This qualitative synthesis study follows the ENTREQ (Enhancing Transparency in Qualitative Research Synthesis) (27) and PRISMA (Preferred Reporting Items for Systematic Reviews and Meta Analyses) (28) guidelines. The study protocol has been registered with PROSPERO (Registration Number: CRD42023405581).

The literature search was performed per the Joanna Briggs Institute (JBI) Methodology Guide, following a three-step procedure. Step 1 comprised an initial search of the PubMed and CINAHL (EBSCO) databases, examining terms in titles, abstracts, and subject headings. Step 2 comprised comprehensive searches in PubMed, Cochrane Library, Embase, CINAHL (EBSCO), Web of Science, and PsycINFO (EBSCO) using a combination of subject terms and free-text words. Step 3 involved snowball searches by reviewing the reference lists of included studies to identify additional relevant research. The search period extended from the inception of each database to October 19, 2025. The search focused on qualitative studies exploring stroke patients' experiences with home exercise rehabilitation. Detailed search strategies are presented in Supplementary File S1, Table S1.

The Population, Phenomenon of Interest, Context, and Study Design (PICoS) framework recommended by the Australian JBI Center for Evidence-Based Healthcare established inclusion and exclusion standards per the study's primary focus. Inclusion criteria: (1) Study subjects: Patients with a confirmed diagnosis of ischemic or hemorrhagic stroke, encompassing the acute, recovery, and sequelae phases. (2) Phenomenon of Interest: Stroke patients' subjective experiences during exercise rehabilitation, including key elements such as physical and psychological sensations during training, motivation for adherence, barriers and challenges, needs and expectations, and perceived outcomes. (3) Context: Exercise rehabilitation conducted in authentic home or community settings, including stroke units, rehabilitation institutions, community rehabilitation centers, outpatient clinics, and home environments. (4) Research Design: Original studies employing explicit qualitative methods, including phenomenology, grounded theory, ethnography, or thematic analysis, and mixed-methods studies from which qualitative data can be independently extracted. Exclusion criteria: Non-English literature; studies with inaccessible full texts, incomplete data, or missing critical information; mixed-method studies in which qualitative content is highly confounded with quantitative data and cannot be extracted separately; and duplicate publications.

Two researchers (XX and HX), trained in evidence-based methods and qualitative research, conducted literature screening and data extraction independently. Any discrepancies were resolved through consultation with a third researcher (MZ). The process included the following steps: (1) Duplicate records were removed using EndNote literature management software; (2) The researchers (XX and HX) independently assessed titles and abstracts against PICos criteria to identify documents for full-text review; (3) Full-text examination determined final inclusion.

Two researchers (XX and HX) independently assessed the included studies using the JBI Critical Appraisal Checklist for Qualitative Research (29). The checklist comprises 10 items, each rated as Yes (Y), No (N), or Unclear (U). Studies were included in the meta-synthesis if six or more of the 10 criteria were rated Yes. Any evaluation discrepancies were resolved through discussion with a third researcher (MZ).

Two researchers (XX and HX) independently extracted data using a standardized tool based on the JBI Qualitative Research Systematic Meta Synthesis Design (30). To address the specific requirements of the study on stroke patients'experiences with exercise rehabilitation, they added dimensions such as nursing and rehabilitation professional involvement, scenario type, and post-stroke timeframe. Any discrepancies were resolved by a third researcher (MZ). Extracted content included study basic information (authors, year, country or region, study design), sample characteristics (sample size, mean age, female proportion, post-stroke time window), methodology (primary analytical methods), and core themes (research focus and main themes).

