Kessler 2018.

Study characteristics
Methods	RCT, multi‐centre (n = 33), France (12), Germany (8), Italy (6), and Spain (7); follow‐up: 12 or 24 months; control group: usual care
Participants	Randomised: 345, I: 172, C: 173 Completed: 265, I: 137, C: 128 Mean age: I: 67.3 years, C: 66.9 years Sex (% male): I: 69.4, C: 69.8 Inclusion criteria: COPD patients aged 35 years or older with post‐bronchodilator forced expiratory volume in 1 second (FEV₁)/forced vital capacity (FVC) ratio ≤ 70%; FEV₁ < 50% of predicted value; 10 pack‐year smoking history or more; ≥ 1 severe exacerbation in the previous year Major exclusions: not expected to survive longer than 6 months; cognitive/psychiatric disease; continuous treatment > 10 mg per day prednisone or equivalent longer than 6 weeks; living in a nursing home; unable to read or speak the country language
Interventions	Multi‐component home‐based COPD disease management intervention, specifically developed for patients with Gold III/IV COPD Intervention components ‐ Patient education (based on “Living Well With COPD”) and motivation by case managers, with the goal of attaining sustainable self‐management skills and behavioural changes ‐ Action plan to prevent exacerbations, with decision‐making and actions to be taken in case symptoms worsen ‐ Self‐monitoring of FEV₁, arterial oxygen saturation measured by pulse oximetry, and heart rate (HR). For patients on long‐term oxygen therapy, daily oxygen use and respiration rate (RR) were recorded by the NOWOX in‐line monitoring device ‐ Care co‐ordination through an e‐health platform for early detection of exacerbations by registration of status of well‐being, worsening, or alarm ‐ Reference to the investigator for same‐day medical assessment and follow‐up when confirmed alarm status ‐ During follow‐up consultation with physician, every 3 months Duration intervention: 12 or 24 months Disciplines involved: case manager, physician
Outcomes	Primary outcome: total number of all‐cause hospital days over 1 year Secondary outcome: COPD‐related hospital days, number of moderate to severe exacerbations, healthcare utilisation, death, HADS score, SGRQ, HRQoL, spirometry, ECG, 6MWD, BODE Index, fatal SAE cost‐effectiveness
Notes	Dominant component: structured follow‐up
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: “patients were allocated to groups in a 1:1 fashion according to a pre‐specified randomisation list generated before the study by a partial‐minimisation computer algorithm under supervision of the study sponsor”
Allocation concealment (selection bias)	Low risk	Quote: “patients were assigned a randomisation number by study staff at each centre in sequential numerical order through a telephone‐based interactive voice response system"
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: “for practical reasons, the study was open; neither the patients nor the investigators were blinded to the COPD management strategy”
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: “hospitalisations were rigorously and blindly reviewed by the end‐point validation committee (EVC) and followed‐up with additional enquiries if necessary, ensuring the reliability of the outcomes”; "EVC members were 3 respiratory physicians independent from the sponsor and investigational sites” Comment: primary outcome assessed blinded
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: missing outcome data were greater in control group (34/162) compared to intervention group (20/157). In control group patient death 23 compared to 3 in intervention group. Differences in missing outcome data potentially related to (absence) intervention. In addition, 23 patients in intervention group lost to follow‐up due to major protocol violations. Likely that outcomes are biased by loss to follow‐up, related to intervention
Selective reporting (reporting bias)	Low risk	Comment: all outcomes pre‐defined in protocol paper reported. However in addition, reported outcomes on smoking habits, daily use of LTOT, days to first exacerbation, number of patients who improved on 6MWD
Other bias	Unclear risk	Comment: main conclusions based on PP analysis instead of ITT. Analysis performed with ITT and PP populations. Potentially high risk of attrition bias with outcomes and reasons for missing data related to intervention. Multiple supportive outcomes that were not pre‐defined in protocol or trial register. ITT based on population at start of follow‐up period, after run‐in (5 weeks with intervention), instead of population at randomisation