Travier 2015.

Methods	RCT, 2 groups Study start and stop dates: conducted between 2010 and 2013 Length of intervention: 18 weeks Length of follow‐up: 18 weeks
Participants	Patients with breast or colon cancer undergoing cancer treatment (204 breast cancer patients of 237 patients in total), intervention started 6 weeks postdiagnosis. Unpublished data (final values) for breast cancer patients used in the analyses of this review
Interventions	Intervention (n = 102): supervised group exercise (aerobic and resistance) program based on Bandura’s social cognitive theory Control (n = 102): asked to maintain their habitual physical activity pattern
Outcomes	Primary outcome: fatigue: MFI and FQL Secondary outcomes: HRQoL: SF‐36 cancer‐specific quality of life: EORTC QLQ‐C30 (Version 3) anxiety and depression: HADS (Dutch version) physical fitness: VO2 peak; peak power output strength: thigh muscle strength, handgrip strength body composition: BMI, body fat distribution physical activity level: Short QUestionnaire to ASsess Health enhancing physical activity (SQUASH) self efficacy about the performance of physical activity (design paper) perceived impact of the disease on participation and autonomy assessed: Impact on Participation and Autonomy (IPA) questionnaire (design paper)
Notes	Trial registration: Current Controlled Trials (ISRCTN43801571), Dutch Trial Register (NTR2138)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Concealed computer‐generated randomisation, 1:1 ratio, stratified per age, adjuvant treatment, use of tissue expander, and hospital by sequential balancing
Allocation concealment (selection bias)	Low risk	"After the patient signed informed consent, the researcher (who was with the patient) called the data management department and provided the participants study number and the information necessary for stratification. The data manager then performed the randomization using a computer program and informed the researcher about the allocation (and also noted the allocation in the randomization log)."
Blinding of participants All outcomes	High risk	"not possible"
Blinding of personnel/care providers All outcomes	High risk	Not reported Comment: probably not done
Blinding of outcome assessment (detection bias) Fitness outcomes	Low risk	Outcome measures were assessed by researchers not involved with the participants
Blinding of outcome assessment (detection bias) All outcomes except fitness outcomes	High risk	Self reported items
Incomplete outcome data (attrition bias) All outcomes	Low risk	Drop‐out after 18 weeks: intervention group: 13/102 (12.7%) control group: 9/102 (8.8%) Follow‐up: high risk Drop‐out after 36 weeks: intervention group: 15/102 (14.7%) control group: 25/102 (24.5%)
Selective reporting (reporting bias)	High risk	Design paper published. EQ‐5D, IPA (Impact on Participation and Autonomy), self efficacy (items based on social cognitive theory) mentioned in the design paper, but not in the publication
Group similarity at baseline	High risk	"were comparable on most characteristics" Comment: More women in the intervention group were highly educated, had triple‐negative breast cancer, and were postmenopausal. Total physical activity levels tended to be higher in the control group
Adherence	Low risk	Participation in 83% (IQR 69% to 91%) of the classes. Reported to be physically active in 11 (IQR 6 to 14) of 18 weeks
Contamination	High risk	High level of physical activity reported by 56% of the controls at 18 weeks