We conducted a parallel-group, multicenter, randomized controlled superiority trial. The study was prospectively registered on ClinicalTrials.gov (NCT05070624) on September 27, 2021, prior to enrollment of the first participant on September 2, 2022.

This study was conducted and reported in accordance with the CONSORT (Consolidated Standards of Reporting Trials) 2025 statement (Checklist 1), the CONSORT E-HEALTH checklist (Checklist 2), and the TIDieR (Template for Intervention Description and Replication) checklist (Checklist 3) [21-23].

Three protocol amendments were made following trial registration and ethics approval. The primary and secondary outcomes remained unchanged throughout the study.

Family caregivers of individuals with neuromuscular disease contributed to the development and refinement of the digital peer support intervention through prior beta-testing and feedback on intervention content, delivery format, and discussion topics. Peer mentors with lived caregiving experience also informed the design of the mentor training procedures and weekly discussion sessions [18]. Patients and the public were not involved in recruitment, data analysis, or interpretation of study findings.

We recruited participants and mentors from 7 sites across Ontario, Canada, through self-referral via social media and national charities and organizations. Participants were recruited from August 2022 to September 2023 (see Multimedia Appendix 1 for study sites, relevant charities and organizations, and trial design). Interested individuals completed an expression of interest form sent via email to the study coordinating center to determine eligibility and were subsequently recruited by research coordinators.

Inclusion criteria for study participants were (1) family caregiver of an adult or child with NMD living in Canada, (2) speaks and reads English, (3) access to the internet and a computer or tablet, and (4) consents to participation. Inclusion criteria for peer mentors were the same as those for participants, with the additional requirement of completing our digital peer support training. No exclusion criteria were applied.

This trial received research ethics approval through Clinical Trials Ontario (CTO-3590). Participants and mentors provided written electronic informed consent using the REDCap (Research Electronic Data Capture; Vanderbilt University) database in accordance with the Declaration of Helsinki [24].

Participant privacy and confidentiality were protected through the use of unique study identification numbers and deidentified data analysis. Study data were securely managed using REDCap hosted at The Hospital for Sick Children, and TouchAway (Aetonix, Ottawa, Canada) data were stored on secure Canadian cloud servers in accordance with Ontario’s Personal Health Information Protection Act. Access to identifiable data was restricted to authorized study personnel. All reported findings are presented in aggregate form to protect participant confidentiality.

Upon study completion, participants were provided a CAD $30 (US $21.11; exchange rate CAD $1=US $0.70, converted on July 6, 2026) online gift card.

No participant-identifying information is included in this manuscript or supplementary materials.

Research coordinators enrolled eligible participants, and participants were randomized through the REDCap randomization module. Participants were randomized in a 1:1 ratio using a computer-generated randomization schedule with random permuted block sizes of 2 or 4 generated by an independent statistician. Allocation concealment was maintained through the REDCap randomization module, which was inaccessible to study personnel until assignment. Due to the nature of the intervention, participants, peer mentors, and study personnel were not blinded to the group allocation. Outcome assessors were also not blinded due to resource limitations.

Our primary study outcome was caregiver mastery measured using the Pearlin Mastery Scale (PMS) at 12 weeks and adjusted for baseline score. The PMS is a valid and reliable 7-item scale [27] with scores ranging from 7 to 49; higher scores indicate higher levels of mastery [28]. The PMS has been used in previous studies of caregivers of people with various conditions, including respiratory failure and cancer [29-32], and studies of web-based interventions [17]. We also measured PMS at 24 weeks.

Secondary outcomes were measured at baseline, 12 weeks, and 24 weeks and included (1) caregiver stress and competence (Caregiving Stress Scale [CSS]), (2) caregiver burden (Zarit Burden Interview [ZBI]), and (3) depression, anxiety, and stress (Depression and Anxiety Stress Scale-21 [DASS-21]).

The CSS is a valid and reliable tool [33,34] comprising 15 independent subscales (score range 1‐4); higher scores reflect better outcomes [7]. This scale is used in caregiving research to assess self-appraisal of efficacy in caregiving [35-37]. The ZBI is a valid and reliable tool widely used to evaluate caregiver burden [38-40]. It has 22 items with total scores ranging from 0 to 88; higher scores reflect higher levels of caregiver burden [41]. The DASS-21 is a valid and reliable 21-item tool [42,43] with subscale scores ranging from 0 to 42; higher scores indicate higher levels of depression, anxiety, or stress [44]. Participants completed the Caregiving Assistance Scale at baseline to assess caregiving contextual factors [45].

Via the aTouchAway app, we collected data on text message, phone, and video call volume sent by mentors and participants and monitored participation in weekly discussion sessions. We established the following criteria to determine fidelity [25]:

Given the nature of the intervention, risks to participants were considered minimal. Potential harms included emotional distress during peer support discussions, including disclosures of caregiver challenges. Peer mentors were trained to recognize participant distress and escalate concerns to the study team, which included social workers and psychiatry professionals.

