Fitzsimmons 2002.

Study characteristics
Methods	Randomised controlled trial (cross‐over design, no wash‐out period) Duration of follow‐up: 2 weeks
Participants	Country: USA (Charlotte County, Florida) Participants were recruited through the Southwest Florida Alzheimer’s Association. Inclusion criteria: age ≥ 65 years, living at home, a medical diagnosis of dementia in the medical record, MMSE score ≤ 24, stable medication Number of participants completing the study: n = 29 Age (mean ± SD) years: people with dementia: 81.3 (range 72 to 90); caregivers: 74.61 (range 52.5 to 89.1) Gender, female: people with dementia: 65.5%; caregivers: 55.2% Cognitive status, MMSE (mean ± SD): 12.93 (range 0 to 23) Care dependency (ambulation status) self n = 14 (48.3%), with device n = 11 (37.9%), with assist n = 3 (10.3%), non‐ambulatory n = 1 (3.4%) Relationship to the participant (spouse): 72.4%
Interventions	Intervention: therapeutic individualised recreational intervention, activities were delivered by a specially trained recreation therapist directly to the study participants (1 to 2 hours for 3 to 4 times per week, for 2 weeks). Control: no intervention (usual care)
Outcomes	Agitation (Cohen‐Mansfield Agitation Inventory), passivity (Passivity in Dementia Scale), blood volume pulse, heart rate (biograph device); no primary outcome defined.
Funding	Grant from the Retirement Research Foundation
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	No information was reported. Comment: answer from study authors: "Use of random number tables"
Allocation concealment (selection bias)	Unclear risk	No information was reported. Comment: answer from study authors: allocation was concealed, but no method was mentioned. No information reported to permit judgement of ‘low risk’ or ‘high risk’.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not blinded to group allocation (not possible), insufficient information to permit judgement of ‘low risk’ or ‘high risk'
Blinding of outcome assessment (detection bias) Subjective outcomes (proxy‐rated)	High risk	Subjective outcomes and assessors (family caregiver) were not blinded to group allocation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	"Thirty subjects were recruited through the Southwest Florida Alzheimer’s Association, and 29 completed the study (one subject was too paranoid to allow the research team into the home)."
Selective reporting (reporting bias)	Unclear risk	Trial was not registered, and no study protocol is available.
Other bias	High risk	No wash‐out period, but no information as to whether this could have led to bias. No paired data were available (risk of unit of analysis bias); we used the (unpaired) data of the complete study period.