This study used a single-center cluster randomized controlled trial (CRCT) to evaluate the effectiveness of PNO2PSP. The control group received routine care, while the intervention group received the comprehensive PNO2PSP intervention.

Convenience sampling was used to recruit newly diagnosed young- to middle-aged patients with breast cancer who were admitted for surgical treatment to the Department of Breast Surgery, Sun Yat-sen Memorial Hospital (Textbox 1).

Based on the sample size calculation formula for experimental studies, using a 2-tailed test, the primary outcome measure of this study was the score on the psychosocial adjustment scale. Referencing previous research [39], the sample size per group was calculated to be at least 18 participants. To account for potential dropouts and other factors, a 15% increase was applied, resulting in 21 participants per group, for a total of 42 participants.

To further accommodate the cluster-randomized design at the enrollment week level, reference was made to previous longitudinal psychosocial intervention studies, which reported a design effect (DE) of 1.472 [40]. By applying this DE to the initial estimate (42×1.472=62.04), it was determined that a minimum of 62 participants was required to maintain adequate power. Ultimately, 70 participants (35/group) were recruited to ensure robust statistical inference and to provide a buffer for attrition.

The PNO2PSP intervention required peer support groups with at least 3 participants. This requirement was primarily grounded in prior peer-support studies, which indicate that a minimum of 3 members is essential to create a collaborative atmosphere and foster a diverse exchange of knowledge and experiential sharing [31]. From a clinical feasibility perspective, this threshold also ensured that the support group remained functional and stable even if 1 member was temporarily unavailable or withdrew. Given these requirements, individual randomization was impractical for ensuring stable group formation among young- to middle-aged patients with breast cancer undergoing treatment. Consequently, a CRCT design was adopted, with weekly cohorts of surgical patients serving as natural clusters [41]. Participants scheduled for surgery in the following week were recruited every Sunday to form a weekly cohort. Random numbers were generated using R (R Core Team) in RStudio (Posit PBC). The allocation results for intervention and control groups were placed in sealed, opaque envelopes prepared in advance by an independent research assistant responsible for maintaining allocation concealment. After recruitment each week, the corresponding envelope was opened to determine group assignment. Since most patients were discharged by Friday of their surgery week, this cohort-based grouping minimized the risk of intervention contamination between the intervention and control groups. Due to the intervention’s nature, blinding of participants and providers was not feasible, and self-reported outcomes precluded blinding of outcome assessors. Thus, a single-blind design was used, with only statistical analysts blinded to group assignments to reduce bias.

The control group received routine care, which included standard discharge education and follow-up. The discharge education was primarily provided by nurses through verbal health education, covering topics such as postoperative wound care, protection of the affected limb, and rehabilitation exercises. Follow-up involved sending reminders to participants about regular dressing changes, encouraging adherence to chemotherapy schedules, and prompting them to attend follow-up visits.

The “Peer Support Intervention Manual” was distributed to the participants in the intervention group. Research assistants categorized participants and peer supporters into groups based on general demographic information, such as age, educational background, marital status, and reproductive history. Each group consisted of 2 peer supporters, 1 intervention nurse, and 3‐5 group members. The groups were then assigned to implement the intervention. The intervention was implemented according to the finalized version of the peer support intervention program developed in the earlier stages. The intervention details are shown in Multimedia Appendix 1.

This study uses both quantitative and qualitative methods for outcome assessment. The quantitative evaluation primarily focuses on changes in the primary and secondary outcome measures through survey scales, while the qualitative evaluation involves interviews to understand participants’ experiences.

In addition to the standardized scales, additional data were collected to provide information on the demographic and clinical characteristics of the participants and to assess treatment burden, as follows: (1) demographic information: sociodemographic data, including age, education level, marital status, residence, employment, and family income; (2) disease information survey: cancer type, disease stage, surgery type, treatment methods, lymph node involvement, and axillary lymph node dissection; and (3) treatment burden survey: patients’ treatment burden was assessed during the final follow-up (week 12) through a self-reported survey. This included treatment costs, unplanned hospital visits, outpatient visits, and total hospitalization time.

