A pragmatic, decentralized trial of the home-based InTandem neurorehabilitation system: analyses of engagement, safety, and effectiveness from the OrcHESTRAS trial

The OrcHESTRAS trial demonstrates that the autonomous, home-based InTandem neurorehabilitation system achieved high engagement and significantly improved walking endurance and functional mobility in adults with chronic stroke-related gait impairment, with greater usage correlating to better outcomes.

The Gist

Imagine you've had a stroke. For many people, the road to recovery doesn't end when they leave the hospital. The hardest part is often the long, lonely stretch of "chronic" recovery that happens at home, months or even years later. Walking might feel slow, unsteady, and exhausting.

This paper is about a new tool called InTandem that tries to solve a major problem: How do you get people to keep doing their physical therapy when no one is watching?

Here is the story of the study, explained simply.

🎵 The Problem: The "Gym Membership" of the Brain

Think of traditional physical therapy like a personal trainer. You go to the clinic, and a trainer counts your steps, claps a rhythm, and pushes you to walk faster. It works great, but it's expensive, hard to get to, and you can only do it a few times a week.

Once you go home, you have to be your own trainer. Most people stop because it's boring, tiring, or they just forget. It's like buying a gym membership and never going back.

🤖 The Solution: The "Smart DJ" for Your Walk

The researchers created InTandem, a system that acts like a smart DJ for your brain and feet.

• How it works: You wear special sensors on your shoes. The system listens to how you walk and plays music that matches your steps.
• The Magic: It uses a trick called "rhythmic entrainment." Just like you can't help but tap your foot to a catchy beat, your brain starts to sync your walking to the music.
• The Autopilot: The best part? It does this automatically. If you start to stumble, the music slows down to keep you safe. If you get stronger, the music speeds up to challenge you. You don't need a therapist in the room; the system is your coach, your music player, and your safety net all in one.

🏠 The Experiment: The "Home Workout" Challenge

The researchers wanted to see if this "smart DJ" could work in the real world, without a doctor watching. They invited 204 adults with chronic stroke to use InTandem at home for 12 weeks.

The Rules:

• Put on the shoes.
• Play the music.
• Walk for 30 minutes, 3 times a week.
• Do this for 3 months.

📊 What They Found: The Results

1. Did People Stick With It? (Engagement)

This was the big question. Would people actually use it?

• The Result: Yes! About 82% of the people used the system for at least a month.
• The Analogy: Imagine if you gave 100 people a new video game console and told them to play for 3 months. If 82 of them were still playing after a month, that's a huge win. It proves that people want to use this tool when it's easy and fun.
• The "Super Users": About 58% of the participants were "super users," sticking with it for over 9 weeks.

2. Did It Help Them Walk Better? (Effectiveness)

They measured two things:

• Endurance (The 6-Minute Walk Test): How far can you walk in 6 minutes without stopping?
• Agility (The Timed Up and Go): How fast can you stand up, walk a few steps, turn around, and sit back down?

The Results:

• Distance: On average, people walked 26 meters (about 85 feet) farther in 6 minutes. That's like walking an extra lap around a small park block!
• Speed: They got up and moved about 1.5 seconds faster. In the world of walking with a stroke, that's a massive improvement.
• The "More is Better" Rule: The study found a clear dose-response. The more weeks you used the system, the more your walking improved. It's like a savings account: every week you "deposit" effort, your walking "balance" grows.

3. Was It Safe?

Since these participants had strokes, safety was the top priority.

• Falls: They tracked every fall. Most falls were mild and happened because of the stroke itself, not the device.
• The Verdict: The device was very safe. In fact, the fall rate was lower than what is usually seen in people with chronic strokes. The only serious issue was two people with a history of epilepsy who had seizures, but the researchers noted this was likely due to their pre-existing condition, not the music itself.

💡 The Big Takeaway

This study is like a proof-of-concept that technology can bring high-quality therapy into your living room.

• Before: You needed a therapist to guide your music and steps.
• Now: You can have a "smart DJ" in your pocket that guides you, keeps you safe, and helps you get stronger, all while you watch TV or walk around your neighborhood.

The Bottom Line:
If you have a stroke and struggle to walk, this study suggests that using a music-based, shoe-sensor system at home is not only safe but actually helps you walk farther and faster. The key is consistency: the more you use it, the better you get. It turns the lonely work of recovery into a rhythmic, manageable, and effective daily habit.

Technical Summary

1. Problem Statement

Chronic stroke-related gait impairment is a leading cause of long-term disability, resulting in slow, asymmetric, and metabolically costly walking. While clinical guidelines recommend continued rehabilitation, many survivors face barriers to outpatient therapy, including cost, transportation, and limited skilled care.

• The Gap: Existing evidence-based interventions, such as Rhythmic Auditory Stimulation (RAS), rely on in-person delivery by skilled clinicians, limiting scalability and accessibility.
• The Challenge: While home-based neurorehabilitation offers a solution, patient engagement (sustained adherence) remains a critical determinant of real-world effectiveness. Previous studies show adherence rates often drop to 50–75%, even in coached programs.
• The Objective: To evaluate the feasibility, safety, and effectiveness of InTandem, an autonomous, home-based neurorehabilitation system that delivers individualized, progressive, music-based RAS without clinician oversight.

