Periodised power training in community-dwelling adults with chronic stroke: a protocol for the Power Exercise for Stroke Recovery (POWER) pilot randomised controlled trial

Abstract

Introduction

Resistance training (RT) can improve muscle strength in people with stroke. However, muscle power may be more important than strength to preserve mobility after stroke. Power training (ie, RT with the intent to move as fast as possible) has shown to be feasible in people with stroke but trials to date were limited in their approaches to exercise prescription, namely providing minimal emphasis on periodisation strategies (eg, developing strength before power). We developed a periodised power training intervention, Power Exercise for Stroke Recovery (POWER), and recently showed that POWER was feasible and could improve muscle power, mobility and health-related quality of life in people with stroke. We propose the next stage in investigation—a pilot multisite randomised controlled trial to test the feasibility of the trial design before embarking on a fully scaled trial.

Methods and analysis

This study will be a pilot multisite randomised controlled trial conducted at McMaster University and the University of British Columbia in Canada. 60 people ≥6 months following mild to moderate stroke will be randomly allocated to one of two, thrice-weekly, 10-week RT interventions: (1) periodised power training intervention (POWER) or (2) conventional RT intervention following current physical activity recommendations (STRENGTH). Both training programmes begin with a 1-week familiarisation period consisting of low-intensity versions of the programmes. The POWER group will begin with an intense strength-building phase (weeks 2–5), followed by a power training phase (weeks 6–10). The STRENGTH group will engage in a moderate-intensity strength-building programme from weeks 2–10. The primary outcome of this trial is feasibility of the process (eg, randomisation, blinding), resources (eg, assessor burden) and scientific feasibility (eg, adverse events), using a priori criteria for acceptance. Clinical outcomes (primary: comfortable and fast Timed Up and Go Tests) and questionnaires will be assessed at baseline, post-intervention and 8-week follow-up.

Ethics and dissemination

This trial has received ethical approval at both institutions (Hamilton Integrated Research Ethics Board #18235, University of British Columbia Clinical Research Ethics Board #H25-00325). Trial results will be published and made available to local stroke recovery groups, clinicians and researchers in our network.

Trial registration number

NCT06780995.

Strengths and limitations of this study.

• This trial will include prospectively selected feasibility indicators with a priori success criteria that will inform the design of a fully scaled trial.

• As a pilot trial, the sample size calculation was based on feasibility outcomes, which may be insufficient to detect changes in clinical outcome measures.

Introduction

Stroke is the third leading cause of death and the leading cause of neurological disability in the world,¹ and commonly leads to reduced mobility that is partially driven by skeletal muscle adaptations.² People with stroke can experience a loss of muscle cross-sectional area,³ muscle fibre atrophy² and changes in muscle architecture and neural drive,⁴ often resulting in muscle weakness. The collective reduction in muscle mass, strength and function after stroke is referred to as post-stroke sarcopenia,⁵ which impacts up to 50% of individuals with stroke.⁶

Resistance training (RT) is a common form of exercise-based rehabilitation and is recommended in physical activity guidelines to improve muscular fitness in people with stroke.⁷ Systematic reviews have unanimously agreed about the benefits of RT on muscle strength after stroke^8–10 but with mixed results on mobility outcomes,^{8 10 11} limiting conclusions that can be drawn about the efficacy of RT on stroke recovery. Research is emerging that supports power training, a type of RT that is performed with the intent to move as fast as possible,¹² in people with stroke. Given its superiority over traditional RT in older adults,¹³ power training is now supported in position statements^{14 15} and exercise guidelines.¹² Similarly, in people with stroke, power training may also be superior to traditional RT for improving fast walking speed,⁹ though the evidence is sparse. Power training is likely to be functionally relevant after stroke, since people with stroke experience atrophy of type II muscle fibres,^{2 16} which are responsible for high force-generating contractions needed for functional movements such as stair climbing, standing from a chair and fast walking. Indeed, lab-based muscle power assessments are related to mobility outcomes in older adults¹⁷ and people with stroke,¹⁸ indicating that it is an important therapeutic target for exercise interventions.

