This 2-group pilot randomized controlled trial (RCT) of a tele-exercise program was conducted in Hong Kong in 2024. About 27 participants were randomized; 24 completed the baseline assessment and were included in the modified intention-to-treat (mITT) analyses. The study was reported in accordance with the CONSORT (Consolidated Standards of Reporting Trials) extension for randomized pilot and feasibility trials (Checklist 1), and the evaluation was guided by the RE-AIM (reach, effectiveness, adoption, implementation, and maintenance) framework. Outcome assessors and data analysts were blinded to group allocation. A multidisciplinary task force (health care professionals, exercise science and psychology experts, clinical trial specialists, data analysts, coaches, and a local breast cancer survivor association) supported intervention development and delivery.

Ethical approval was obtained from the Research Ethics Committee of the Hospital Authority of Hong Kong (KC/KE-23‐0122/ER-4). The trial was registered on ClinicalTrials.gov (NCT06382441) and conducted in accordance with the Declaration of Helsinki and ICH-GCP (International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use – Good Clinical Practice). All participants provided written informed consent and were informed of their right to withdraw at any time.

Confidentiality was protected through restricted access to study files, secure storage of paper records in locked cabinets, and encryption of electronic data (including SSL/TLS [Secure Sockets Layer/Transport Layer Security] encryption for any web-based data transfer). Personally identifying information and the ID linkage file were stored separately from the analysis dataset, which was deidentified prior to analysis and reporting. Study documents are retained for 5 years after study completion in line with hospital authority/hospital policies and are securely destroyed thereafter. No identifiable participant information is presented in this paper or multimedia appendices.

Breast cancer survivors in the adjuvant-treatment period (within 6 weeks after completion of chemotherapy, with radiotherapy as per usual care) were recruited from the Breast Centre at Kwong Wah Hospital, Hong Kong, between November 2023 and April 2024. Inclusion criteria included (1) females aged 40‐64 years, (2) within 6 weeks after completion of chemotherapy, (3) without severe anemia, (4) without cancer metastasis, (5) able to read and communicate in Cantonese or Mandarin, and (6) smartphone users. Those with any medical, physical, or psychological conditions that may limit participation were excluded from the study, such as uncontrolled severe cardiovascular disease, schizophrenia, and severe neurological dysfunction. Eligibility was assessed by a case manager of breast cancer (a nurse consultant), consulting with a surgeon or a physiotherapist if needed.

Previous studies have suggested that 10% of the estimated sample size of the main study (n=130) is recommended for a simple-armed pilot (ie, intervention only). Given that a control group was included, the estimated sample size for the pilot was 24. Several effective strategies for participant recruitment and retention suggested by previous studies were adopted, including emphasizing study benefits to participants (eg, free assessments of the rehabilitation process on physiological and psychological functions) [12].

Participants were randomized in a 1:1 ratio to the intervention or control group using permuted block randomization with randomly varying block sizes (4, 6, or 8). The randomization schedule was computer-generated prior to recruitment by an independent staff member not involved in enrollment, intervention delivery, or outcome assessment.

Allocation concealment was implemented using sequentially numbered, opaque, sealed envelopes (SNOSE) prepared in advance and stored by independent research staff. After eligibility confirmation and written consent, the next envelope was opened to reveal the group assignment. Baseline assessments were scheduled immediately after randomization; however, 3 participants withdrew after allocation and before completing the baseline assessment. Outcome assessors and data analysts remained blinded to group allocation at both baseline and postintervention assessments. Intervention coaches were not blinded because they delivered the program, but they were not involved in outcome assessments or data analysis. Coaches did not have access to the randomization sequence/allocation list or outcome data beyond participant contact details and scheduling information required for intervention delivery.

The tele-exercise (Table S1 in Multimedia Appendix 1), consisting of both aerobic and resistance exercises with intensity progressively increasing to a moderate level, was delivered to patients in the intervention group for 7 weeks (150 minutes per week). Delivery mode was transitioned gradually from supervised in week 1 (3 sessions/week, 30‐50 minutes/session) to mixed mode with 1 supervised session and 2 unsupervised sessions in week 7 (Table S2 in Multimedia Appendix 1).

