Nyrop 2017.

Methods	RCT, duration: 6‐week intervention, plus 6‐month follow‐up
Participants	Inclusion Stage 0‐III BC, age > 21 years (although 16%, n = 10, had stage IV disease at baseline) Compliant on AI ≥ 4 weeks ≥ 3/5 score on PROMIS pain intensity‐short form Exercising ≤ 150 min/week Exclusion Concurrent chemotherapy or radiation therapy Identified through review of appointment schedule at BC clinic in tertiary care hospital. Recruited between February 2014‐August 2015 n = 62 (31 in exercise, 31 in waiting list group) Median age: 63.8 ± 8.3 years Intervention group had more prior use of tamoxifen (11 vs 5) and vitamin D supplement (28 vs 19); more AI non‐compliance in control group (forgets once a week = 2 vs 9). Baseline joint symptoms balanced
Interventions	Intervention Home‐based, self‐directed walking, 6 weeks (Walk with Ease‐Breast Cancer programme) Given written information, workbook encouraging at least 150 min exercise/week, walking as comfortable pace Physical activity log No contact with investigators during the 6‐week intervention period Comparator Waiting list control‐ no intervention. Asked to await further contact from the research team at 6 weeks after study baseline After 6 weeks given same materials and instructions as intervention group
Outcomes	(Assessed at baseline, 6 weeks and 6 months after intervention. Waiting list control group had an extra questionnaire 6 weeks after completing intervention) Physical activity levels (min/week) VAS for pain, stiffness, fatigue WOMAC index FACT‐G RAI Adherence to AI ASE scale Outcome expectations from exercise Self‐efficacy for physical activity Feasibility, tolerability, adverse events
Notes	Study funding from National Cancer Institute
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Quote: "Randomized", however statements indicating serious problems ‐ "misrandomised", "inadvertent randomisation errors". Unclear what the errors were however. Imbalances in baseline demographics, which may be a reflection of selection bias.
Allocation concealment (selection bias)	High risk	No information provided regarding allocation concealment. 5/62 participants were "mis‐randomised". It is unclear whether this was a result of the randomisation process, or allocation process. To be identified as being "mis‐randomised", it indicates that the allocations were not concealed from investigators.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Unable to blind this intervention
Blinding of outcome assessment (detection bias) All outcomes	High risk	All outcomes were PROs, completed by unblinded participants.
Incomplete outcome data (attrition bias) All outcomes	High risk	ITT analysis prospectively planned. However 23% of intervention arm did not complete 6‐week data, vs 6% control arm. Only 77% completed 6‐month data, but no mention about dropout between arms. No description of reason for not completing 6‐week or 6‐month questionnaires
Selective reporting (reporting bias)	Low risk	All outcomes reported
Adherence	Low risk	Mean walking time in min/week at end of 6 weeks intervention: 108.7 for intervention arm (aim was 150 min/week)
Contamination	Low risk	Mean increase in activity time in min/week between baseline and 6 weeks was 76.21 in intervention arm, and 10.52 in the waiting list control arm. The study authors did not collect information on the type or intensity of exercise that these women engaged in, only the number of min/week.
Other bias	Low risk	No other sources of bias identified