Stroke is one of the most debilitating brain disorders, often leading to death or severe functional impairment. The global burden of stroke is exceptionally high in low- and middle-income countries, where increased incidence, mortality, disability, and economic strain are prevalent [1]. Stroke is the second leading cause of death and disability worldwide [2], with over 116 million healthy years lost annually due to disabilities and fatalities associated with stroke [3]. Among stroke survivors, more than 80% experience motor impairments in the upper extremities (UEs) [4], significantly limiting their ability to engage in social and daily activities. While approximately two-thirds of patients who have experienced a stroke can walk independently 6 months after the stroke, nearly half remain unable to functionally use the affected arm despite receiving treatment [5].

UE impairment is a common challenge for stroke survivors, often leading to significant motor dysfunction and impacting daily living activities [6,7]. Improving UE function through targeted rehabilitation is essential for enabling stroke survivors to perform daily tasks independently [8,9]. Constraint-induced movement therapy (CIMT) is a highly effective intervention for enhancing UE function after stroke [10] and is considered one of the most essential rehabilitation approaches for stroke recovery [11]. Numerous systematic reviews provide strong evidence supporting CIMT’s use across acute, subacute, and chronic phases of stroke rehabilitation [12-15]. Consequently, CIMT is recommended as a core intervention in clinical guidelines worldwide, including in the United Kingdom [16], Australia [16], and Canada [17]. In addition, research indicates that CIMT can enhance functional performance compared to traditional rehabilitation and increase the use of the most affected UE [18].

Despite the prevalence of motor impairments following a stroke, neuromuscular electrical stimulation (NMES) interventions are designed to enhance the quality of life and motor function by contracting weak muscles in the affected extremity. This approach aims to address paresis, restore fractionated movements, improve coordination and joint control, and decrease abnormal movement patterns, such as reduced spasticity and abnormal muscle synergies [19]. Clinical evidence supports the therapeutic efficacy of NMES as a complementary approach to traditional physiotherapy. Specifically, NMES for finger, arm, and wrist extension has shown the most significant clinical and neuroprosthetic benefits. Research indicates that individuals with moderate impairments within 2 years after a stroke experience essential improvements in UE function from NMES, especially when the stimulation is synchronized with the individual’s movement intent [19,20].

Motor function recovery is a critical aspect of stroke rehabilitation. However, in the chronic stage of stroke, the pace of functional recovery often slows considerably. This slow recovery rate, combined with the inherent tendency of humans to rely on bilateral arm coordination for functional activities, may lead to the development of compensatory movement strategies that hinder optimal recovery of the affected UEs. Humans instinctively use both arms in a coordinated fashion to carry out daily tasks. Following a stroke, this natural inclination to rely on bilateral arm coordination, coupled with the slow recovery of the affected arm, may lead patients to develop compensatory movements. While enabling individuals to cope with their impairments, these strategies limit the potential for optimal recovery of the affected UE. As patients become increasingly reliant on compensatory movements, they may fail to engage the impaired limb effectively, leading to a plateau in functional improvement. Consequently, rehabilitation professionals must recognize and address these compensatory strategies to promote optimal recovery and prevent long-term functional limitations in chronic stroke survivors. To overcome the challenges posed by compensatory strategies and promote recovery in the chronic stage of stroke, a combined approach integrating motor rehabilitation techniques, such as CIMT, with sensorimotor stimulation methods, such as NMES, may be beneficial [21]. Combined therapeutic interventions demonstrate significant potential for activating diverse brain regions involved in motor control, enhancing recovery following stroke. The clinical implications of these findings suggest that implementing multimodal interventions is essential for optimizing motor recovery in chronic stroke survivors. This approach is substantiated by functional magnetic resonance imaging studies revealing neuroplastic changes associated with combined therapy modalities, demonstrating enhanced cortical reorganization when peripheral stimulation is paired with intensive motor training [21-23]. The neuroimaging evidence provides a neurophysiological basis for the efficacy of combined interventions, offering valuable insights into evidence-based rehabilitation protocols that may inform clinical practice guidelines and future research directions in UE rehabilitation after stroke.

Although CIMT is an established intervention for stroke recovery, limited research exists on the synergistic effects of combining it with NMES. No clinical studies have reported the effectiveness of CIMT combined with NMES in comparison to CIMT alone or conventional therapy for treating UE dysfunction in patients who have experienced a stroke. Therefore, we aim to combine CIMT with NMES to explore the effect of their combination on UE dysfunction in chronic stroke survivors and compare it to CIMT alone and conventional therapy.