The results of the included studies were meta-synthesized using the System for the Unified Management, Assessment and Review of Information (SUMARI), a web-based systematic review and meta-analysis tool developed by JBI (31), following these steps: (1) One researcher (XX) entered the results, while another researcher (MZ) conducted a review; (2) Two researchers (XX and HX) independently summarized findings into individual statements and assigned credibility ratings of clear, plausible, or lacking support to ensure consistency between citations and findings; (3) Findings rated clear or plausible were categorized by content, while those rated lacking support were excluded from further analysis; (4) Categories were subsequently consolidated into comprehensive findings. All authors reviewed the appropriateness of these findings, and any disagreements were resolved through discussion with a third researcher (MZ). To enhance transparency, the synthesis process followed the standard JBI meta-aggregation hierarchy: primary findings were extracted from the original qualitative data, grouped into conceptual categories based on similarity, and finally integrated into overarching synthesized findings to inform clinical practice.

The ConQual evidence grading method (32) was employed to assess the quality of evidence in the integrated findings. All integrated findings were initially considered high quality and downgraded based on dependability and credibility ratings, ultimately establishing the final evidence quality level.

As shown in Figure 1, the database search identified 7,291 studies. After duplicates were removed, 2,081 studies remained. Following title and abstract screening, 2,003 studies were excluded, leaving 78 for full-text evaluation. Ultimately, 10 studies (16, 33–41) satisfied the inclusion criteria.

The closed-loop model in this study illustrates the continuous interaction between the hospital, community, and home settings in stroke rehabilitation. As shown in Figure 2, the hospital provides initial guidance and discharge planning, which is followed by community support and, finally, home-based self-regulation and feedback. This cyclical process ensures continuous care, with feedback flowing between settings to maintain rehabilitation adherence. The model emphasizes the importance of continuity of care across these stages, facilitating sustained patient engagement and rehabilitation success.

The results of the quality assessment for the included studies are presented in Table 1. All studies demonstrated strong performance in methodological consistency (items 1–5), presentation of participant perspectives (item 8), ethical compliance (item 9), and evidence supporting conclusions (item 10), achieving 100% consistency (10/10). Research methods, data analysis, and interpretation of conclusions exhibited logical coherence; participant perspectives were fully articulated; and all studies received ethical approval.

However, notable deficiencies were observed in the reflective research entries (Entries 6 and 7): Entry 6—whether the researcher' s perspectives and value positions that may influence the study were clarified—showed an uncertainty rate of 50% (5/10); Entry 7—description of the relationship between the researcher and participants and its impact on the study—had an uncertainty rate of 60% (6/10). These findings indicate that, while existing research provides robust empirical evidence on motivational maintenance and continuity of care in home-based rehabilitation, systematic reflection on the researcher's positionality remains limited.

Table 2 presents the basic characteristics of the 10 included qualitative studies, covering 263 stroke survivors from eight countries between 2018 and 2023. The mean age of participants was 63.2 years (±7.5), with 47.8% female. Most participants were in the subacute to chronic phase post-stroke, ranging from 2 months to 22 years after onset. The primary settings were home-based (60%) and community rehabilitation (40%) programs.

Methodologically, 80% of the studies employed thematic analysis to examine the intrinsic construction of meaning and behavioral motivation mechanisms in collective experiences. All studies underwent ethical review and received informed consent. Sample sizes ranged from 10 to 50 participants (mean 26.3), with most studies employing semi-structured interviews or focus group discussions to balance data depth and diversity.

Regarding research content, 90% of the literature (16, 33–37, 39–41) examined maintaining exercise motivation, self-efficacy, and belief reconstruction in patients within home environments. 80% (16, 34–38, 40, 41) highlighted the role of family and social support systems in promoting exercise adherence. 70% (16, 33–35, 37–39) explicitly emphasized the critical importance of nursing continuity, including community nurse follow-ups, rehabilitation guidance, and telephone supervision. Collectively, these studies indicate that, in home settings without hospital supervision, sustained support from healthcare providers and individual regulation of motivation are central elements for promoting rehabilitation adherence.

The meta-aggregation process was used to synthesize findings from the included studies, as shown in Figure 3. This process involves transforming illustrative data (participant quotes) into findings, grouping similar findings into categories, and ultimately synthesizing them into synthesized findings. These steps ensure that data is systematically integrated into coherent, actionable insights. The flowchart provides a clear visual representation of how individual findings were categorized and synthesized to derive overarching conclusions for practice.