A purposeful sampling strategy was used to recruit participants from the intervention group. To ensure rich and diverse perspectives, participants of different ages, disease stages, and levels of psychosocial adjustment based on quantitative scores were invited. Semistructured, face-to-face, one-on-one interviews were conducted using an interview guide. Core questions included the following: (1) What is your impression of this project? (2) How has this project impacted your psychosocial adjustment? (3) What are the strengths and weaknesses of this project? The entire interview process was audio-recorded.

Participants who were admitted to the hospital wards were selected according to the inclusion and exclusion criteria of this study. Once eligible patients were identified, the purpose and significance of the study were explained to them, informed consent was obtained, and a baseline data questionnaire was completed. The completed forms were checked for missing items on site, and once verified, they were numbered to ensure data quality. Eligible participants were randomly assigned to the intervention group or the control group, and interventions were conducted accordingly. Based on previous studies, this study established follow-up data collection at 3 time points: 4 weeks post surgery (T1), 8 weeks post surgery (T2), and 12 weeks post surgery (T3) [23]. At baseline, demographic information, disease information, psychosocial adjustment, self-efficacy, social support, and coping modes were collected. At T1, T2, and T3, psychosocial adjustment, self-efficacy, social support, and coping modes were reassessed. At T3, additional information on treatment costs, number of doctor visits, hospitalizations, and unplanned readmissions was collected.

For the qualitative part, participants were contacted 1‐2 days before the interview to explain the study’s purpose and procedures. Interviews were conducted after obtaining written informed consent. All interviews took place in a quiet, private room in the ward and were conducted individually by trained researchers with extensive qualitative interviewing experience. At the beginning of each interview, participants were informed that there were no right or wrong answers and were encouraged to share their experiences openly. They were also told that they could pause or stop the interview at any time if they felt uncomfortable. All data were anonymized to ensure confidentiality. The sample size for the qualitative phase was determined by data saturation. Data analysis was conducted alongside data collection, and recruitment ceased when no new themes emerged in 3 consecutive interviews [45].

Descriptive statistics were used to summarize participants’ baseline characteristics. Continuous variables were compared using independent samples t tests or Mann-Whitney U tests, while categorical variables were analyzed using Pearson chi-square (χ²) tests or Fisher exact tests. Baseline balance was assessed between the intervention and control groups, as well as between participants who completed the study and those lost to follow-up. This study followed the intention-to-treat principle, including all participants who were randomized. No formal imputation was performed for missing data. Generalized estimating equations (GEEs) use all available observations under the missing-at-random assumption and provide valid population-averaged estimates without requiring explicit imputation.

The GEE model was used to evaluate the intervention effect. “Week” was defined as the cluster variable to adjust for cohort-level correlations, and “TIME” was set as the repeated factor. Data were sorted by participants’ hospital identification numbers to ensure the integrity of the repeated measures structure in the model. TIME was treated as a categorical variable, and the group × time interaction effect was tested. The models specified a Gaussian identity link function with an exchangeable working correlation structure. To ensure the reliability of the inferences, robust (sandwich) SEs were used to provide unbiased parameter estimates. Clustering was explicitly modeled at the week level to account for potential correlation within cohorts. This approach ensures that the SEs are adjusted for the nested structure of the data.

The primary analysis consisted of unadjusted GEE models, which included group, time, and the group × time interaction as predictors. The control group and baseline (T0) served as the reference categories. To test the robustness of the primary findings against potential confounding—specifically the baseline imbalance observed in coping modes (avoidance)—sensitivity analyses were performed using adjusted GEE models. In these models, baseline coping modes (avoidance) were included as a covariate. These analyses confirmed that the statistical significance and direction of the intervention effects remained consistent across both models, reinforcing the reliability of the primary results. Further analysis revealed no significant interaction between group size and the intervention effect on the primary outcome. Statistical significance was defined as P<.05. All analyses were performed using R.