Twelve participants formed two groups (6 participants per group) according to their individual exercise intensity levels determined prior to the intervention for each participant. Each group received a total of 15 online supervised training sessions. Each session consisted of warm-up, aerobic exercise, resistance exercise, stretching, and cool-down, with the exercise period progressively increasing from 30 minutes to 50 minutes and exercise intensity progressively increasing from low to moderate levels. Lighter intensity was adopted when there was a need (eg, for those with severe conditions). Target heart rate during exercise was monitored by a wearable activity monitor watch (Mi Band 5, Xiaomi Corporation) to ensure the real exercise intensity (%HRR, heart rate reserve) fulfilled the individualized requirements. Cameras were requested to be turned on all the time. Sessions were delivered by a certified fitness instructor (coach) and an assistant coach. Both were coinvestigators of the study. The lead coach held a Certificate in Fitness Instruction (Hong Kong Baptist University), and the assistant coach held a master’s degree in Sport and Leisure Management (Hong Kong Baptist University). For quality control, YG (the principal investigator) participated in the first 3 sessions in week 1 and attended 1 session per week in weeks 2‐4 without prior notification. Figure 1 illustrates some exercise examples. Starting from week 4, unsupervised sessions were introduced, with one session in weeks 4 and 5 and two sessions in weeks 6 and 7. Participants were asked to complete these sessions at home by watching pre-made videos for them. They were also asked to record their exercise (eg, session duration) in an exercise log and to record heart rate using the Mi Band 5.

In order to promote habit formation of exercise, in each supervised session, we included effective behavior counseling elements. These elements were guided by an integrated theoretical framework targeting motivational, volitional, and habit-formation processes, drawing on dual-process accounts of behavior (reflective vs automatic processes) and habit formation theory to support both behavior change and maintenance. The counseling content was developed based on established psychological theories and BCTs used in our previous studies [12-15] and tailored to unique facilitators and barriers to exercise in breast cancer survivors. Seven psychological counseling sessions (30 minutes per session) were delivered by the same coach and assistant from the tele-exercise sessions at a frequency of once per week for 7 weeks. The lead coach had previous experience as a research assistant in a tele-exercise intervention for patients recovering from COVID-19, co-led by the principal investigator (YG), where she supported the delivery of educational workshops. For the present trial, the coach completed three 1-hour training sessions focused on counseling content and techniques tailored to breast cancer survivors and undertook several practice sessions to further refine counseling delivery.

The psychological strategies aimed at changing motivation, self-regulation skills, and habit formation of exercise. Motivational content targeted reflective motivation (eg, strengthening intentions and outcome expectancies), volitional content targeted self-regulation and action control (eg, planning and self-monitoring), and habit-formation content targeted automaticity by establishing stable contextual cues and repetition. For example, the participants were advised on how to work toward their goals and maintain motivation in the face of barriers (motivational strategy); they were encouraged to set outcome-specific goals and self-monitor their progress toward the goals by monitoring how much time they spent in exercise (self-regulation strategy); and they were asked to recognize and identify situational, contextual, and time-based cues that could prompt them to be active (habit development strategy). Table S3 in Multimedia Appendix 1 details the topics, theoretical strategies, and BCTs involved in each session, and Multimedia Appendix 2 provides the detailed content of session 1 (Chinese only).

On the day of baseline measurements, participants in group A received a face-to-face briefing on installing and using Zoom and on intervention procedures. Additional support was provided before, during, and after the intervention via a hotline and social media (eg, WhatsApp). Participants were asked not to share intervention-related materials (including videos) with other patients to minimize potential contamination.

Safety and potential adverse effects of intervention: It is critical to monitor and record harm or unintended consequences in exercise interventions for breast cancer survivors in order to ensure that the interventions are safe and appropriate. In the intervention arm, a pulse oximeter (Innovo Premium Fingertip Pulse Oximeter) was used to monitor oxygen saturation in the initial exercise sessions to avoid exercise-induced oxyhemoglobin desaturation [16]. Participants with lymphedema wore a compression sleeve during exercise. Possible adverse effects, such as lymphedema, fracture, and other arm/shoulder symptoms, suggested by previous studies were monitored and remedied promptly via diverse means (eg, to inform participants of the possible adverse consequences in advance and provide feasible and practical preventive strategies to them) [16,17]. According to the American College of Sports Medicine’s guidelines for exercise, the tele-exercise was stopped immediately in the following situations during exercise: (1) deteriorated upper extremity symptoms; (2) chest pain; (3) central nervous system symptoms (eg, dizziness or near syncope); (4) fatigue, shortness of breath, wheezing, muscular cramps, or claudication; (5) heart rate during exercise ≥90% HRR (the upper limit of vigorous intensity); (6) an absolute decrease in oxygen saturation SpO2≥5% or SpO2≤80% [16].

After the intervention, all participants in the intervention group were interviewed regarding acceptability and feasibility. Secondary health outcomes were measured among all participants before and after the intervention to investigate preliminary effects of the study and to calculate the sample size of the main study.