Despite the potential relevance of muscle power and power training for individuals with stroke, power training is not yet recommended in clinical practice guidelines as research to date has been limited. Of the available evidence in people with stroke,^19–32 power training studies have reported high attendance rates and no adverse events, but variable effects on upper and lower extremity strength, physical function and health-related quality of life. Notably, none of these power training programmes involved periodisation strategies, which may be important ingredients to exercise prescription.¹⁵ Indeed, building a strong foundation of muscle mass and strength is a prerequisite to developing muscle power,^{33 34} but the previous studies started with endurance walking or low-intensity skill acquisition or progressed immediately into power training protocols at lower workloads. It may be that individuals with stroke require a foundation of muscle strength prior to engaging in high-velocity movements.

To address this gap in the literature, we designed the Power Exercise for Stroke Recovery (POWER) programme.³⁵ POWER is a 10-week high-intensity periodised RT programme with goals to improve muscle strength followed by muscle power to improve mobility. The initial pre-post feasibility study of POWER was designed to understand the intervention’s feasibility and its preliminary effects on health and mobility outcomes in 15 people who were >6 months post-stroke. We reported high attendance rates (96%) and no serious adverse events or unsafe blood pressure responses across 2301 exercise sets.³⁵ POWER also resulted in moderate to large effects (Cohen’s D: 0.55 to 1.53) on muscle strength, power and clinical measures such as the Timed Up and Go (TUG) test, fast-paced walking speed, and 30-Second Chair Stand Test (30sCST).³⁵

The POWER-Feasibility trial demonstrated that certain design elements (eg, intervention, recruitment sources) can be done, with some modifications (eg, intensity monitoring processes).³⁵ However, the trial did not implement other processes to be undertaken by the fully scaled trial (eg, randomisation, a comparator arm, inclusion of additional sites), highlighting the need for a randomised pilot trial—ie, a subtype of feasibility study that reflects the design features of the anticipated future randomised study, but on a smaller scale.³⁶ Thus, following the POWER-Feasibility study, the next step for POWER is to embark on a randomised pilot trial (POWER-Pilot) to test the feasibility of multisite randomised trial design before embarking on a fully scaled trial. The research questions for the POWER-Pilot trial are twofold:

1. What is the feasibility of a multi-site randomised design to evaluate POWER in community-dwelling people ≥6 months following mild to moderate stroke? Hypothesis: We expect the POWER-Pilot trial will demonstrate sufficient feasibility for a multisite randomised design using prespecified process, resource and scientific indicators, which will support a future fully scaled efficacy trial.

2. What are the preliminary estimates of the effect of POWER compared with conventionally recommended RT (Strength Training Engaging Guidelines To improve Health (STRENGTH)) in community-dwelling people ≥6 months following mild to moderate stroke on clinical outcomes of mobility, post-stroke fatigue, psychological well-being, cognition and health-related quality of life? Hypothesis: We hypothesise that clinical outcomes will favour POWER compared with STRENGTH, recognising that the clinical comparisons are exploratory in a pilot trial.

Methods and analysis

This protocol was reported according to the 2025 Standard Protocol Items: Recommendations for Interventional Trials checklist and the pilot and feasibility trials extension of the 2010 Consolidated Standards on the Reporting of Randomized Trials checklist. The trial was registered prospectively on ClinicalTrials.gov (NCT06780995). The trial is expected to recruit from September 2025 to July 2027, with trial completion estimated for December 2027.

Patient and public involvement

The trial team is comprised of eight academic members, two people with lived experience of stroke and two community exercise programme coordinators. We used the International Association for Public Participation Spectrum of Public Participation³⁷ and Contributor Role Taxonomy to describe the engagement levels of all team members (online supplemental appendix 1).

Trial design

This study is an assessor-blinded multisite pilot randomised controlled trial with aims to reflect the design features of the anticipated fully scaled trial.³⁶ After baseline assessment, participants will be randomised 1:1 to either the POWER or STRENGTH group, stratified by study site with variable block sizes of two to six. The sequence will be generated by an independent statistician who will be blind to the study objectives and outcomes. The sequence will be uploaded by the statistician onto an electronic data collection platform (Research Electronic Data Capture, REDCap, Vanderbilt University, Nashville, Tennessee, USA) to preserve allocation concealment. Within 2 weeks of randomisation, participants will begin their respective 10-week exercise programmes. Clinical outcomes will be measured at baseline (week 0, T0), within 1 week following the final exercise session (week 11, T1) and at the follow-up time point, approximately 8 weeks later (week 19, T2). The full trial duration will last approximately 20 weeks. The participant schedule is shown in figure 1.