Qualitative data collection and analysis were conducted concurrently. All interviews were transcribed verbatim within 24 hours after completion. Data were analyzed using content analysis. Two researchers independently performed open coding, grouped similar codes into categories, and further abstracted them into core categories. Relationships among categories were examined during the abstraction process. Discrepancies were resolved through team discussion until consensus was reached, and triangulation was used to integrate qualitative and quantitative findings, allowing for a more comprehensive interpretation of how and why the intervention produced its observed effects [46].

A standardized protocol for the implementation of the intervention was developed, and all team members involved in the intervention received uniform training to minimize inconsistencies during the execution of the plan. Research assistants participated in both online and offline group discussions throughout the intervention process. All offline interventions were video-recorded, online interventions were screen-recorded, and all voice chats were audio-recorded to ensure comprehensive documentation. The research team held meetings at least once a week to identify issues and difficulties in the implementation of the research protocol. All peer supporters underwent uniform training and assessment. Additionally, intervention guidelines were provided to assist them in the implementation process.

To minimize contamination between groups, a cluster enrollment strategy based on admission week was adopted. Patients in the intervention and control groups were recruited in different weeks, and due to the fixed surgical admission and discharge schedule, participants from different weeks did not overlap during hospitalization. This temporal separation reduced opportunities for cross-group interaction. In addition, all participants signed a confidentiality agreement before the intervention, which emphasized that all information about intervention content and personal experiences of group members should not be disclosed to individuals outside the group. Intervention delivery was restricted to designated facilitators, and no crossover of intervention staff occurred between study arms.

This study complied with the principles of the Declaration of Helsinki and relevant Chinese regulations governing clinical research. It was approved by the Ethics Committee of the School of Nursing, Sun Yat-sen University, Guangzhou, China (approval no L2023SYSU-HL-019) and registered with the China Clinical Trial Registry. All relevant departments in the hospital approved the study. Participants in the study signed a paper copy of the informed consent before enrollment. They were provided with detailed information about the study’s objectives, process, potential risks and benefits, their right to withdraw at any time, and data collection methods. The data obtained in the study were handled in accordance with confidentiality principles. All data were anonymized before analysis. Personal and clinical information was securely stored, and access was restricted to authorized members of the research team only. Identifiable features or images of patients are not included in the paper and supplementary materials. Participants did not receive any financial compensation for their involvement in the study. However, they were provided with free educational support as part of the research process. To ensure that all participants benefit equitably, after the data collection was completed, the participants in the control group received the intervention manual and were added to online peer support groups.

During the study, 90 young- and middle-aged patients with breast cancer who met the inclusion criteria were approached. Among them, 20 refused to participate, 10 expressed no interest, 7 were unwilling to complete the questionnaires, and 3 did not have time to participate in the intervention. As a result, a total of 70 participants from 6 recruitment clusters (weeks) were included in the final analysis. For the primary outcome, the observed intraclass correlation was 0.0303, which is below the threshold of 0.05 and therefore indicates a small degree of clustering according to previous studies [47]. Given an average cluster size of 11.67, the DE was calculated to be 1.323. Post hoc power analysis using PASS (NCSS LLC) demonstrated that with this DE and sample size, the study achieved a statistical power of 0.875 to detect the intervention effect on the primary outcome, confirming the adequacy of the sample size for robust inference.

At T1, 33 control group participants and all 35 intervention group participants completed the follow-up, with 2 losses in the control group. At T2, 31 control and 32 intervention participants completed the follow-up, with 4 and 3 losses, respectively. At T3, 28 control and 27 intervention participants completed the follow-up, with 7 and 8 losses, respectively. Overall, the attrition rate for the control group was calculated as 12.4%. The main reasons for attrition were loss of contact or unwillingness to complete the follow-up questionnaire. The attrition rate for the intervention group was 10.5%. The main reasons for attrition were unwillingness to continue participation, lack of time to complete the questionnaire, or loss of contact. A flowchart of patient recruitment, enrollment, follow-up, and attrition is shown in Figure 1.