Participants in the control group received 7 educational essays on exercise and health (1 essay per week for 7 weeks). All participants in both groups continued to receive usual care from the study hospital during the study period. Multimedia Appendix 3 provides an example essay (Chinese only).

The pilot was assessed following the RE-AIM framework. Both quantitative and qualitative data were collected via semistructured, face-to-face interviews [18]. The interview included a structured 14-item acceptability and feasibility questionnaire, with each item rated on a 5-point Likert scale (“1=strongly disagree” to “5=strongly agree”). For each item, participants were also invited to elaborate through an open-ended follow-up prompt to explain their rating and provide additional comments.

The structured Likert-scale responses were analyzed quantitatively to summarize acceptability and feasibility across domains, including the individualized exercise design, exercise-intensity progression, session frequency/duration, perceived difficulty of the online guided exercises, scheduling and staffing arrangements, transition to independent activity, perceived usefulness and safety, overall satisfaction, willingness to recommend the intervention, and acceptability of the assessment procedures. The open-ended responses were used to complement and contextualize the quantitative findings.

Given the pilot nature of the study and the supportive role of the qualitative data, we conducted a descriptive content summary. One researcher (SY) reviewed the written interview notes, extracted relevant text, and organized comments into broad categories guided by the interview domains (eg, perceived benefits, barriers, facilitators, and recommendations). The qualitative summaries were used to support the interpretation of the quantitative results rather than to generate formal qualitative themes. No qualitative analysis software was used; data were managed in Microsoft Word. Table 1 details the indicators of acceptability and feasibility.

Health outcomes were assessed pre- and postintervention for all participants, including cardiopulmonary fitness (primary outcome of the main study), muscular strength, balance, body composition, affected-side shoulder range of motion (ROM), lymphedema, and health-related quality of life. These outcomes were carefully selected based on literature and clinical importance. Table 2 lists the details of the measurements. A self-administered questionnaire was developed to collect personal information, health-related quality of life, physical activity level, and other related factors.

Quantitative data were summarized descriptively to evaluate the acceptability and feasibility of the tele-exercise. Qualitative open-ended comments were summarized descriptively to contextualize and supplement the quantitative findings. To preliminarily estimate the health effects of the tele-exercise protocol, changes in health outcomes (eg, cardiopulmonary fitness) were compared between the intervention and control groups using the independent-samples t test. Analyses followed an mITT approach including all randomized participants with baseline data. Participants who withdrew after randomization but before baseline assessments were excluded; as 3 participants withdrew postallocation (all before baseline), this may introduce selection bias. In the definitive trial, baseline assessments will be completed prior to allocation disclosure (or allocation will be concealed until after baseline) to minimize postrandomization attrition. Given the pilot nature and limited power, analyses were exploratory and focused on estimation. We report mean differences with 95% CIs and effect sizes (Cohen d; interpreted as small 0.2, medium 0.5, large 0.8), and treat P values as descriptive with no prespecified threshold for statistical significance [21]. No adjustment for multiple comparisons was applied because outcomes were exploratory. The sample size for the main trial was estimated using the repeated measures ANOVA within-between interaction approach (G*Power, Version 3.1.9.7), with parameters of α=.05, β=.8, two groups and three measurement waves. Cohen d was converted to Cohen f using f=d/√3 [22].

Of 34 individuals assessed for eligibility, 27 provided consent and were randomized, yielding a recruitment rate of 79.4% (27/34). Of those randomized, 3 participants withdrew after allocation upon learning they had been assigned to the control group and did not complete the baseline assessment; thus, 24 participants completed the baseline assessment and were included in the mITT analyses (12 per group). The remaining 7 eligible individuals declined participation prior to randomization (most commonly due to being busy or lack of interest). All 24 participants who completed baseline also completed the postintervention assessment, resulting in a baseline-to-post retention rate of 100% (24/24). Figure 2 shows the flowchart of the pilot study.

Attendance and compliance were summarized separately for supervised and unsupervised sessions (Table 3). Attendance was 88.3% (159/180) for supervised sessions and 86.1% (62/72) for unsupervised sessions. Compliance was 86.1% (155/180) and 83.3% (60/72), respectively. As shown in Figure 3, attendance and compliance were highest at the start (12/12 participants in sessions 1‐2) and then fluctuated modestly, with no obvious decline following the introduction of unsupervised sessions (diamond markers); attendance generally ranged from 9 to 12 participants and compliance from 8 to 12 participants across sessions.