Figure 1. Participant timeline: schedule of enrolment, interventions and assessments.

Study setting and ethics

This trial will be conducted at McMaster University (lead site) in Hamilton, Ontario, Canada, and at The University of British Columbia in Kelowna, British Columbia, Canada. Ethical approvals have been obtained for both sites (Hamilton Integrated Research Ethics Board #18235, Clinical Research Ethics Board #H25-00325), and all protocol amendments will be handled by the respective ethics boards. All participants will be screened over the telephone and provide informed, written consent in person by study staff. All participant data, including assessment results, will be confidential and stored on a cloud-based server at the lead site.

Eligibility criteria

Participants will be eligible for the trial if they are (1) ≥19 years old, (2) ≥6 months post-stroke, (3) able to walk >10 m independently (with or without the use of an assistive device), (4) living in the community and (5) have mild to moderate stroke severity, as determined by the telephone version of the modified Rankin Scale (<4).³⁸ Individuals will be excluded if they have (1) any contraindications to exercise for people with cardiovascular disease according to the American College of Sports Medicine,¹² such as unstable angina, uncontrolled hypertension or uncontrolled arrhythmias; (2) difficulty following instructions or (3) are actively engaged in stroke rehabilitation services, defined as a personalised exercise programme for a set duration of time, which is supervised and prescribed by a rehabilitation professional.

Interventions

The POWER and STRENGTH interventions were reported in accordance with the Template for Intervention Description and Replication checklist (online supplemental appendix 2). Both interventions involve structured upper and lower limb resistance exercises, three times per week for 10 weeks in accessible exercise gyms designed for older adults, persons with disability and/or chronic conditions. To limit the risk of contamination, the POWER and STRENGTH sessions will be performed at different times of day with a time buffer between sessions. Each session will begin with an 8–10-minute warm-up using varied modalities, including a recumbent cycle ergometer (eg, Life Fitness 95RI, Rosemont, Illinois, USA), arm crank ergometer (eg, Monark 871E Rehab Trainer, Monark Sports and Medical, Vansbro, Sweden) or recumbent stepper (eg, NuStep T4r, NuStep LLC, Ann Arbor, Michigan, USA). Each session will be 60–90 min in length, and both programmes will be matched for the number of exercises per session. Both programmes will also gradually progress in resistance or movement complexity (eg, bilateral to unilateral) to meet the prescribed intensity. An exercise training manual, showing example exercises, progressions and regressions for both programmes, was developed for trainers to ensure treatment fidelity. A brief summary of exercises is shown in table 1.

Table 1. Basic exercises included in the POWER and STRENGTH interventions.

	Day 1	Day 2	Day 3
POWER Programme
Phase 1 Familiarisation (Week 1) Tempo: 2:1:2:1	Leg press	Chest press	Leg press
	Knee extension	Lat pulldown	Knee extension
	Knee flexion	Seated row	Knee flexion
	Sit to stand	Shoulder flexion	Shoulder press
	Calf raises	Shoulder extension	Calf raises
Phase 2 Strength (Weeks 2–5) Tempo: 2:1:2:1	Leg press	Chest press	Leg press
	Knee extension	Lat pulldown	Split squats
	Knee flexion	Seated row	Chest press
	Sit to stand	Shoulder flexion	Step ups
	Calf raises	Shoulder extension	Shoulder press
	Dorsiflexion	Bicep curl	Hip extension
	Hip abduction	Triceps extension	Calf raises
Phase 3 Power (Weeks 6–10) Tempo: 2:0:1:1	Leg press	Chest press	Leg press
	Knee extension	Lat pulldown	Forward lunges
	Knee flexion	Lateral raises	Shoulder press
	Sit to stand	Shoulder flexion	SL glute bridge
	Foot drops	Shoulder extension	Knee raises
	Lateral lunges	Bicep curl	Walk on toes
	Marching	Shoulder adduction	Pull-down
STRENGTH Programme
Phase 1 Familiarisation (Week 1) Tempo: 2:1:2:1	Leg press	Chest press	Leg press
	Knee extension	Lat pulldown	Knee extension
	Knee flexion	Seated row	Knee flexion
	Sit to stand	Shoulder flexion	Shoulder press
	Calf raises	Shoulder extension	Calf raises
Phase 2 Strength A (Weeks 2–5) Tempo: 2:1:2:1	Leg press	Chest press	Leg press
	Knee extension	Lat pulldown	Glute bridge
	Knee flexion	Seated row	Chest press
	Sit to stand	Shoulder flexion	Hip flexion
	Calf raises	Shoulder extension	Shoulder press
	Dorsiflexion	Bicep curl	Hip extension
	Hip abduction	Triceps extension	Calf raises
Phase 3 Strength B (Weeks 6–10) Tempo: 2:1:2:1	Leg press	Chest press	Leg press
	Knee extension	Lat pulldown	Glute bridge
	Knee flexion	Seated row	Chest press
	Squat	Shoulder abduction	Hip flexion
	Calf raises	Shoulder adduction	Shoulder press
	Dorsiflexion	Bicep curl	Hip extension
	Hip abduction	Triceps extension	Calf raises