The baseline data comparison between the follow-up completion group and the lost-to-follow-up group showed no statistically significant differences in general information, disease, and treatment data (P>.05; shown in Table 1). However, there was a statistically significant difference in social support (P=.038). The comparison of baseline data between the intervention group and the control group revealed no statistically significant differences in general information, disease, and treatment data (P>.05). However, there was a statistically significant difference in the avoidance dimension of coping strategies (P=.003; shown in Table 2). Subsequently, the GEE analysis was conducted, with these scores used as a covariate to adjust the results (shown in Table 3).

The results of the GEE model, intergroup comparison analysis, and within-group comparison analysis in this study are shown in Table 3.

The intervention group had significantly lower psychosocial adjustment scores than the control group at T2 (P<.001) and T3 (P=.006).

Additionally, the intervention group had significantly lower scores in the vocational environment dimension at T3 (P=.020), in the sexual relationship dimension at T2 (P=.001) and T3 (P=.002), as well as in the psychological distress dimension at T2 (P<.001) and T3 (P<.001; Table 4). The intervention group reported significantly higher social support scores than the control group (P=.006) at T2. Confrontation coping scores were also significantly higher in the intervention group (P=.019) at T2, (P=.002) at T3. In contrast, avoidance coping scores were significantly lower in the intervention group (P=.004, .006, and .017) at T1, T2, and T3, respectively. No significant differences were observed between groups in self-efficacy or acceptance-resignation coping at any time point (Table 4, Figures 2-7, and Multimedia Appendix 2).

There was a statistically significant difference between the 2 groups in unplanned hospital visits (Z=−0.354; P=.01), indicating that the intervention group had fewer unplanned hospital visits than the control group (shown Multimedia Appendix 3).

Young- to middle-aged patients with breast cancer face significant psychosocial challenges, which can impair immune function and worsen prognosis [8,48]. Addressing these issues with effective interventions is crucial—the PNO2PSP notably improved psychosocial adjustment, supported by qualitative interviews. A similar trial by Høybye et al [49] found improvements in helplessness and depression, though without affecting broader mental adjustment or mood disturbance. The difference likely stems from the lack of trained peer supporters and structured interventions in that study [50]. In contrast, this study used various support formats, including peer education and meetings with nurses, ensuring more stable social networks and addressing real concerns. Peer support groups also offered a space for participants to communicate anytime and anywhere, addressing adjustment issues promptly [25,51].

In terms of feasibility, this study recruited participants preoperatively, with a total of 90 patients contacted and 70 (77.8%) consenting to participate. Attrition rates were 12.4% in the control group and 10.5% in the intervention group, primarily occurring during the follow-up phase after the intervention. During the first 8 weeks of the intervention, the overall participation rate in the intervention group was high at 86.3%. This high rate can be attributed to the intervention themes being tailored to the specific psychosocial adjustment challenges faced by young- to middle-aged patients with breast cancer. Additionally, the blended online and offline intervention model reduced logistical and financial burdens, allowing participants to participate from home [52]. The intervention’s feasibility was further enhanced by its design. Participants scheduled for surgery were recruited during their hospital stay, allowing ample time for baseline assessments and the initial offline intervention session, which facilitated participant engagement. Expert consultations and focus group interviews determined the blended intervention format. The initial offline session established connections and trust between peers and participants, while subsequent online sessions minimized the need for hospital visits, increasing participation and motivation. The study established online support groups for participants, where the actual frequency of group discussions exceeded the initially planned schedule, reflecting a high level of participant engagement. This enthusiasm was largely driven by the peer supporters, who regularly followed up with participants in the online support groups, provided necessary information, and responded promptly to participants’ questions. For issues requiring professional expertise beyond the scope of peer supporters, nurses stepped in to provide timely assistance. Furthermore, each online support group included at least 2 peer supporters, ensuring that if 1 was unavailable, the other could respond without delay.

