Aboumatar 2019.

Study characteristics
Methods	RCT; follow‐up: 6 months; control group: usual care
Participants	Eligible: 417 Randomised: 240, I: 120, C: 120 Completed: 187, I: 93, C: 94 Mean age (SD): I: 63.9 years; C: 66.0 years Sex (% male): I: 40, C: 37 Inclusion criteria: hospitalised patients and their family‐caregivers who were admitted with a COPD‐related condition. Additional eligibility criteria included age 40 years or older; history of smoking more than 10 pack‐years; understands English language; has no terminal illness (< 6 months' life expectancy) unless end‐stage COPD; no severe cognitive dysfunction (able to follow simple instructions) Major exclusions: severe cognitive dysfunction; terminal illness (< 6 months' life expectancy) that is non‐COPD‐related; homelessness
Interventions	Hospital‐initiated programme that combines transition and long‐term self‐management support to patients and their family caregivers (the BREATHE programme). The BREATHE transitional care programme, which was co‐developed with COPD patients, family‐caregivers, and stakeholders Intervention components ‐ Tailored hospital‐to home transition support ‐ Individualised COPD self‐management education and support ‐ Facilitated access to community programmes and healthcare services The intervention is delivered by a new team member called "COPD Nurse Transition Guide". The new team member works with both hospital and outpatient care teams, is a registered nurse with home care service experience, and has received additional training in COPD self‐management and motivational interviewing. The nurse meets participants in the hospital and then follows up with them via home visits and phone calls. The intervention involves both patients and family caregivers (if available), is literacy adapted, and follows a tailored approach based on patient needs, priorities, and preferences Invervention duration: 3 months Disciplines involved: COPD nurse, treating physician Dominant component: none
Outcomes	Combined number of COPD‐related hospitalisations and ED visits per participant at 6 months (primary outcome); quality of life (SGRQ); combined number of ‘all‐cause’ hospitalisations and ED visits and individual components (hospital and ED visits separately); time to first event (re‐hospitalisation, first ED visit death); dyspnoea (mMRC); anxiety and depression; patient activation score; self‐efficacy and self‐care behaviours; patient perceptions of family‐caregivers' support; family‐caregivers' preparedness for caregiving and coping
Notes	This article was retracted and re‐published due to programming error and other errors that affected the results of our article. This article was re‐published with complete corrected findings. These findings have been used in this review Dominant component: self‐management (investigator's judgement)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "participants were randomised in a 1:1 ratio to either study group based on a pre‐generated sequence of assignments. Randomization was stratified by hospital unit, and a computer algorithm was used to perform blocked randomisation assignment..."
Allocation concealment (selection bias)	Unclear risk	Quote: "participants were randomised in a 1:1 ratio to either study group based on a pre‐generated sequence of assignments" Comment: unclear whether people screening for eligibility were aware of pre‐generated sequence of study group assignment
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: "...due to the nature of the intervention, participants and clinicians were not blinded" Comment: plausible that people being aware of group allocation could have biased results on SGRQ outcome, being more subjective. As noted by investigators, "increased communications with clinicians about exacerbation signs might have led to increased referrals to the emergency department (and subsequent hospitalizations)"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Comment: "data collectors and outcomes assessors were blinded to group allocation"
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: loss to follow‐up balanced across arms (27 in intervention arm and 25 in usual care arm). Reasons for loss to follow‐up comparable between groups
Selective reporting (reporting bias)	Low risk	Comment: all reports specified in protocol reported on. Additional post‐hoc analysis well supported by arguments.