All study data were collected using case report forms developed in the REDCap database by trained research personnel. Participants completed study measures either independently via this database accessed through an email link with up to 3 reminder prompts or with assistance from a research coordinator via telephone, according to their preference.

Basing our assumptions on a previous study of a web-based intervention for family caregivers [17], we assumed a minimum clinically important difference in the PMS of 2.95 points, an SD of 4.5, with 90% power and α of 5%. We required a sample size of 50 study participants in each group (n=100 participants total). Therefore, the study was appropriately powered to detect a clinically meaningful difference in the PMS. We completed a post hoc power analysis to identify the minimum detectable difference for the primary outcome. As the initial sample size calculation was performed using a version of the PMS with scores ranging from 7 to 28 as opposed to 7 to 49, the minimum clinically important difference for this study was adjusted to 5.16 with an SD of 7.87 for the post hoc analysis.

Mean (SD) PMS scores at 12 weeks were 29.8 (9.3) for the intervention group and 29.4 (8.7) for the control group, with an adjusted mean difference of 0.67 (95% CI −1.75 to 3.10; Table 2).

We found no difference between intervention and control group scores for PMS measured at 24 weeks or the ZBI, DASS-21, or CSS measured at 12 and 24 weeks (Table 3).

Controlling for baseline PMS score, we found no association between the 12-week PMS score and caregiver age, years of caregiving, medical diagnosis, ventilation type, and randomization arm (Table 4).

Exploratory analysis of secondary outcomes (Multimedia Appendix 2) found the following associations, controlling for baseline scores: fewer caregiving years (estimate −0.23, 95% CI −0.45 to −0.01) was associated with a lower 24-week PMS score, indicating lower mastery; older caregiver age (estimate 0.11, 95% CI 0.1 to 0.22) and care recipient SMA diagnosis (reference: amyotrophic lateral sclerosis [ALS]; estimate 5.21, 95% CI 0.11 to 10.31) were associated with higher 12-week and 24-week (SMA compared to ALS; estimate 7.07, 95% CI 1.04 to 13.09) DASS anxiety scores, indicating more psychological morbidity. Conversely, being a man was associated with lower 24-week DASS anxiety scores (estimate −3.01, 95% CI −5.98 to −0.03) and 24-week DASS stress scores (estimate −4.82, 95% CI −8.15 to −1.49), indicating less psychological morbidity. Caregivers of those with a myopathy diagnosis had lower 24-week CSS caregiver competency scores (reference: ALS; estimate −2.68, 95% CI −5.32 to −0.05), indicating less caregiver stress.

Thirty-one (62%) intervention participants and 11 (92%) mentors engaged with at least 1 program component for 8 of the 12 weeks. Among the specific elements, participants and mentors interacted most frequently through app-based messaging. Twenty-one (42%) participants contacted their mentor during 8 of the 12 weeks. Peer mentors sent a mean of 21.3 (SD 33.3) messages per participant they mentored, while participants sent a mean of 17.7 (SD 33.0) messages to their respective mentor (Table 5).

Fifteen (30%) intervention participants and 2 (17%) mentors attended the weekly discussion sessions for 8 of 12 weeks. Only 4 (8%) participants posted on the asynchronous discussion forum for 8 of 12 weeks. One (2%) participant withdrew after completing the 12-week program but before completing postintervention questionnaires.

In this randomized controlled trial of a 12-week digital peer-support intervention for family caregivers of children and adults with NMD, we found no difference between the intervention and control arms for our primary outcome of caregiver mastery measured at 12 weeks and adjusted for baseline score. Similarly, there were no differences in our secondary outcomes, including caregiver stress, competence, burden, anxiety, and depression. Our exploratory analyses found that fewer caregiving years were associated with less mastery, while older caregiver age and caregiving for an individual with SMA were associated with worse psychological outcomes. These findings suggest that digitally delivered peer support alone may be insufficient to improve psychosocial outcomes in a heterogeneous population of caregivers with established caregiving experience.

Our study did not observe an effect on caregiver mastery, an outcome that has been infrequently examined in previous trials of digital peer support interventions [48,49]. Caregiver mastery is defined as a positive belief in one’s own ability to manage the challenges of caregiving [28]. Mastery is derived from a sense of ability to handle caregiving demands and perceived self-efficacy and is protective against stress and the negative mental health sequelae of caregiver burden [17,28,31]. We therefore chose caregiver mastery as our primary outcome measure, as this best reflected the intended benefits of our virtual peer support program. However, other previous studies on digital peer support for caregivers have focused on outcomes such as caregiving competence, burden, stress, depression, and anxiety [50-56]. Among studies that did measure caregiver mastery, the interventions were mainly structured educational programs or personalized coaching [48,49]. For example, mastery, measured as a secondary outcome, was significantly improved using an 8-week self-management program for caregivers of individuals with dementia, combining face-to-face coaching by trained psychologists with online forums for peer interaction [52]. In a trial of a web-based intervention for stroke patients and their caregivers that included both peer and professional support via online chat sessions, no direct effect on mastery (secondary outcome) was found; however, reductions in caregiver depression were significantly associated with increased mastery [56]. These studies suggest that caregiver mastery may be more responsive to interventions incorporating self-management or psychoeducational strategies rather than to peer support alone.