PNO2PSP significantly improved the psychosocial adjustment of young- to middle-aged patients with breast cancer, a finding further supported by the results of qualitative interviews. Similar findings were reported by Høybye et al [49], who conducted a randomized controlled trial of an internet peer support group for patients with breast cancer. However, their study showed improvements in helplessness, confusion, and depression, with no significant effects on other dimensions of mental adjustment to cancer or total mood disturbance. The difference in results likely stems from the previous study’s reliance on mutual support among participants without trained peer supporters [50]. Additionally, the intervention lacked clear themes and fixed schedules, and the online-only format made it difficult for participants to establish a stable social network. In contrast, this study adopted a variety of peer support formats. Peer education involved trained peers sharing their personal experiences and coping modes with the disease. Peer group meetings provided a platform for participants to interact with peer supporters and nurses, offering an opportunity for emotional expression. Peer support groups offered a space for participants to communicate anytime and anywhere, addressing adjustment issues promptly [25,51]. Moreover, the intervention themes in this study were developed based on specific adjustment issues identified through qualitative research, ensuring that they addressed real and relevant concerns faced by young- to middle-aged patients with breast cancer.

At the same time, the analysis of the subdimensions of psychosocial adjustment showed that the intervention improved participants’ scores in the sexual relationship dimension. A previous study found that peer support did not effectively improve sexual relationships among patients with breast cancer [29]. The discrepancy may lie in the fact that previous peer support studies, even when addressing sexuality-related topics, focused only on knowledge dissemination, offering participants limited actionable advice [53]. In this study, through the use of intervention manuals, peer supporters’ experience sharing, and nurses’ education, participants received knowledge related to sexual health. This not only informed participants that they could engage in sexual activities postsurgery but also encouraged them to actively discuss and resume sexual activities [54]. Participants were provided with techniques to alleviate discomfort during sexual intercourse, and peer supporters shared practical ways to maintain intimacy beyond intercourse, such as kissing and hugging [55]. These actionable strategies significantly improved sexual relationships. Influenced by traditional Chinese culture, sexual health education has long been a challenge in educating patients with breast cancer [14].

However, because peer supporters in this study were all women of similar age and with similar experiences, participants could freely discuss sexuality-related issues without feeling embarrassed [55]. Therefore, peer support could be a viable approach for improving sexual health among patients with breast cancer in the future. The intervention also improved participants’ psychological distress scores. Interviews revealed that participants believed peer support helped alleviate negative emotions, reducing their suffering, anxiety, and fear, which aligns with the findings of previous studies. A previous systematic review on peer support for patients with cancer indicated that peer support interventions effectively reduce depression and anxiety [56]. In many Asian societies, cultural norms surrounding emotional expression, illness disclosure, and discussion of family matters may influence how openly patients communicate within peer support groups. This cultural context may partly explain the limited effects observed in family-related dimensions of psychosocial adjustment [57]. However, there was no statistically significant improvement in the social relationship dimension, possibly due to the short follow-up period. At the end of follow-up, many participants were still in treatment and had not yet resumed social reintegration. Nonetheless, qualitative interviews showed that participants believed peer support could promote their occupational and social adjustment.

The findings revealed that the intervention could enhance the levels of social support, consistent with the results of previous studies [50,58]. In the design of this study, peer support provided participants with informational, emotional, and appraisal support. When participants had questions about their disease or treatment, peer support offered informational support and guided them to seek professional support, family support, and other social support resources [25]. When participants experienced negative emotions due to illness and treatment, peer supporters provided emotional support through listening, comforting, and encouraging, which improved individuals’ perceived levels of social support [59]. Additionally, appraisal support helped participants reevaluate the support resources they possessed, enhancing their subjective perception of social support [60]. This program also facilitated the adoption of problem-focused coping strategies while reducing the use of avoidance coping strategies. The interview results with participants supported this finding. A prior study on peer support for adolescent patients with cancer [61] yielded similar results, where participants who received peer support adopted more coping strategies 3 months postsurgery. The reasons for these outcomes include peer supporters sharing their own experiences in addressing psychosocial issues, providing participants with reference points that reduced their fear and avoidance of disease-related challenges [62]. Through the intervention, participants’ attitudes toward their illness underwent a significant transformation—they became more willing to confront cancer directly, ceased avoiding cancer-related issues, actively sought ways to resolve problems, and approached their condition with greater positivity and optimism [63].