Table 3 presents an overview of the comprehensive findings. Across 10 studies, 48 research findings were identified and categorized into 12 and four overarching themes. Among these findings, 36 were rated as clear and 12 as plausible.

Table 4 presents the ConQual summary of the synthesized findings. All ten articles included in this study satisfied the core criteria of the JBI qualitative research quality assessment, including methodological consistency, ethical compliance, and evidence supporting conclusions; therefore, all were retained. For the dependability rating, we followed the method described in Section 2.8.1 by evaluating Items 2, 3, 4, 6, and 7 of the JBI Critical Appraisal Checklist for Qualitative Research. As shown in Table 1, Six studies (34, 36, 38–41) achieved 4–5 “Yes” ratings and received no downgrade, whereas four studies (16, 33, 35, 37) achieved 3 “Yes” ratings and were downgraded by one level, mainly due to unclear reporting on Items 6–7 (reflexivity-related items). Accordingly, the dependability of the synthesized findings was downgraded by one level. Regarding credibility, the synthesized findings were supported by unequivocal evidence, consisting of direct participant quotes and clear illustrations from the original studies that established a logical link between the data and the findings. Therefore, no downgrade was applied to credibility. Consequently, all four synthesized findings received an overall ConQual rating of moderate, which should be considered when interpreting the strength of practice implications.

The transition from hospital care is not a seamless continuation of the medical pathway but a high-risk period of discontinuity. Our findings reveal three forms of discontinuity post-discharge: information, relational, and resource gaps. First, the information gaps identified in this study highlight the pervasive challenges in translating clinical guidelines into practice. Although the European Stroke Organisation (ESO) (42) explicitly recommends providing standardized rehabilitation education and personalized prescriptions before discharge, our synthesis indicates that static education is insufficient to address the complex challenges of the home environment. This qualitative finding complements objective data, observing a rapid decline in adherence among stroke patients (43). Our qualitative evidence further elucidates the underlying psychological mechanism. This decline stems not only from functional limitations but also from safety-related anxiety caused by a lack of real-time guidance. This suggests that existing health education models fail to fill the behavioral adaptation void during the transition from a controlled hospital setting to a complex home environment.

Second, the relational gaps emphasized in this study reflect significant differences between stroke rehabilitation and other chronic disease management models. Unlike the well-established home-based exercise models in cardiac rehabilitation (44), stroke rehabilitation exhibits a higher degree of technical dependence and a critical need for real-time feedback (45). In the absence of continuous follow-up by specialist nurses or physical therapists, stroke survivors may develop learned helplessness and abandon their training programs due to a lack of immediate correction and positive reinforcement (46). This reflects a rupture in professional support during the transition home. The inadequacy of existing primary healthcare systems in providing high-intensity, specialized follow-up remains a systemic cause of suboptimal long-term exercise adherence.

Third, resource gaps expose the global challenge of inequitable medical service distribution. In centralized systems such as China, discontinuity often stems from the concentration of high-quality rehabilitation resources in tertiary hospitals, leaving primary care facilities ill-equipped to provide specialized follow-up (47). Conversely, even in decentralized systems like Denmark, where community-based services are more robust, patients still encounter transition fatigue due to fragmented communication between municipal providers and acute care sectors (48). In the context of this study, this implies that rehabilitation adherence is no longer merely a matter of individual motivation but an ecological issue constrained by the accessibility of systemic resources.

In summary, the discharge transition is a phased, dynamic process involving behavioral adaptation, reconstruction, and integration (16). Based on a study of community stroke patients in Hong Kong, Chau et al. (37) proposed a rhythmic continuity of care model encompassing pre-discharge preparation, discharge-day handover, and post-discharge follow-up, which provides a practical reference for this study. Building on this, this study further proposes that transitional care should be temporally continuous and experientially continuous. Consequently, future discharge support should shift from education to behavioral support, achieving this transformation through remote follow-ups, hybrid care pathways, and home-adaptive technologies.

