Grossberg 2020a.

Study characteristics
Methods	Study design: randomised controlled trial Study grouping: parallel group Study duration: 12 weeks Rescue medication: allowed: antidepressants (if the dose was stable for 30 days prior to randomisation and did not change during the study) and benzodiazepines (during the first 4 weeks of the randomised phase only (limited to 4 days/week with a maximum dose of 2 mg/day of lorazepam or equivalent)
Participants	Number randomised: 433 (297 brexpiprazole, 136 placebo) Mean age: 73.8 years Sex (female): 55.2% Type of dementia: Alzheimer's disease Severity of dementia: mild‐severe Indication: Agitation (baseline CMAI 70.9) Setting: care facility or community‐dwelling, provided the patient was not living alone Country: Russia (29.1% of randomised patients), the USA (27.9%), Ukraine (14.8%), Serbia (12.2%), Croatia (8.5%), Spain (4.4%), and Germany (3.0%). Efficacy results of 1 study group (0.5mg/day) not reported
Interventions	Brexpiprazole: 0.5, 1 mg and 2 mg/day Placebo Doses were titrated over a period of 2−4 weeks (days 1−3, 0.25 mg/day; days 4−14, 0.5mg/day; days 15−28, 1 mg/day; day 29 onwards, assigned dose). Patients unable to tolerate their assigned dose (or matching placebo) were discontinued from the trial.
Outcomes	Agitation: Cohen‐Mansfield Agitation Inventory (CMAI) Extrapyramidal symptoms Death Any adverse event Any serious adverse event Discontinuation (any reason) Discontinuation due to adverse events Cognitive function: Mini‐Mental State Examination (MMSE)
Identification
Notes	Sponsorship Source: Otsuka Pharmaceutical Development & Commercialization Inc (Princeton, NJ, USA) and H. Lundbeck A/S (Valby, Denmark).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Treatments were assigned by an interactive voice/web response system based on a fixed‐block, computer‐generated randomization code provided by the study sponsor and stratified by study center."
Allocation concealment (selection bias)	Unclear risk	No information provided
Comparability of groups (selection bias)	High risk	There were various differences between the study groups, including a higher score on the CMAI and on the MMSE in the placebo group than in (all three) drug groups. These differences were not (all) taken into account in the analysis
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"Treatment assignments were blinded to patients, investigators, and sponsor personnel..." "Brexpiprazole and matching placebo tablets were provided by the sponsor, packaged in numbered, weekly blister cards."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"Treatment assignments were blinded to patients, investigators, and sponsor personnel..."
Incomplete outcome data (attrition bias) All outcomes	High risk	No efficacy data for one study group (0.5mg/day). Overall and differential drop‐out low. No ITT analysis of efficacy.
Selective reporting (reporting bias)	High risk	Many outcomes reported on clinical trials.gov were not reported
Other bias	High risk	Run‐in of 6 weeks to wash out of antipsychotics, and other types of medication