Acceptability of the intervention was assessed through interviews using 14 five-point Likert items among the 12 participants who completed the intervention. All 12 participants completed the acceptability ratings; 5 participants additionally provided optional open-ended written comments. The items covered various aspects, including the individualized approach, exercise intensity, frequency and duration of sessions, difficulty level, time scheduling, staffing arrangement, the transition from supervised to unsupervised delivery modes, usefulness of the psychological elements, safety, perceived effectiveness, satisfaction, willingness to recommend the program to others, and the assessment arrangement. All participants either agreed or strongly agreed with the acceptability of the intervention for half of the items (Table 4). For the remaining items, most responses were positive, with only one neutral response per item. In the optional open-ended comments (n=5), participants reported high satisfaction with the intervention, intention to recommend it to others, and a desire for continuation. One participant described perceived health improvements, including reduced symptoms and improved emotional and mental well-being, and expressed gratitude for the staff’s professionalism and support. Regarding the neutral responses, one participant attributed her rating to missing several sessions due to scheduling conflicts with chemotherapy appointments. Another participant indicated a preference for more sessions per week and therefore rated “frequency and duration” as neutral.

Fidelity of the intervention was evaluated in terms of content, frequency, duration, and coverage of the intervention.

The results indicated that the intervention was implemented with high fidelity to the original protocol.

This pilot RCT evaluated the feasibility, acceptability, and preliminary effects of a 7-week home-based, group-based tele-exercise intervention (with a progressive transition from supervised to unsupervised sessions) plus brief psychological counseling among breast cancer survivors undergoing adjuvant therapy. The intervention demonstrated high feasibility, with a recruitment rate of 79.4% and baseline-to-post retention of 100%, and good engagement, with an overall attendance rate of 87.7% and compliance of 85.3% in the intervention group. The intervention was also highly acceptable, with all participants reporting agreement or strong agreement across key acceptability items. Although the study was not powered for definitive effectiveness testing, between-group differences in change generally favored the intervention across cardiopulmonary fitness (6MWT, primary outcome) and showed positive trends in lower-extremity strength, balance, shoulder ROM on the affected side, and several health-related quality-of-life outcomes, warranting confirmation in a fully powered trial.

The pilot demonstrated strong feasibility, evidenced by high recruitment and retention rates. These rates appear higher than those reported in many exercise and tele-exercise trials among breast cancer survivors [2,4,5,23], which may reflect the study’s pragmatic design and local tailoring (eg, highlighting participant benefits such as free assessments and aligning delivery with local care routines). Engagement was also high: overall attendance was 87.7%, and compliance was 85.3%, comparing favorably with prior exercise interventions [4,5]. The progressive transition from supervised to unsupervised sessions was rated as acceptable (Table 4), and adherence remained broadly comparable between supervised and unsupervised sessions during the intervention period (Table 3 and Figure 3). Figure 3 further suggests that attendance remained stable across the intervention, including after the introduction of unsupervised sessions, which supports the feasibility of reducing supervision without an obvious loss of participation during adjuvant therapy.

In contrast, heart rate–based compliance showed modest session-to-session fluctuations and a slight downward tendency over time. One possible explanation is physiological adaptation: as participants became more familiar with the prescribed workload, the same exercises may have elicited a lower heart-rate response, reducing the likelihood of meeting the original target heart-rate zone despite continued participation [24,25]. This suggests that more dynamic adjustments to the exercise prescription will be necessary in the definitive trial. The brief psychological counseling component may also have helped participants sustain engagement by supporting motivation and self-regulation, although its specific contribution cannot be isolated in this pilot. Tele-exercise delivery at home likely reduced common participation barriers (eg, fatigue, transportation, and time constraints), although longer-term maintenance beyond the intervention period cannot be inferred without follow-up.

Preliminary between-group estimates generally favored the intervention across cardiopulmonary fitness, lower-limb function, balance, shoulder ROM, and health-related quality of life, consistent with meta-analytic evidence in breast cancer survivors [2,4,5,26]. The estimated effect on 6MWT (Cohen d=0.65; 95% CI includes 0) suggests the primary outcome is promising to prioritize in a definitive trial. While functional capacity is clinically relevant, this pilot cannot establish clinical effectiveness or downstream outcomes (eg, morbidity/mortality) [27]. From a clinical perspective, the between-group difference in change in 6MWT in this 7-week pilot was approximately 15 m, which is below commonly cited minimal clinically important differences (MCIDs) of 20‐30 m in oncology populations [28]; this might be due to the relatively short duration of the intervention. Therefore, these findings should be interpreted as exploratory; the definitive 12-week trial is needed to determine whether the intervention can achieve clinically meaningful improvements in functional walking capacity.