Beyond systemic gaps, home-based rehabilitation relies on the transition from external compliance to autonomous self-regulation, a process aligned with the COM-B model (49). In the Capability and Opportunity dimensions, our findings suggest that environmental constraints and psychological readiness are barriers to adherence. This qualitative insight complements quantitative data (50) showing that environmental factors significantly predict post-stroke exercise persistence. Unlike studies focusing solely on movement techniques (41), we emphasize life-embedded exercise to alleviate the psychological burden, treating the home environment as an active affordance for recovery. Regarding motivation, home-based training lacks immediate feedback, making patients more susceptible to frustration, pain, and emotional fluctuations (39). This explains the quantitative trend (45) of declining adherence post-discharge. To rebuild the psychological scaffolding, we propose using objective feedback metrics [e.g., Borg RPE or wearable data (51)] to provide the immediate reinforcement previously supplied by clinicians. Therefore, a closed-loop strategy that integrates goal stratification, real-time feedback, and problem-solving follow-ups is essential.

When rehabilitation transitions from institutional settings to the home environment, motor rehabilitation for stroke patients evolves from an individual health behavior to a social practice influenced by family structures, social relationships, and community resource s (35). The integrated findings of this study indicate that a three-tiered support network—comprising family, peers, and community—is essential for patients' sustained engagement. Through three interlinked mechanisms, the network reinforces adherence to rehabilitation. Emotional support acts as the primary driver of rehabilitation motivation. While quantitative evidence frequently highlights the strong correlation between family companionship and recovery outcomes (52, 53), our qualitative synthesis further elucidates the underlying mechanism: family affirmation does not merely provide encouragement but helps patients reconstruct their identity and alleviate the sense of illness and self-doubt. This emotional buffering is particularly critical in stroke recovery to counter the psychological stress of solitary training. Unlike cardiac rehabilitation (54) that emphasizes functional monitoring, stroke rehabilitation requires a higher degree of relational continuity to maintain the patient's intrinsic drive.

Additionally, Structural support provides the practical conditions necessary for sustained rehabilitation implementation. Community-level resources and peer connections trigger long-term psychosocial effects through peer connections, fostering role identification and cohesion. This aligns with quantitative findings suggesting that group-based interventions significantly enhance self-efficacy compared to isolated home exercise (55). Through mutual reminders and shared achievements, patients undergo an identity transformation from passive care-receivers to active rehabilitators (56), which ensures adherence even after professional supervision is withdrawn. Therefore, nursing practice should evolve from singular patient education to supporting the construction of a comprehensive ecosystem. Nurses should instruct patients in training techniques and design supportive structures, such as integrating family members as rehabilitation collaborators and facilitating community peer support networks (36). This semi-supervised, semi-social participation model expands the accessibility of rehabilitation services, transforming recovery from an isolated task into a life practice embedded within relational and management continuity, thereby enhancing the sustainability of interventions.

Effective home-based exercise rehabilitation is unlikely to be achieved through a single intervention alone; rather, it requires establishing a guidance–monitoring–motivation closed-loop system coordinated by nurses and implemented collaboratively across hospitals, communities, and homes.

During the guidance phase, nurses must transition from traditional educators to rehabilitation facilitators. This process necessitates thoroughly considering the patient's home environment and caregiving resources to prevent the disconnect when hospital prescriptions are directly applied to home settings (37). This aligns with recent stroke rehabilitation guidelines (42), which emphasize that personalized, goal-oriented prescriptions are more effective than generic education in preventing the post-discharge service gap.