The interventions will be led by study staff who were trained by investigator (KN). All staff will have experience in exercise testing and prescription in clinical populations and persons with disability. Sessions will also be supported by volunteers and students to ensure at least one trainer to two participant ratios. New staff and supporting volunteers will attend a 3-hour training session to ensure consistent intervention delivery between sites and among trainers.

POWER

The POWER intervention³⁵ involves three progressive phases. Phase 1 Familiarisation (week 1) will allow participants to begin with a low-intensity version of the exercises to acclimate participants to the movements, using body weight resistance or light weights at Borg CR-10 rating of perceived exertion (RPE) 2–3 for 8–10 repetitions. Phase 2 Strength (weeks 2–5) will progress to achieve volitional fatigue between six and eight repetitions (RPE 7–9). Phase 3 Power (weeks 6–10) will use loads to elicit an RPE 4–6 for 15–20 repetitions and execute each exercise at the highest possible velocity to focus on muscle power development. The exercises were selected to target relevant muscle groups and replicate functional movements relevant to daily activities.^39–41

STRENGTH

The STRENGTH programme consists of two phases: phase 1 Familiarisation (week 1) will be the same as the POWER programme, involving a low-intensity version of the training programme; phase 2 Strength (weeks 2–10) will focus on progressively building muscle strength according to best practice recommendations.⁴² Exercises will be organised into sets of 10–15 repetitions at a moderate intensity (RPE 4–5) throughout the entire programme.

The intensity progressions for the POWER and STRENGTH interventions are shown in figure 2A and B, respectively. Participants in the POWER-Feasibility trial noted difficulties interpreting traditional RPE scales due to stroke-specific impairments that confounded their perceived exertion rating (eg, coordination and balance vs weight achieved).³⁵ Therefore, we will use a modified version of the Borg CR-10 RPE scale,⁴³ which asks: “On this scale from 0 to 10, how heavy was that set?”. The RPE scale will be used alongside instructor-perceived and participant-perceived repetitions in reserve (RIR) to monitor set intensity, which is the number of remaining repetitions the participant could have completed before reaching failure. The instructors will estimate the participant’s RIR from their perspective (i-RIR) and participant-perceived RIR (p-RIR) will be determined by asking participants, “How many more reps could you have done at the same weight, speed, and with good form?”. Intensity monitoring and exercise prescription will be based on all three metrics: RPE, i-RIR and p-RIR. We will consider intensity to be ‘within range’ if both participant and instructor ratings are within the prescribed range. For example, if RPE and i-RIR match the prescribed intensity, no change in intensity will be applied. However, if only RPE and p-RIR match the prescribed intensity, a change in load or movement complexity will be applied to ensure i-RIR is within range to consider both participant and instructor perspectives. Theorised conversions^{44 45} between RPE, RIR and % 1-repetition maximum are described in table 2.

Figure 2. POWER and STRENGTH programme progression approaches.

Table 2. Theorised intensity conversion table.

Modified Borg CR-10	Nominal interpretation	Corresponding 1-RM	Corresponding RIR
0	Rest
1	Very light	15%
2	Light	30%
3	Moderate	40–45%
4	Somewhat heavy		4–6 repetitions
5	Heavy	60–65%
6
7	Very heavy	75–80%	3 repetitions
8			2 repetitions
9	Very very heavy		1 repetition
10	Heaviest I could lift	100%	0 repetitions

RIR, repetitions in reserve; 1-RM, 1-repetition maximum.

Measures

A detailed account of the feasibility indicators and their a priori thresholds for success or revision and clinical outcomes are shown in table 3.