This study is an assessor-blinded, cluster, randomized controlled trial (RCT) comparing 3 intervention groups: 2 experimental groups (CIMT combined with NMES and CIMT alone) and 1 control group (conventional therapy). The trial will be conducted across 3 National Stroke Association of Malaysia (NASAM) rehabilitation centers. NASAM, a primary referral center for stroke survivors in Malaysia, operates 8 centers nationwide, providing rehabilitation services, peer support, and information to stroke survivors.

The study will be reported according to the CONSORT (Consolidated Standards of Reporting Trials) guidelines. The trial schedule is presented in Figure 1, while the trial design flowchart is provided in Figure 2. In addition, the study has been registered prospectively with the Australian New Zealand Clinical Trials Registry under registration number ACTRN12624000564550.

The principal researcher will recruit eligible participants using the predefined inclusion and exclusion criteria, as described in Textbox 1 [24-28].

Participants will be recruited from the NASAM. Three NASAM centers will be randomly assigned to 1 of 3 groups: the CIMT combined with the NMES group, the CIMT-only group, or the control group. Due to the specific nature of the therapies and the distinct characteristics of the centers, randomizing individual participants within centers is impractical. A cluster randomization approach will be used to minimize contamination among patients attending the same center. In addition, to minimize the risk of contamination between intervention groups, a cluster randomization approach will be used, wherein health care centers will serve as the unit of randomization. Each participating center will be randomly allocated to 1 of the 3 intervention arms using a computer-generated sequence administered by research personnel not involved in intervention delivery or outcome assessment. This concealed allocation process ensures allocation concealment and methodological rigor. Following randomization, all eligible patients who have experienced a stroke attending a given center will receive only the intervention protocol assigned to that facility. This methodological approach enables consistent intervention delivery within each site while substantially reducing the likelihood of cross-contamination between intervention groups—a critical consideration when evaluating health care provider–delivered rehabilitation protocols with behavioral components.

Eligible participants will provide written informed consent before their involvement in the study. The principal investigator or a trained research staff member not involved in therapy delivery will be responsible for obtaining consent from each participant. This process will take place only after participants have received a comprehensive explanation of the study procedures, ensuring that they clearly understand their involvement and the associated risks and benefits.

Following the informed consent process, baseline demographic information, including the participants’ names, ages, and gender, will be collected. Once demographic data have been obtained, the remaining baseline assessments will be conducted. Throughout the study, the confidentiality of all participants’ personal information and data will be strictly maintained in accordance with established research ethics guidelines and data protection regulations.

Outcome assessors will be blinded to the group assignments to ensure an unbiased evaluation of the results. However, given the significant differences in the therapies and the nature of the interventions, it is not feasible to blind the therapists or the patients involved in the study. Clinical staff must be aware of the specific therapy they are administering. Therapists will be informed that a study is being conducted on different modalities in other NASAM centers, but no particular details about the treatment regimens or locations will be provided. They will be aware of the treatment regime at their center but are reminded not to disclose this information to other therapists, either within or at different centers.

Recognizing that therapist-related variability represents a significant potential source of bias in rehabilitation trials, we have developed comprehensive standardization procedures to minimize therapist-level confounding. While the complete elimination of therapist effects remains impossible in hands-on interventions, our multifaceted approach ensures maximum consistency across sites and health care providers.

All participating therapists must complete a 3-day training program before commencing study treatments. This standardized curriculum encompasses theoretical foundations underlying each intervention approach, extensive hands-on skill development with practice participants, and a detailed review of all protocol procedures. Competency assessment requires therapists to achieve a minimum score of 90% on written examinations covering theoretical knowledge and protocol specifications and satisfactory performance on practical skills assessments evaluated by master trainers. Weekly refresher sessions throughout the trial will maintain intervention quality, address emerging questions, and ensure continued protocol adherence.

To enhance baseline comparability across sites, participation is restricted to therapists possessing at least 5 years of stroke-specific rehabilitation experience. This requirement ensures foundational clinical expertise while creating greater homogeneity in health care provider characteristics. Each therapist will exclusively deliver 1 intervention type throughout the study, eliminating within-center variability and reducing contamination risk between treatment arms. The same therapist will treat all participants within their assigned center, maintaining consistency in intervention delivery style and therapeutic relationship development.

Structured peer observations using standardized feedback forms will maintain consistent intervention quality across sites. Detailed session plans specifying exercise sequences, progression criteria, and acceptable modifications will limit unwanted variability while preserving necessary clinical flexibility for individual participant needs. This comprehensive standardization approach minimizes therapist-related variability while enabling appropriate statistical control for unavoidable health care provider effects, strengthening our trial’s internal validity and interpretability.

The interventions in this study will take place on 3 consecutive days each week for 4 weeks. The study flowchart (Figure 2) provides a detailed overview of participant flow through the clinical trial.

