Berglund 1994.

Study characteristics
Methods	Study design: RCT Number randomized: 199; 98 to the exercise group and 101 to the control group Study start and stop dates: not reported Length of intervention: 7 weeks Length of follow‐up: 1 year after end of the intervention
Participants	Type cancer: 80% breast cancer, 7‐8% ovarian cancer, remaining were other types Time since cancer diagnosis: not reported Time beyond active treatment: not reported  Inclusion criteria: < 75 years old curative treatment for a primary tumor within 2 months after postoperative treatment with radiation therapy or chemotherapy Eligibility criterion related to interest or ability to exercise, or both: none Exclusion criteria:  none Gender: not reported Current age: not reported Age at cancer diagnosis: not reported Ethnicity/race: not reported Education level: not reported SES: not reported Employment status: not reported Comorbidities: not reported Past exercise history: not reported On hormone therapy: not reported
Interventions	98 participants assigned to the exercise intervention, including: physical training, information, and coping skills training. The physical training component included exercises to increase mobility, muscle strength, general fitness, and relaxation in the form of progressive muscle relaxation or deep relaxation with positive images. Patients were given instructions for progressive relaxation at home Type exercise (aerobic/anaerobic): aerobic Intensity of experimental exercise intervention: not reported Frequency: twice per week during the first 4 weeks (once for physical training and once for information), then once per week for coping skills training Duration of individual sessions: not reported Duration of exercise program: 7 weeks Total number of exercise sessions: 11 Format: group Facility: facility Professionally led: professionally led by an oncology nurse 101 participants assigned to control group, including: information and coping skill training Adherence: the mean absenteeism among participants was 1 session, representing a variation of the number of participants per session between 3 and 7 (mean 4.9). Contamination of control group: not reported
Outcomes	No primary outcome identified.  HRQoL outcomes included: fatigue ‐ measured as part of a nonspecified larger scale body image ‐ measured as part of a nonspecified larger scale pain ‐ measured as part of a nonspecified larger scale global health ‐ measured as part of a nonspecified larger scale anxiety ‐ measured using the shortened HADS depression ‐ measured using the shortened HADS problems with QoL ‐ not specified how this was measured MAC scale Physical outcomes included: physical strength physical training tiredness body image pain global health Outcomes were measured at baseline; end of the intervention; and 3 months, 6 months, and 12 months: exercise group: n = 98 at baseline, n = 90 at end of the intervention, n = 90 at 3 months, n = 88 at 6 months, n = 87 at 12 months control group: n = 101 at baseline, n = 98 at end of the intervention, n = 93 at 3 months, n = 91 at 6 months, n = 89 at 12 months Subgroup analysis: not reported Adverse events: no cancer recurrences or adverse events reported
Notes	Country: Sweden Funding: Swedish Cancer Foundation
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Used Efron's method of randomization of small samples (Hjelm‐Karlsson 1991)
Allocation concealment (selection bias)	Unclear risk	Whether the treatment assignment was concealed from study personnel and participants was not described
Blinding of participants and personnel (performance bias) All outcomes	High risk	Owing to the nature of the intervention, it was not possible to blind the participants; however, it is unclear whether the outcome was influenced by a lack of masking
Blinding of outcome assessment (detection bias) All outcomes	High risk	Study personnel and outcome assessors were not masked or blinded to the study interventions
Incomplete outcome data (attrition bias) All outcomes	High risk	There was no ITT analysis, it is unclear how missing values were handled, there were large losses to follow‐up
Selective reporting (reporting bias)	Low risk	There appears to be no selective reporting of outcomes
Other bias	Low risk	The trial appears to be free of other problems that could put it at a high risk of bias