Schultz 2021.

Study characteristics
Methods	Study design: RCT Total duration of study: 5 months (plus additional 12 months follow‐up) Details of any run‐in period: Participants assessed at randomisation (T0) and 4 weeks later before commencing rehabilitation (T1). Number of study centres and location: Single site. Germany Study setting: Hospital Withdrawals: 12 participants from each group withdrew from the study. Date of study: June 2015 to August 2017
Participants	Number recruited: Randomised to intervention group: 214. Randomised to control group: 222 Number completed: Completed intervention group: 202. Completed control group: 210 Mean (SD) age: Intervention group 50.7 (8.8) years. Control group 51.6 (8.7) years. Age range: Not stated Gender (M/F): Intervention group 120/82. Control group 117/93 Mean (SD) BMI: Not stated Severity of condition: Not stated. Uncontrolled asthma ACT < 19 was an inclusion criterion. Diagnostic criteria: Physician diagnosis of asthma (ICD‐10; J45). Every asthma diagnosis was verified by a pulmonologist at admission to the rehabilitation unit. Baseline lung function: Mean (SD) % FEV1 postbronchodilator: Intervention group 87.5 (20.7)%. Control group 87.3 (20.5)% Smoking history: Current smokers: Intervention group 34 (16.8%). Control group 34 (16.2%) Asthma treatment: Not stated Inclusion criteria: Approved for pulmonary rehabilitation. Physician diagnosis of asthma (ICD‐10; J45). Uncontrolled asthma based on ACT < 20 Exclusion criteria: Cognitive impairment. Inadequate German language ability. Severe concomitant disease that might mask results of rehabilitation (e.g. cancer, orthopaedic, cardiac, psychiatric comorbidities). If initial diagnosis of asthma could not be confirmed by pulmonologist at admission to rehabilitation
Interventions	Intervention: 3‐week inpatient PRP consisting of physical training (endurance training 5 units per week 45 to 60 min each time, strength training 3 sessions per week), education, respiratory physiotherapy and (if needed) psychosocial support, smoking cessation, nutritional counselling, inspiratory muscle training. Delivered by a multidisciplinary team Comparison: Waitlist control group Concomitant medications: Not stated Excluded medications: Not stated
Outcomes	Primary outcomes: Asthma control (ACT) Secondary outcomes: Quality of life (SGRQ, AQLQ). Subjective health (Global Rating of Change). Symptoms (severity of dyspnoea cough sputum, pain on numerical rating scale). Fatigue (BFI). Depression and anxiety (PHQ‐9, GAD‐7). Illness representation (Brief IPQ). Subjective self‐management (Skill and Technique Acquisition Scale). Medication adherence and medical beliefs (MARS‐D). Lung function. Health‐related resource use (FIMA‐Lu questionnaire). Exercise capacity (6MWD). Work ability (work ability index) Time points reported: At randomisation (T0). 4 weeks after randomisation (T1). 7 weeks after randomisation (intervention group finish PR) (T2). 20 weeks after randomisation (intervention group 3 months after completing PR; control group begin PR) (T3). Also follow‐up at 6, 9, 12 months after inpatient rehabilitation in the intervention group, and 3, 6, 9 months after rehabilitation in the control group Data reported as endpoint rather than as change from baseline. Adjusted mean differences also reported.
Notes	Funding: German Statutory Pension Insurance of South Bavaria. Funding covers costs for staff and materials and travelling expenses. Notable conflicts of interest: No competing interests declared. Other: Data at T2 and T3 extracted for review. 6MWD data were available pre‐ and post‐PRP for both intervention and control groups, with end of PRP occurring 3 months before the control group started PRP.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The randomization list (with computer‐generated random numbers) is created by the Department of Medical Psychology and Psychotherapy, Medical Sociology and Rehabilitation Sciences at the University of Würzburg (concealed allocation).”
Allocation concealment (selection bias)	Low risk	The randomization list (with computer‐generated random numbers) is created by the Department of Medical Psychology and Psychotherapy, Medical Sociology and Rehabilitation Sciences at the University of Würzburg (concealed allocation).”
Blinding of participants and personnel (performance bias) Subjectively reported outcomes	High risk	High for participants due to nature of intervention. Low for personnel blinding. "Patients themselves cannot be blinded due to the timepoint of the start of their inpatient rehabilitation. However, those who deliver the rehabilitation treatment are unaware whether the patient is a study participant, a participant of the IG or the CG, or a regular inpatient outside the study".
Blinding of participants and personnel (performance bias) Not subjectively reported outcomes	High risk	High for participants due to nature of intervention. Low for personnel blinding. "Patients themselves cannot be blinded due to the timepoint of the start of their inpatient rehabilitation. However, those who deliver the rehabilitation treatment are unaware whether the patient is a study participant, a participant of the IG or the CG, or a regular inpatient outside the study".
Blinding of outcome assessment (detection bias) Subjectively reported outcomes	High risk	"Blinding was not possible"
Blinding of outcome assessment (detection bias) Not subjectively assessed outcomes	High risk	"Blinding was not possible"
Incomplete outcome data (attrition bias) All outcomes	Low risk	436 subjects randomised. 412 subjects included in intention to treat analysis.
Selective reporting (reporting bias)	Low risk	All primary and secondary endpoints stated in protocol are reported upon.
Other bias	Low risk	No other suggestions of bias evident.