A blinded assessor unaware of group allocations will conduct baseline and postintervention assessments to prevent bias. The assessor will be trained in standardized assessment techniques and cannot access baseline data during postintervention evaluations.

A single trained assessor will perform all assessments at the respective rehabilitation centers while remaining blinded to group allocation and treatment throughout the study. This assessor will travel between centers to ensure consistent assessment procedures.

All data will be meticulously recorded on case report forms and verified by the researcher. Raw data will be securely stored and accessible to researchers upon request, with appropriate measures to protect patient confidentiality. The principal researcher will be primarily responsible for ensuring the accuracy, integrity, and timeliness of the data entry process. To protect participant privacy, all files will be secured with passwords and only accessible to authorized research team members.

We used the G*Power software (version 3.0.10; University of Düsseldorf) to calculate the required sample size. A mixed model ANOVA, a statistical method used to analyze the effects of interventions over time and between groups, was selected to analyze the interventions’ time, group, and interaction effects. The F test (ANOVA repeated measures within and between interactions), a specific type of statistical test used in the mixed model ANOVA, was selected for this analysis. With a study power of 80%, an α level of .05, and an effect size of 0.25 (based on the primary outcomes from a previous CIMT study [47]), a minimum sample size of 42 participants (14 in each group) is required. If a group has more than 14 eligible participants, the first 14 to provide consent will be recruited based on the referral date. This first-come, first-served approach ensures transparency in participant selection while maintaining cluster integrity.

An intention-to-treat analysis, which includes all randomized participants, regardless of their adherence to the protocol or completion of the study, will be implemented for this RCT. Under this approach, all recruited participants will be included in the final analysis regardless of protocol adherence, withdrawal, or discontinuation. For participants who withdraw or discontinue the treatment, their baseline data will be used as exit data in the primary analysis. This conservative approach acknowledges the potential impact of missing data while maintaining randomization integrity. Sensitivity analyses using multiple imputation techniques, which are statistical methods used to handle missing data, will be conducted to assess the robustness of findings against different missing data assumptions. This analytical strategy aligns with CONSORT guidelines for reporting randomized trials and addresses potential selection bias that might occur with complete-case analysis.

Data will be analyzed using IBM SPSS Statistics software (version 29.0), adhering to the intention-to-treat principle, which includes all randomized participants regardless of protocol adherence. Participant characteristics and baseline measures will be summarized using descriptive statistics, including mean, SD, median, and IQR.

Before primary analyses, data will be assessed for normality using the Shapiro-Wilk test and homogeneity of variance using the Levene test. For outcomes meeting parametric assumptions, a 2-way mixed model ANOVA will be used to compare the effects of the interventions’ interactions between and within groups. Appropriate transformations or nonparametric alternatives will be applied in cases of violated parametric assumptions. Potential confounding variables, such as baseline motor function and stroke severity, will be included as covariates to provide adjusted treatment effect estimates. Statistical significance will be set at P<.05 for all analyses.

The linear mixed-effects models that include the center as a fixed effect covariate and random intercepts at the participant level will be used to address residual center-specific effects without invoking a design effect. This modeling framework accommodates hierarchical data structures in single-cluster designs while preserving valid inference for group-by-time interactions.

Missing data will be addressed using multiple imputation techniques, with sensitivity analyses comparing results to those obtained using the baseline observation carried forward method. This comprehensive approach ensures a robust interpretation of intervention effects while preserving statistical power.

This study will be conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. Ethics approval has been granted by the Research and Ethics Committee of the Universiti Kebangsaan Malaysia (RECUKM; the National University of Malaysia; study ID UKM PPI/111/8/JEP-2023-596). All participants will provide informed consent before participating, and their privacy will be kept confidential.

The study findings will be disseminated through publication in peer-reviewed journals and presentations at national and international congresses. A summary of the trial results will also be made available on the Australian New Zealand Clinical Trials Registry database. All necessary regulatory approvals for the study have been obtained. Upon trial completion, participants will receive personalized feedback regarding their involvement and the study outcomes.

This study evaluates the effects of combining CIMT with NMES on UE function among chronic stroke survivors. We hypothesize that the combined CIMT+NMES intervention will produce superior improvements in motor function and perceived functional use compared to CIMT alone or conventional therapy. This dual approach leverages complementary mechanisms: CIMT addresses learned nonuse through intensive task practice [48], while the NMES facilitates muscle activation and enhances proprioceptive feedback to support voluntary movement.