Table 3. List of outcomes.

Feasibility indicators
Domain	Outcome	Description
Process	Randomisation	100% participants allocated as assigned, Yes/No Balance of prognosis: no clinically important differences between groups at baseline, Yes/No
	Allocation concealment	Sequence remained concealed to study team, Yes/No
	Assessor blinding	Assessors unblinded for <5% of participants, Yes/No
	Contamination	No occurrences of contamination between POWER and STRENGTH groups, Yes/No
	Recruitment rate	Mean ≥1 participant/month, per site, Yes/No
	Retention rate	Complete follow-up data on ≥80% of participants, Yes/No
Resource	Participant and assessor burden	≥80% completed assessments are ≤2 hours in duration, Yes/No Median participant burden ratings ≤3 points on 10-point scale, Yes/No Median assessor burden ratings ≤3 points on 10-point scale, Yes/No
Scientific	Safety	No serious adverse events attributable to the interventions or assessment procedures, Event rate
	Missing data	≥90% available data points at post-intervention (ie, n=54/60), Yes/No

Clinical outcomes assessed T0 (0 weeks), T1 (10 weeks) and T2 (18 weeks)
Domain	Performance-based measures and questionnaires
Mobility	Timed Up and Go Test, fast and comfortable 10-Meter Walk Test, fast and comfortable
Functional strength	30-Second Chair Stand Test
Balance and balance confidence	Berg Balance Scale Brief-Balance Evaluation Systems Test Activities-specific Balance Confidence scale
Post-stroke fatigue	Fatigue Severity Scale
Emotional well-being	General Health Questionnaire-28
Health-related quality of life	EuroQoL 5D-5L Stroke Impact Scale, V.3.0
Cognition	Montreal Cognitive Assessment
Exercise barrier self-efficacy	Exercise Self-Efficacy Scale
Quality of participation	Measure of Experiential Aspects of Participation
Perceived intervention benefit	Global Rating of Change Questionnaire

Demographic information

All demographic information will be collected at baseline (T0). We will collect data on age, sex, gender identity, education level, race/ethnicity, self-reported general health status (item 1 from the 36-item Short Form Survey [SF-36]⁴⁶) and number/type of other chronic conditions. We will also ask participants about the type of stroke (ie, ischaemic and/or haemorrhagic), their most affected side and number of strokes they experienced, as well as their list of medications.

Feasibility outcomes

The primary outcome of the trial is feasibility. In brief, we will examine process (randomisation, allocation concealment, blinding, contamination, recruitment and retention), resource (participant and assessor burden) and scientific (safety, availability of clinical outcomes) feasibility indicators⁴⁷ based on consultations with the research team and prior data.³⁵

Clinical outcomes

The effects of the interventions will be examined using several performance-based measures and questionnaires, assessed at T0, T1 and T2, by a trained assessor blinded to group allocation.

Timed Up and Go Test

The primary clinical outcomes will be mobility, using the comfortable and fast TUG Test (seconds).⁴⁸ The TUG has demonstrated good to excellent inter-rater reliability and concurrent validity, is predictive of falls and is also a recommended outcome measure for research and clinical practice in people with stroke.^49–51 A standard chair (~46 cm in height) will be placed at the start of a flat walking course, and a cone will be placed 3 m from the chair. Participants will stand from the chair, walk 3 m, turn back towards the chair and sit back down, without assistance. The task will be timed and measured in seconds. Participants will be provided with one practice trial to familiarise themselves with the course prior to the assessment, followed by two trials at a comfortable pace (comfortable TUG) and two trials at fastest speed (fast TUG). The average times will be used for analysis.

Walking speed

We will use the 10-Meter Walk Test to measure comfortable and fast-paced walking speed (metres/second). Four cones will be placed on a flat non-slippery surface at 0, 2, 8 and 10 m. Participants will be asked to walk from end to end at their comfortable speed and as fast as they can safely. We will measure the time in the middle 6 m (cones at metres 2 and 8) to account for the acceleration and deceleration phases. The average walking speed across two trials will be used for analysis.

Functional strength

The 30sCST⁵² is a measure of lower extremity strength and endurance. A ~43 cm chair with a straight back will be placed against the wall. Participants will cross their arms and keep feet flat on the floor. With their back straight, they will stand up and sit down as many times as possible in 30 s. Only one trial is performed. The total number of repetitions completed is the primary outcome of this test.

