TABLE 1.

Characteristics of included studies

Study (country)	Study design	Participants	Intervention	Comparison	Outcomes
Ringbaeket al. [21] (Denmark)	Quasi-randomised trial (allocation by alternation)	Intervention: 46 participants, age 68±9 years, 55% male, BMI 24±5 kg·m−2, FEV1 31±9% predicted, current smokers 14%, long-term oxygen therapy 7%, CAT score 20±7, MRC dyspnoea score 4 (2–5). Control: 69 participants, age 69±11 years, 39% male, BMI 25±6 kg·m−2, FEV1 35±11% predicted, current smokers 17%, long-term oxygen therapy 13%, CAT score 19±6, MRC dyspnoea score 4 (2–5).	Use of a wireless tablet computer (Nexus 7) with mobile-based internet connectivity plus usual care. The tablet software consisted of 1) a training diary (including the type, frequency, duration of exercises, and registered dyspnoea after each exercise using the Borg dyspnoea score), 2) video recordings of all the training exercises used and 3) training results for patients to monitor their own training. Staff had access to this information, and used it discuss progress, individual barriers and opportunities as well as to encourage further training.	Usual care: pulmonary rehabilitation twice weekly for 7 weeks at one of two hospitals (Gentofte Hospital and Hvidovre Hospital) or 10 weeks at Nordsjælland Hospital with each session lasting 2 h. Each session consisted of 1 h of supervised exercise and 1 h of education. The supervised training sessions consisted of walking and cycling. Patients were instructed to exercise at a level equal to 85% of predicted peak oxygen uptake as calculated from the incremental shuttle walk test.	Dropout rates. 1) Completion of rehabilitation: intervention 42 out of 46 participants; control 57 out of 69 participants. 2) Sessions attended: intervention 631 out of 784 sessions; control 509 out of 724 sessions.

Data are presented as mean±sd or mean (range), unless otherwise stated. BMI: body mass index; FEV₁: forced expiratory volume in 1 s; CAT: COPD Assessment Test; MRC: Medical Research Council.