Travers 2017.

Study characteristics
Methods	Study design: randomised controlled trial (ACTRN12613000296730) Duration of follow‐up: 8 weeks Study period: not reported
Participants	Country: Australia Setting: four nursing homes, southern suburbs of Brisbane, Queensland, 48 to 126 beds. All catered for residents with high and low care needs and all were accredited with the National Accreditation Agency (Aged Care Standards and Accreditation Agency Ltd). The facilities were managed by religious (two), independent not‐for‐profit (one) and private (one) organisations and all facilities employed activities staff and provided a variety of organised activities for residents. Participants Inclusion criteria: living in the nursing home for at least 3 months, being able to communicate in English, mild to moderate dementia (sMMSE score of 10 or higher), symptoms of depression (GDS‐12R score of four or higher). Exclusion criteria: no diagnosis of dementia, participants receiving psychotherapy Number of participants randomised: n = 19; intervention group n = 10, control group n = 9 Number of participants lost to follow‐up: control group n = 1 (died) Number of participants completing the study: n = 18; intervention group n = 10, control group n = 8 Baseline characteristics Age (mean ± SD) years: intervention group 87.2 ± 7.7, control group 85.5 ± 10.9 Gender, female: intervention group 80%, control group 100% Cognitive status, sMMSE (mean ± SD): intervention group 19 ± 3.87, control group 16.13 ± 3.7 Care dependency: not reported Mobility status: intervention group: 10% independently ambulatory, 50% ambulatory with assistance, 40% non‐ambulatory; control group 37.5% independently ambulatory, 50% ambulatory with assistance, 12.5% non‐ambulatory
Interventions	Intervention: BE‐ACTIV Control: walking and talking intervention
Outcomes	Primary (defined in the trial registration): Depression (GDS‐12R) Quality of life (QoL‐AD nursing home version) Secondary: agitation (CMAI ‐ short form)
Funding	Funded by the JO & JR Wicking trust
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"The Project Coordinator assigned participants a unique study identification number and randomly allocated them to either the BE‐ACTIV or the Walking and Talking intervention using the SPSS randomization function."
Allocation concealment (selection bias)	High risk	"The Project Coordinator assigned participants a unique study identification number and randomly allocated them to either the BE‐ACTIV or the Walking and Talking intervention using the SPSS randomization function." Not concealed.
Blinding of participants and personnel (performance bias) All outcomes	High risk	"In the first instance, facility staff (activities staff in particular) and volunteers were invited to attend two 90‐min depression training sessions that were conducted at the commencement of the study in each facility (...) An overview of the project including its rationale and methods was also provided." The therapist and the nursing staff was aware of the group allocation; no information about the blinding of the participants was reported. Since the same nurses cared for participants in both study groups there is a risk of contamination. We judged risk of performance bias to be high.
Blinding of outcome assessment (detection bias) All outcomes	High risk	"The measures of QOL (QOL‐AD‐NH), and depression (GDS‐12R) were re‐administered following completion of the interventions by the Project Coordinator only, who was not blinded regarding participant’s group allocation."
Incomplete outcome data (attrition bias) All outcomes	Low risk	"Ten participants were allocated to the BE‐ACTIV intervention and nine to the Walking and Talking intervention. One resident who had been allocated to the Walking and Talking intervention, however, died prior to commencement of the intervention and his data were excluded from all analyses. All remaining participants completed the eight‐week interventions."
Selective reporting (reporting bias)	Unclear risk	Only the primary outcomes defined in the trial register were reported.
Other bias	Low risk	‐