Balance and balance confidence

We will measure balance using the Berg Balance Scale (BBS)^{53 54} and the Brief-Balance Evaluation Systems Test (Brief-BEST).⁵⁵ The BBS is a performance-based balance measure using 14 tasks: sitting to standing, standing unsupported, sitting unsupported, standing to sitting, transfers, standing with eyes closed, standing with feet together, reaching forward, retrieving an object from the floor, turning to look behind, turning 360 degrees, placing alternate foot on a stool, standing with one foot in front, standing on one leg. Each task is scored on a 5-point ordinal scale, ranging from 0 (lowest level of balance) to 4 (highest level of balance). The highest possible score is 56. The Brief-BEST⁵⁵ is another balance assessment involving eight tasks measuring hip/trunk lateral strength, forward reaching, single leg standing, compensatory lateral stepping, eyes closed standing on foam surface and the TUG test. Each task is scored on a 4-point ordinal scale, ranging from 0 (lowest level of balance) to 3 (highest level of balance). The highest possible score is 24. We will measure balance confidence using the activities-specific balance confidence (ABC) scale.⁵⁶ The ABC is composed of 16 items to assess confidence in maintaining balance when performing everyday tasks and is rated on an 11-point scale ranging from 0% (no confidence) to 100% (completely confident) in 10% increments for each task. Values across the 16 items are averaged to obtain a total score. The highest possible score is 100%.

Fatigue

We will use the 9-item Fatigue Severity Scale to assess fatigue symptom levels over the past week. Participants are asked to indicate on a 7-point Likert scale whether they agree or disagree with statements about their fatigue levels, where 1 indicates strongly disagree and 7 indicates strongly agree. The highest possible score is 63, with higher scores indicating greater fatigue severity.

Emotional well-being

We will measure emotional well-being using the General Health Questionnaire-28 (GHQ-28).⁵⁷ The GHQ-28 is a 28-item questionnaire that evaluates somatic symptoms, anxiety/insomnia, social dysfunction and severe depression. Participants will be asked to indicate how they feel in relation to each question, using positively worded or negatively worded response options, such as (1) better than usual, (2) same as usual, (3) worse than usual or (4) much worse than usual. The GHQ-28 is scored in two ways. The traditional scoring method treats each item binarily, where responses 1 and 2 will be attributed a score of 0, and responses 3 and 4 will be attributed a score of 1. The alternative scoring method uses a Likert scoring method, where the items are scored from 0 (best emotional well-being) to 3 (worst emotional well-being). In both methods, scores will be summed for a total score (traditional method max score: 28, alternative method max score: 84). Higher scores indicate lower emotional well-being.

Health-related quality of life

The EuroQoL 5D-5L (EQ-5D-5L)⁵⁸ and the Stroke Impact Scale, V.3.0 (SIS-3.0)⁵⁹ will be used to measure health-related quality of life. The EQ-5D-5L contains five items: mobility, self-care, usual activities, pain/discomfort and anxiety/depression. Participants will indicate whether they experience any problems with each of these items with five response options, where 1 indicates no problems and 5 indicates extreme problems. We will analyse the EQ-5D-5L using the time trade-off derived value set for Canada,⁶⁰ where higher values indicate greater quality of life. The SIS-3.0 consists of 59 items that measure eight different health domains: strength, hand function, (instrumental) activities of daily living, mobility, communication, emotion, cognition and participation. Items are scored on an ordinal scale, ranging from 1 to 5, where 1 and 5 indicate the lowest and highest scores for each item, respectively. Domain scores will be calculated, which range from 0 (no recovery) to 100 (full recovery).

Cognitive function

We will assess global cognitive function using the Montreal Cognitive Assessment. Participants will be asked to complete a series of cognitive tasks, which assess visuospatial abilities, executive function, short-term memory, delayed recall, phonemic fluency, verbal abstraction, attention, concentration, working memory, language and orientation.⁶¹ Each task is individually scored and summed for a total score, with a maximum of 30 points. Higher scores indicate greater cognitive function.

Self-efficacy for exercise

We will measure exercise barrier self-efficacy using the 10-item Exercise Self-Efficacy Scale,⁶² where participants rate their confidence in exercising under certain conditions on a 4-point scale (1=Not always true, 4=Always true). Item scores are summed to obtain a composite score, with a highest possible score of 40. Higher scores indicate greater exercise self-efficacy.

