This study evaluates whether a home-based telerehabilitation program (tele-CPASS) is as effective as an in-person rehabilitation program (CPASS) for improving arm and hand function after stroke. Stroke often leads to long-term difficulty using the affected arm in daily activities, and access to in-person therapy can be limited. This study aims to determine whether therapy delivered remotely can provide similar benefits to standard in-clinic care.

Participants who recently experienced a stroke will be randomly assigned to receive 20 hours of upper extremity rehabilitation therapy either in person or through a telehealth platform. Both groups will receive the same type and amount of therapy focused on practicing meaningful, patient-selected daily activities. Participants will complete assessments before treatment, immediately after treatment, and at 6 and 12 months to measure recovery of arm function, real-world arm use, and participation in daily life.

The results of this study will help determine whether telerehabilitation can improve access to effective stroke recovery interventions while maintaining clinical effectiveness comparable to in-person therapy.

Stroke is a leading cause of long-term disability, with a substantial proportion of survivors experiencing persistent upper extremity impairment that limits independence in daily activities. While rehabilitation can improve outcomes, access to consistent, high-dose therapy remains a major barrier, particularly in the early months after stroke when recovery potential is highest. The Critical Periods After Stroke Study (CPASS) previously demonstrated that 20 hours of intensive, task-specific upper extremity training delivered within the first six months after stroke leads to clinically meaningful improvements in motor function compared to standard care.

Despite this evidence, many patients face challenges accessing in-person outpatient rehabilitation due to transportation limitations, cost, and geographic barriers. Telerehabilitation offers a potential solution by enabling delivery of structured therapy in the home environment. Prior studies suggest that home-based rehabilitation can achieve outcomes comparable to in-clinic care, while increasing accessibility and adherence. However, the efficacy of a telerehabilitation version of CPASS (tele-CPASS) relative to the established in-person CPASS intervention has not been rigorously tested.

This study is a randomized, controlled, parallel-group trial designed to evaluate whether tele-CPASS is comparable in efficacy to in-person CPASS for improving upper extremity recovery after stroke. Participants will be enrolled in the early post-stroke period and randomly assigned to one of two groups: (1) in-person CPASS delivered in a clinical setting, or (2) tele-CPASS delivered remotely in the participant's home using a secure telehealth platform. Both groups will receive the same total dose (20 hours) of therapist-guided, task-specific training over approximately 2-3 weeks. Therapy is individualized and based on patient-prioritized functional goals, emphasizing repetitive practice of meaningful daily activities.

Study procedures include baseline assessments followed by post-intervention and longitudinal follow-up assessments. In addition to standardized clinical evaluations, this study incorporates objective measurement of real-world arm use using wrist-worn accelerometers. These devices capture continuous movement data during daily life, allowing quantification of functional upper extremity use outside the clinical setting. A structured activity protocol performed in the laboratory and at home, combined with video annotation and machine learning methods, will be used to classify functional versus non-functional arm movements and enhance the ecological validity of outcome measurement.

The trial uses a non-inferiority framework to determine whether outcomes achieved with tele-CPASS are not meaningfully worse than those achieved with in-person CPASS. This approach is appropriate given prior evidence supporting the efficacy of the in-person intervention. Demonstrating comparable effectiveness would support tele-CPASS as a scalable and accessible model for delivering high-intensity stroke rehabilitation.

By directly comparing remote and in-person delivery of an evidence-based intervention, this study addresses a critical gap in stroke rehabilitation. If successful, tele-CPASS may expand access to effective therapy, reduce barriers to care, and inform future models of rehabilitation delivery that integrate clinic-based and home-based approaches.

Inclusion Criteria:

• Age 21 years or older
• Neuroimaging-confirmed stroke within 40 days prior to enrollment
• Persistent hemiparesis resulting in impaired upper extremity function, defined by a score of 0 to 3 on the NIH Stroke Scale motor arm item
• Able to participate in all study-related activities, including 1-year follow-up
• Short Blessed Memory Orientation and Concentration Scale score less than 8
• Able to follow 2-step commands
• Montreal Cognitive Assessment (MoCA) score greater than 25
• No upper extremity injury or condition that limited upper extremity use prior to the stroke

Exclusion Criteria:

• Inability to provide informed consent
• Persistent disabling neurologic condition, such as multiple sclerosis, Parkinsonism, amyotrophic lateral sclerosis, or dementia requiring medication
• Clinically significant fluctuations in mental status within 72 hours prior to randomization
• Active psychosis, psychosis within the prior 2 years, active substance abuse, or prior substance abuse
• Dense sensory loss, defined by a score of 3 or 4 on the NIH Stroke Scale sensory item
• Ataxia out of proportion to weakness in the affected arm, defined by a score greater than 1 on the NIH Stroke Scale ataxia item
• Not expected to survive 12 months because of other illnesses
• Pregnancy
• Severe aphasia
• Receipt of botulinum toxin or expectation of receiving botulinum toxin during the study period

Individual participant data will not be shared because this is a small, single-site rehabilitation trial involving individuals recovering from stroke, and the dataset includes potentially identifiable clinical, functional, accelerometry, and video-derived movement data. The protocol includes video recordings of upper extremity activity scripts and therapy sessions, which are used for frame-by-frame annotation and machine-learning classification of functional arm use. Although study data are coded with randomly generated participant IDs and identifiers are stored separately, the combination of clinical characteristics, timing post-stroke, functional outcomes, accelerometry, and video-derived data may increase the risk of deductive identification. Deidentified aggregate results may be shared through publications and presentations.