We examined current rehabilitation paradigms for chronic stroke to contextualize our approach within existing evidence. CIMT has established efficacy for improving UE function in selected populations of individuals who have experienced a stroke through forced use and mass practice tasks [49]. Similarly, NMES demonstrates clear benefits for muscle strengthening and motor unit recruitment in paretic limbs [50]. However, a few studies have investigated their combined therapeutic potential [10]. Recent evidence indicates that multimodal rehabilitation approaches may amplify neuroplastic changes through synergistic cortical reorganization mechanisms. Functional magnetic resonance imaging studies have documented enhanced cortical activation patterns when peripheral electrical stimulation accompanies intensive motor training, suggesting accelerated neural adaptation compared to single interventions [21].

This study advances this emerging field by providing the first systematic comparison of combined CIMT+NMES therapy against established monotherapies. This investigation addresses a critical gap in stroke rehabilitation evidence, particularly for chronic stroke survivors who typically experience limited recovery with conventional approaches. By examining whether combined interventions overcome the ceiling effects often observed with single modalities, our findings will inform evidence-based treatment selection for this challenging population.

This trial incorporates several methodological strengths that enhance scientific rigor and clinical applicability. Blinded outcome assessors evaluate all participants using validated instruments spanning the International Classification of Functioning, Disability, and Health framework, ensuring a comprehensive assessment of motor impairment, activity limitations, and participation restrictions. Our standardized intervention protocols, supported by detailed treatment manuals and competency-based therapist training, promote consistent implementation across sites. The pragmatic trial design, conducted within existing rehabilitation centers using available clinical resources, enhances external validity and facilitates future implementation.

We acknowledge specific limitations inherent in our study design. While pragmatically necessary given resource constraints, the single-cluster-per-arm randomization strategy prevents traditional design effect calculations in our sample size calculation. We have addressed this limitation through conservative effect size selection and comprehensive statistical modeling that includes the center as a fixed effect covariate. Our mixed ANOVA approach protects against type I error inflation while maintaining adequate sensitivity for detecting clinically meaningful differences. The absence of long-term follow-up assessments limits our understanding of treatment durability; however, immediate postintervention outcomes will establish the foundation for future longitudinal investigations.

Despite rigorous standardization procedures, therapist-delivered interventions inevitably introduce some variability. We minimize this concern through mandatory competency assessments and structured fidelity monitoring. These mitigation strategies ensure that observed treatment differences reflect intervention efficacy rather than health care provider variability.

Our findings will directly inform clinical decision-making regarding optimal rehabilitation strategies for chronic stroke survivors with persistent UE deficits. Should the combined intervention demonstrate superiority, clinicians will gain evidence supporting integrated treatment protocols that maximize functional recovery potential. The comprehensive outcome battery, including motor function (FMA-UE), activity performance (Wolf Motor Function Test), perceived use (Motor Activity Log), and self-efficacy (UPSET), provides multidimensional insights into treatment effects across the International Classification of Functioning, Disability, and Health domains.

Future research should expand upon our foundation through several critical avenues. Large-scale multicenter trials incorporating cluster randomization with appropriate design effects will establish definitive efficacy estimates. Extended follow-up periods of 6 to 12 months will determine whether initial gains translate into sustained functional improvements. Neurophysiological assessments, including transcranial magnetic stimulation and electroencephalography, could elucidate the cortical mechanisms underlying potential synergistic effects [51].

We have developed a comprehensive dissemination strategy to ensure rapid translation of findings into clinical practice. Primary dissemination will occur through publication in high-impact rehabilitation journals with open-access options to maximize accessibility. Conference presentations at international forums will facilitate direct engagement with clinical and research communities.

To accelerate clinical implementation, we will develop practitioner-oriented resources, including evidence summaries, clinical decision algorithms, and video training modules demonstrating proper technique. Collaboration with professional organizations will support integration into clinical practice guidelines and continuing education curricula. Patient-centered knowledge translation activities include creating accessible educational materials explaining treatment options and expected outcomes, available in multiple languages to serve diverse populations. Partnerships with stroke survivor organizations will ensure broad distribution through established communication channels, empowering patients to advocate for evidence-based care.

This protocol addresses a critical gap in stroke rehabilitation evidence through rigorous evaluation of combined therapeutic approaches, integrating CIMT with NMES to target both voluntary movement control and direct muscle activation through complementary neuroplastic mechanisms. By systematically examining whether multimodal interventions produce synergistic effects exceeding those of single treatments, our study advances rehabilitation science through a comprehensive methodological framework that incorporates standardized protocols, quantitative fidelity monitoring, and cluster-appropriate statistical modeling, ensuring reliable evidence generation for clinical practice guidelines. Beyond establishing comparative efficacy, this research will elucidate optimal dosing parameters and practical implementation strategies for complex rehabilitation interventions, enabling clinicians to make informed decisions about resource allocation and treatment selection for chronic stroke survivors while contributing essential evidence toward developing personalized, evidence-based rehabilitation strategies that maximize functional recovery in this underserved population.