Regarding monitoring, our integrated findings suggest that minimal datasets of key metrics are crucial for sustaining long-term adherence (41). Specific indicators may include daily training steps/duration, Borg-RPE fatigue scores, etc (51). These metrics are easily understood by patients and families while effectively reflecting training efficacy. Nurses can maintain continuity in rehabilitation services through remote or in-home follow-ups at key time points. Consistent with quantitative evidence on telerehabilitation (57), scheduled remote follow-ups provide a rhythmic structure that reduces patient dropout rates by offering professional presence even in the absence of face-to-face contact.

Regarding motivation, this study highlights that stroke patients' rehabilitation adherence is driven more by value-based factors than by purely medical rationality (39, 40). Therefore, a key strategy is proactively establishing explicit connections between training and life's meaning. For example, interviews can guide patients in recalling family roles they wish to resume post-recovery, linking training to personal value goals (58). Additionally, group-based incentives, such as peer support groups and rehabilitation experience-sharing sessions, can strengthen patients' recovery identity and social reinforcement effects (59).

The Guidance-Monitoring-Motivation model constitutes a replicable, scalable, continuous care pathway framework. This includes nurses developing personalized home rehabilitation prescriptions before discharge, implementing tiered follow-ups, establishing peer support networks, and using patient training adherence and functional improvement as quality feedback indicators for continuous program optimization. Future research should employ implementation science and mixed methods to evaluate this model's effectiveness, scalability, and equity across diverse settings, thus advancing the practice translation of evidence-based care models.

To effectively translate these findings into practice, it is essential to clarify the multidisciplinary synergy between nursing and physical therapy (PT). While stroke rehabilitation in many health systems is primarily led by PTs, who specialize in assessing movement impairments and prescribing task-specific training, our proposed nurse-centered closed-loop model should be interpreted as a coordination framework rather than a replacement. The primary distinction lies in their functional focus: PTs drive the clinical progression of exercise, whereas nurses complement this by addressing the continuity of care. Specifically, nurses bridge the discharge transition by monitoring adherence and safety, facilitating goal reviews, and connecting families to community resources—functions that patients in this synthesis identified as critical gaps.

Guided by this multidisciplinary framework, we propose the following considerations for optimizing home-based exercise rehabilitation. To bridge the care gap identified during the discharge transition, clinicians should prioritize an intensified follow-up schedule, particularly within the first two weeks post-discharge, followed by periodic reassessments at approximately one, three, and six months to address the fluctuating motivation observed in stroke survivors. To provide the professional supervision that patients reported as lacking, multimodal remote feedback, such as telephone calls, text messages, or remote monitoring sessions, should be employed to offer specific guidance on technique and safety. Furthermore, to bolster patient self-efficacy, clinicians should adopt a collaborative goal-setting approach, co-producing achievable short-term goals and reviewing progress through simple indicators (e.g., perceived exertion or step counts). Finally, given the vital role of social support networks, practitioners should explicitly define caregiver roles and facilitate stronger information handover between hospitals and community-based services to ensure that professional support and PT expertise are effectively extended into the home environment.

This study, a meta-synthesis grounded in qualitative research, provides a comprehensive summary of stroke patients' experiences with home-based exercise rehabilitation from the perspective of nursing continuity. However, several limitations require objective consideration. First, the literature demonstrates considerable heterogeneity. The ten included studies spanned eight countries, revealing significant variations in healthcare systems and home care cultures across regions. Although methodological consistency was maintained through category consolidation and thematic extraction during synthesis, these contextual differences may still restrict the generalizability of the findings. Second, the qualitative reports varied in detail. Some studies inadequately described the researchers' data saturation process, making it challenging to assess their methodological rigor fully. Third, this study exclusively included English-language literature, potentially overlooking regional studies in other languages, particularly evidence regarding nursing practices in Latin America and Eastern Europe. Furthermore, the nature of meta-synthesized qualitative research precludes the quantitative measurement of intervention effects, allowing only experiential-level inferences about impact mechanisms. Future studies could integrate mixed-methods approaches and longitudinal tracking to validate the practical efficacy of continuity-of-care interventions in enhancing adherence and functional recovery.