2. Methodology

Study Design:

• Type: Pragmatic, decentralized, single-arm, longitudinal clinical trial (OrcHESTRAS study).
• Population: 232 enrolled adults (≥6 months post-stroke) with residual gait deficits. 204 initiated the intervention.
• Inclusion Criteria: Age 18–85, able to ambulate without physical assistance (assistive devices allowed), capable of walking ≥0.4 m/s.
• Exclusion Criteria: Non-reciprocal gait patterns, significant comorbid neurologic conditions (e.g., Parkinson's), >2 falls in the prior month, hearing impairments, or medical instability.

Intervention (InTandem System):

• Mechanism: Uses shoe-worn inertial sensors and decision-making algorithms to deliver music-based RAS. The system leverages auditory-motor entrainment, synchronizing movement to rhythmic cues.
• Adaptivity: The system operates in a closed-loop, adjusting musical tempo and beat salience in real-time based on the user's gait quality and stability. It ensures progression only when stable entrainment is demonstrated.
• Protocol: Participants utilized the system at home for 12 weeks, targeting 30-minute sessions, 3 times per week.

Data Collection & Metrics:

• Primary Endpoint (Engagement): Proportion of participants achieving "moderate-to-high" engagement (defined as ≥4 weeks of use). Engagement was categorized by weeks, sessions, and cumulative minutes.
• Secondary Endpoints (Clinical):
  • 6-Minute Walk Test (6MWT): Measuring walking endurance.
  • Timed Up and Go (TUG): Measuring functional mobility.
• Safety: Monitoring of Adverse Events (AEs), Serious Adverse Events (SAEs), and falls.
• Statistical Analysis: Intention-to-treat framework using linear mixed-effects models (adjusting for age, sex, time since stroke, and assistive device use). Engagement was analyzed as both a continuous variable and categorical subgroups (Low: ≤3 weeks; Moderate/High: ≥4 weeks).

3. Key Contributions

• Autonomous Delivery: Demonstrates that a complex, sensor-driven neurorehabilitation intervention can be successfully deployed autonomously in the home without direct clinician supervision.
• Engagement Benchmarking: Establishes a pragmatic benchmark for home-based rehabilitation, showing that the majority of chronic stroke survivors can sustain engagement with an unsupervised digital therapy.
• Dose-Response Evidence: Provides quantitative evidence linking the duration of engagement directly to clinical outcomes, validating the "dose" of the intervention.
• Safety Profile: Offers robust safety data for a large-scale, decentralized trial involving over 100,000 minutes of autonomous gait training.

4. Results

A. Engagement (Primary Endpoint)

• Success Rate: 81.9% (167/204) of participants achieved moderate-to-high engagement (≥4 weeks), significantly exceeding the pre-defined benchmark of 60% ($p < 0.001$).
• Breakdown:
  • High Engagement: 58.1% (≥9 weeks).
  • Moderate Engagement: 23.9% (4–8 weeks).
  • Low Engagement: 18.1% (≤3 weeks).
• Session/Minute Metrics: 67.6% completed ≥12 sessions; 55.4% completed ≥360 minutes.

B. Clinical Effectiveness

• Walking Endurance (6MWT): Significant improvement from baseline to 12 weeks (+26.1 ± 5.6 m; $p < 0.001$).
  • Dose-Response: Each additional week of engagement predicted a 5.82 m greater gain in 6MWT distance ($p < 0.005$).
  • Subgroup Analysis: The Moderate-to-High engagement group improved by +31.6 m, while the Low engagement group showed a non-significant decline (-27.0 m). The between-group difference (~58.6 m) exceeded the Minimal Clinically Important Difference (MCID) of 34.4 m.
• Functional Mobility (TUG): Significant improvement in time (-1.45 ± 0.31 s; $p < 0.001$).
  • Engagement did not significantly moderate TUG improvements in the continuous model, but the Moderate-to-High group showed significant improvement (-1.39 s) compared to the Low group.

C. Safety

• Adverse Events: During the 12-week intervention, 137 AEs were reported. Only 12.4% were Serious AEs (SAEs).
• Relatedness: 88.2% of SAEs were unrelated to the intervention. Two SAEs (seizures in a participant with a history of epilepsy) were "possibly related."
• Falls: 27 falls were reported. 85.2% were unrelated to the device. Only 4 falls (14.8%) were possibly or definitely related. The fall rate (<14%) was lower than typical rates for community-dwelling chronic stroke survivors.

5. Significance and Conclusion

• Scalability: The study proves that InTandem is a scalable solution for chronic stroke rehabilitation, overcoming the "access gap" of traditional RAS therapy.
• Clinical Impact: The intervention produced clinically meaningful improvements in walking endurance, particularly for those with higher engagement. The observed gains (approx. 26m average, >40m predicted for high engagement) translate to improved community ambulation capacity (e.g., navigating parking lots, grocery stores).
• Target Population: While baseline walking speed did not statistically moderate the overall effect, the most significant and clinically meaningful gains were observed in "Limited Community Ambulators" (<0.8 m/s), suggesting the system is highly effective for those with greater initial impairment.
• Future Standard of Care: By transforming a well-established but traditionally unscalable therapy into a safe, home-delivered, autonomous system, InTandem offers a pathway to a new standard of care that addresses both access and engagement barriers in post-stroke rehabilitation.

Trial Registration: ClinicalTrials.gov NCT06051539.