Our qualitative findings supported that the intervention contributed to increasing treatment confidence among participants. However, the intervention did not produce a statistically significant improvement in self-efficacy among young- to middle-aged patients with breast cancer. Similarly, a prior study [64] reported that a 6-month peer counseling intervention for patients with breast cancer failed to significantly enhance their self-efficacy. A systematic review on peer support interventions for patients with cancer also found that peer support interventions did not improve participants’ self-efficacy [56]. However, a previous study demonstrated that a 6-week dyadic peer intervention improved patients with breast cancer self-management self-efficacy among patients with breast cancer [65]. The potential reasons for this discrepancy may be related to the measurement tool used in this study. The GSES used may lack the sensitivity needed to detect changes in the self-efficacy specific to coping with illness [66]. Mechanistically, self-efficacy relies on structured mastery experiences, whereas our program emphasized peer empathy and informational support over skill-based tasks. Furthermore, the robust external support from nurses and peers may have improved adjustment but temporarily bypassed the need for participants to develop independent self-reliance. Alternatively, peer support may primarily enhance disease-specific self-efficacy, which has a limited impact on overall self-efficacy [65]. Another possible explanation is that self-efficacy tends to be a stable personal attribute that is difficult to change over a short period [56].

The study results also showed that the intervention significantly reduced the number of unplanned hospital visits among participants. Similarly, a previous study [67] found that peer support interventions for patients after coronary artery bypass surgery led to reduced health care usage 3 months postsurgery, with fewer visits to general practitioners and emergency departments. This reduction is likely attributable to the synergy between peer supporters, who provided immediate experiential guidance for common postdischarge concerns, and intervention nurses, who offered professional triage and symptom management advice. This collaborative approach effectively addressed minor complications that might otherwise have triggered unnecessary hospital visits. This reduction in hospital readmissions may help alleviate patients’ financial burden and reduce the strain on health care systems. Although the result confirmed the robustness of this finding, the reliance on self-reported health care usage data remains a limitation. Future large-scale trials should integrate clinical records with a formal economic framework to conduct a comprehensive cost-utility analysis.

This study has certain limitations. This study used a single-center CRCT with a limited sample size (N=70), which may restrict the generalizability of the findings. Furthermore, as the primary outcomes were based on self-reported scales, this lack of blinding may have introduced potential self-report bias. In addition, participant attrition during the follow-up period may further limit the robustness and generalizability of the results. Future research should consider multicenter, large-sample trials to further evaluate the effectiveness of peer support. Although this study focused on early postoperative intervention and follow-up, extending the intervention period in future research may further enhance participants’ psychosocial adjustment during the recovery phase. Furthermore, this study only collected self-reported patient expenditures without performing a comprehensive economic analysis.

Peer support represents a potentially more cost-effective and sustainable form of support. This study developed a PNO2PSP for psychosocial adjustment in young- to middle-aged patients with breast cancer, guided by a complex intervention framework. The PNO2PSP demonstrates strong scientific validity and feasibility. It effectively improves the psychosocial adjustment levels of young- to middle-aged patients with breast cancer, enhances social support, and promotes positive coping modes. By continuously training peer supporters and conducting peer support interventions, the psychosocial adjustment of young- to middle-aged patients with breast cancer can be further promoted, and this PNO2PSP could be standardized and widely implemented in clinical practice in the future.