Changes in physical function were also encouraging. Improvements in the timed stand test and single-leg stance are plausible given the repeated lower-body functional tasks and progressive loading embedded in the sessions. For the timed stand test, lower values indicate faster completion and better performance. These measures are clinically relevant because lower-limb weakness and balance impairment can limit mobility and confidence in daily activities. Shoulder ROM on the affected side also improved, which is notable given the frequency of posttreatment shoulder stiffness; embedding brief mobility work within each session may be a pragmatic way to address this without substantially increasing intervention burden.

Upper-extremity strength (handgrip strength) showed only small between-group differences, which may reflect limited progressive upper-body loading in the 7-week protocol and the need for more targeted upper-body resistance progression training in the main trial. BMI and percent body fat decreased modestly; however, body composition was not a primary target, and large changes were not expected over this short, nondiet intervention.

Quality-of-life measures (FACT-B Trial Outcome Index, FACT-G, and FACT-B total score) also tended to improve in the intervention group. Importantly, the between-group differences in change were approximately 9.33 (FACT-B Trial Outcome Index), 11.54 (FACT-G), and 15.04 (FACT-B total), which exceed commonly cited MCIDs for these scales (FACT-B Trial Outcome Index: 5‐6 points; FACT-G: 5‐6 points; FACT-B total: 7‐8 points) [29]. These preliminary clinically meaningful improvements may reflect both gains in physical function from the combined aerobic-resistance training and the brief behavioral counseling elements (eg, motivation, self-regulation, and habit-formation strategies), which may have helped participants manage barriers and sustain activity. Given the pilot design, findings are exploratory and require confirmation in the definitive trial.

Finally, we did not observe lymphedema exacerbation over the intervention period, and perceived safety ratings were high, which is reassuring and consistent with prior evidence that appropriately prescribed exercise is safe for individuals with breast cancer-related lymphedema [27].

This pilot study has several strengths, including a randomized design, blinded outcome assessment, the use of both objective functional outcomes and patient-reported health-related quality of life measures, and a pragmatic tele-exercise intervention that is feasible to deliver during adjuvant therapy.

Several limitations should be considered when interpreting the findings. First, as a pilot study, the trial was designed to evaluate feasibility and acceptability and was not powered to draw definitive conclusions about effectiveness. Accordingly, the small sample (n=24) from a single center in Hong Kong limited the precision of estimates and the generalizability of the findings. Second, the pilot intervention period was limited to 7 weeks, as time and resource constraints did not allow completion of the planned 12-week intervention for the main study. This reduced intervention dose may have limited changes in outcomes and may have led to an underestimation of the potential effectiveness and effect sizes achievable with the intended intervention duration. Attendance and compliance observed over 7 weeks may not be representative of adherence over the planned 12-week intervention, as participation often declines with longer intervention duration. Third, follow-up assessments were not conducted; therefore, the longer-term maintenance of behavior change and durability of outcome improvements could not be evaluated. Fourth, eligibility criteria (language and smartphone use) may limit generalizability; future studies should broaden inclusion and provide support for participants with lower digital access and literacy. Fifth, randomization occurred before baseline assessment. This design may increase postallocation withdrawal due to treatment preference and can introduce selection bias into the analyzed sample. In the future study, baseline assessments will be completed prior to randomization (or, alternatively, randomization will be retained but allocation will be disclosed only after baseline assessment) to minimize postrandomization attrition.

The results of this pilot study provide a foundation for the design and implementation of a larger-scale main study. Based on the observed primary outcome effect size, the estimated sample size for the main study is 102 participants (51 per group). Future research should (1) include postintervention follow-up to evaluate whether benefits persist after supervision is reduced, (2) strengthen fidelity and adherence monitoring and reporting, and (3) consider enhancing upper-body progressive resistance training if upper-limb strength is a priority outcome. In addition, the integration of psychological theories and BCTs in the intervention design may have contributed to habit formation and sustained exercise participation. These elements should be refined and tailored to address the needs and barriers faced by breast cancer survivors across different contexts and settings.

This pilot RCT demonstrates that a home-based, group-based tele-exercise intervention with brief psychological counseling is feasible and acceptable for breast cancer survivors undergoing adjuvant therapy, with high engagement and no apparent safety concerns. Exploratory between-group estimates for cardiopulmonary fitness, functional performance, affected-side shoulder ROM, and health-related quality of life (effect sizes and 95% CIs) provide parameters to inform the design of a fully powered RCT with longer follow-up to evaluate effectiveness and sustainability.