In our exploratory analyses, we found that a shorter caregiving duration was associated with lower mastery. This may reflect the challenges faced by newer caregivers who are still adapting to their roles and lack established coping strategies [57]. Older caregiver age and an SMA diagnosis were associated with higher anxiety. Anxiety has been linked to increased health concerns and care demands, physical and mental fatigue, social isolation, financial strain, and difficulty engaging with digital platforms [58-65]. Men caregivers reported lower psychological morbidity compared to women caregivers, consistent with prior research [66-68]. Myopathy diagnosis correlated with lower caregiver competence, likely reflecting the condition’s complexity and limited support [69]. It is also important to note that the CIs for these findings were quite wide, limiting our confidence in the precision of the estimates and their effect on the outcomes of interest.

Several intervention-related and participant-related factors may explain the absence of measurable intervention effects on caregiver mastery. First, the 12-week intervention duration may be too short to produce an effect on mastery. Second, inclusion of caregivers for children and adults with any form of NMD introduced sample heterogeneity, potentially limiting the ability to foster meaningful peer connections [70]. Third, the digital format may have reduced interaction quality and, therefore, homophily [71,72]. Fourth, some peer mentors were active caregivers themselves while responsible for supporting multiple participants, which may have diluted their capacity to provide consistent or individualized peer support. Fifth, baseline mastery scores were in the midrange, suggesting participants may already possess a strong sense of mastery, limiting the potential for improvement. Sixth, although it is possible that the lack of effect observed in this trial was due to insufficient power, this is unlikely, as post hoc analyses suggested the study was adequately powered to detect a clinically meaningful difference in caregiver mastery.

Intervention fidelity and participant engagement may also have contributed to the absence of measurable effects. Engagement with several intervention components was low, particularly the asynchronous discussion forum and weekly group sessions. Our qualitative feedback suggested barriers such as time constraints and difficulty feeling connected to mentors due to differences in caregiving characteristics [73], indicating that more tailored matching or flexible delivery approaches warrant further exploration. While digital interventions help to address barriers such as social isolation and geographical distance, they may not consistently support meaningful connections [49,74].

A key strength of this study is its rigorous randomized controlled design, which minimizes bias through randomization and allocation concealment. However, as participants and outcome assessors were not blinded, there is potential for performance and detection bias, particularly given the use of self-reported psychosocial outcomes. As no significant between-group differences were observed, any such bias was unlikely to have substantially affected the findings. Additionally, a large proportion of eligible caregivers declined participation, introducing potential selection bias, which may limit the generalizability of our findings. It is possible that those who agreed to participate were more confident or capable in their caregiving roles and thus more willing to engage in peer support. Those who declined may have been the individuals who would have benefited most from the intervention. Allowing participants to choose their peer mentor may have enhanced relational depth and tailored support. However, this was not feasible given our small pool of peer mentors. Most participants were existing caregivers with moderate to high mastery and low distress, suggesting our inclusion criteria could have included measurement of baseline mastery to identify those most in need of the intervention. Although the peer mentor training program was adapted from a previously established caregiver support model and refined through pilot testing with caregivers of individuals with NMD, the original framework was developed for caregivers of children with cancer. While core mentorship principles such as communication skills, boundary setting, and emotional support are transferable, the training may not have fully addressed the unique challenges associated with NMD caregiving, such as progressive functional decline, long-term ventilation support, and complex care coordination [75,76]. For the exploratory component of the analysis, no corrections were made for multiple comparisons or false discovery. This limits the confidence that can be placed in the results of this portion of the study, and future work is necessary to confirm these findings. Lastly, this study did not directly measure stress mediators that social support is theorized to affect by the Stress Process Model, specifically social support, self-efficacy, and competence. This limits our ability to understand why the intervention was not effective within this population. Although the study did capture depression, anxiety, stress, burden, and caregiver sense of mastery, more direct measures would likely have improved our understanding of the direct effects of the intervention.

This randomized controlled trial suggests that digitally delivered peer support alone may be insufficient to improve psychosocial outcomes among caregivers of individuals with NMD, particularly within a heterogeneous population of experienced caregivers. While several factors, including the short intervention duration, potential ceiling effects, lack of homophily, and variability in caregiver engagement, may have contributed to the absence of measurable impact, these explanations remain speculative. Future research should consider focusing on caregivers who are newer to the caregiving role and evaluating longer-duration, individually tailored peer support interventions integrating psychoeducational or skills-based components to better understand their potential to improve caregiver well-being. These findings highlight the challenges of designing scalable digital support interventions that foster meaningful interpersonal connection while addressing the diverse needs of family caregivers.