing agitation/aggression in people with dementia (low-strength evidence); interventions tailored to recipients' skills, interests, or both were similar to usual care in managing agitation/aggression in people with dementia (low-strength evidence). Care delivery–level interventions (dementia care mapping and person-centered care) were similar to usual care in managing agitation/aggression in people with dementia (low-strength evidence). Evidence was insufficient to draw conclusions on the effectiveness of most caregiver-level interventions in managing agitation/aggression in people with dementia; caregiver interventions targeting caregiver skills and behavior were similar to attention control in managing agitation/aggression (low-strength evidence). However, these interventions show benefits in caregiver confidence in caregiving and caregiver distress. Adverse effects were rarely reported.
Conclusions: Although many trials have been conducted to determine effective nonpharmacologic interventions for agitation/aggression in dementia, which is a critical topic, the evidence base is weak because of the variety of comparisons, measurement issues, and other methodological limitations. When evidence was sufficient to draw conclusions about effectiveness for a group of interventions, agitation/aggression outcomes were typically similar to those of control groups. Future research is needed to guide providers and informal caregivers toward effective interventions for agitation/aggression in dementia.