Study designs of evaluations included in the review
Randomised controlled trials (RCT) and single-case designs (SCDs; including those comparing treatments within an individual) were eligible for inclusion.
Eligible RCTs had to fulfil the following criteria from the American Psychological Association's Task Force: clearly stated inclusion and exclusion criteria; reasonable length of follow-up; the determination of clinical as well as statistical significance; use of an agreed treatment manual or standard; evaluation of treatment fidelity; use of a between-group comparison using intention-to-treat analysis; at least 25 participants per arm; clear reporting of drop-outs and treatment refusals.
SCDs had to fulfil the criteria: ABAB design (where A is no intervention and B is an intervention) or multiple baseline designs; continuous assessment with a stable baseline assessment of behaviour (at least 3 data points or predictable variability); standardised NPS assessments; inter-observer agreement data collected; use of appropriate analyses or suitable graphical presentation of results.
Specific interventions included in the review
Studies of non-pharmacological interventions compared with placebo, no treatment, attention control or standard care were eligible for inclusion; studies with additional pharmacological treatment were excluded. The interventions included were an unmet needs intervention (hearing aids), learning and behavioural interventions (functional analysis, cloth barrier, differential reinforcement, and a multi-component intervention combining elements such as functional analysis and contingent reinforcement), caregiving interventions (family visit education programme, caregivers trained in behaviour management, and a combination of caregiver training and an exercise programme for patients), and an environmental/reduced stress-threshold approach (bright light therapy). Control groups, where applicable, received usual care.
Participants included in the review
Studies of patients with dementia were eligible for inclusion. The included patients had various stages of dementia from probable to severe; the majority had Alzheimer's disease or related disorders. The patients were at home, in community dwellings, in day care, in nursing or residential homes, or in a skilled nursing facility.
Outcomes assessed in the review
Studies reporting a reduction in NPS using objective measures with evidence of reliability and validity were eligible for inclusion. The review focused on disruptive behaviours such as agitation, aggression or wandering, and excluded studies that only assessed depression. The outcomes measured in the studies were observations and ratings: hearing aid use; wandering frequency; agitated speech; behaviour counts and behavioural observation by staff; and measures of agitation, behaviour and depression using various disease specific scales such as the Behaviour in Alzheimer Disease scale.
How were decisions on the relevance of primary studies made?
At least two independent reviewers performed the literature searches, with any disagreements resolved by consensus or through consultation with another reviewer.