Quality of participation in exercise

We will use the Measure of Experiential Aspects of Participation⁶³ to measure quality of exercise participation. This measure assesses the degree to which an individual experiences full and effective participation in exercise, across six elements: autonomy, mastery, belongingness, engagement, challenge and meaning. Each element contains two items (ie, 12 total items), which are averaged to generate an element score. Higher scores indicate greater quality of participation.

Perceived responses to interventions

The Global Rating of Change questionnaire, adapted from Kamper and colleagues,⁶⁴ will be used to examine participants’ perceived responses to the interventions in each of the domains listed above (eg, mobility, balance). Participants will be asked: “With respect to the following areas, how would you describe yourself now compared to when you started the study?” Participants will rate their change on a scale ranging from 1 to 5, where 1 indicates much worse and 5 indicates much better compared with baseline (T0).

Sample size

The current pilot investigation is not designed to estimate the efficacy of POWER. We used the following equation⁶⁵ to power our study based on anticipated retention rates from the POWER-Feasibility study: $n=p\left(1-p\right){\left(\frac{Z_{1-\alpha }}{ME}\right)}^2$, where p is the assumed proportion, Z_1-α is the value corresponding to the corresponding CI (ie, 1.96 for a 95% CI) and ME is the margin of error (10%). Based on 93% retention rate,³⁵ we obtained a total sample size of n=49 (ie, n=25 per intervention arm). Conservatively, we will recruit 60 participants into this study (ie, n=30 per intervention arm).

Recruitment

Potential participants will be recruited in four active streams: the outpatient stroke rehabilitation units at the Regional Rehabilitation Centre (Hamilton, ON) and Kelowna General Hospital (Kelowna, BC), local community stroke groups and networks, private community rehabilitation clinics and a list of individuals who have previously participated in our studies and have consented to be contacted for future studies. We will also passively recruit participants via social media outlets (eg, Facebook, LinkedIn), posters on local community posterboards and both campuses to maximise reach.

Statistical analysis

Feasibility outcomes, such as randomisation, allocation concealment, assessor blinding, contamination, recruitment and retention rates, participant and assessor burden, safety and missing data, will be treated as dichotomous variables, quantified as either meeting or not meeting the set feasibility criteria. Raw data (eg, counts, rates, frequencies) will also be presented. All feasibility data will be presented for the entire sample, and by sex and gender identity. For efficacy analyses, linear mixed-effects models will be conducted to model the within-subjects and between-groups effects on our clinical outcomes and questionnaires. The independent variables of interest are time (three levels: baseline, post-intervention, follow-up) and group allocation (two levels: POWER and STRENGTH). An interaction term between time × group will be included in the model, regardless of statistical significance. Random slopes and intercepts will be considered in the analyses. Models will be compared with the Bayesian Information Criteria and likelihood ratio tests. Pairwise comparisons will be examined using a Sidak correction for multiple comparisons. The stratification variable (recruitment site, two levels: Hamilton and Kelowna) and sex⁶⁶ (two levels: male and female) will be entered into the model as covariates. Unadjusted and adjusted will be presented. All analyses will be performed on StataSE (V.17.0, College Station, Texas, USA).

Data management and monitoring

Study outcome data will be stored on REDCap (Vanderbilt University, Nashville, Tennessee, USA), accessible only by the senior author (AT) and lead coordinators (KN, HF). Outcome assessors will have access to participant data at their respective sites. There are no planned interim analyses for clinical outcome measures, which will be analysed at trial completion. Trial-related adverse events will be discussed with the main trial team, including the senior author (AT), site coordinators and leads (KN, BMS, HF), patient partners (BE, KP) and exercise training staff, when relevant.

Ethics and dissemination

This study has received ethical approval from the Hamilton Integrated Research Ethics Board (HiREB #18235) and the University of British Columbia Clinical Research Ethics Board (CREB #H25-00325) and was prospectively registered on ClinicalTrials.gov (NCT06780995).

After the follow-up (T2) assessment, participants will receive the results of their assessments and the details of their allocated exercise programme (eg, list of exercises). The results of the trial will be disseminated to stroke recovery groups, clinicians and researchers in our network to garner interest in the full-scale trial. Data for the trial will be made available upon reasonable request.