Robot-assisted upper-limb training,
does it really help with Superior arm-function recovery after stroke compared with usual rehabilitation?
research showsThe claim that robot-assisted training produces better post-stroke arm recovery than usual rehabilitation is rated D with 30 points. Assessor-masked RATULS randomized 770 participants and included 669 in the three-month primary analysis, leaving 101 missing outcomes (13.1%). Action Research Arm Test success was 44% versus 42%, adjusted odds ratio 1.17 (98.3% CI 0.70 to 1.96), so the primary outcome failed. Positive pooled findings from smaller trials may reflect small-study and publication effects, so the large independent confirmatory trial takes priority.
ads claimThe factual ability to deliver more repetitive movements should not be expanded into superior everyday functional recovery. RATULS also found no superiority over enhanced upper-limb therapy.
Useful facts when choosing a product
- The MIT-Manus system in RATULS used shoulder-elbow, wrist, and hand modules to provide repeated movements.
- The robotic group was assigned 45-minute sessions three times weekly for 12 weeks, totaling 36 sessions. Delivery of repetitions is factual; superior recovery is a separate clinical question.
- The 770 randomized participants must be distinguished from the 669 with primary three-month outcome data.
What the research actually shows
Assessor-masked RATULS randomized 770 participants and included 669 in the three-month primary analysis, leaving 101 missing outcomes (13.1%). Action Research Arm Test success was 44% versus 42%, adjusted odds ratio 1.17 (98.3% CI 0.70 to 1.96), so the primary outcome failed. The 2018 Mehrholz Cochrane review found arm-function SMD 0.32 (95% CI 0.18 to 0.46), I-squared 36%, P<0.0001 across 41 trials and 1,452 participants, and activities-of-daily-living SMD 0.31 (95% CI 0.09 to 0.52), I-squared 59% across 24 trials and 957 participants. When pooled small trials and a large independent trial differ, the confirmatory trial takes priority rather than being labeled repeated conflict. Verdict 1879 is confirmed as C with 45 points because virtual reality lacks a comparable large confirmatory trial and its updated pooled evidence varies; robotics has RATULS, and RATULS failed.
Why this is classified as D (30)
The patient-function primary outcome failed in large, publicly funded RATULS, which takes priority over pooled small-study signals. No major bias defect is counted, but the upper confidence limit of 1.96 still allows a large benefit, giving D with 30 points.
Counterpoint. The system may help deliver additional practice to selected patients. This verdict addresses claimed clinical superiority over usual or purpose-matched rehabilitation.
Rejudgment record. Cross-check applied — Priority given to the failed patient-function primary outcome in large publicly funded RATULS, with earlier small-study signals plausibly reflecting small-study and publication effects and a confidence interval that still permits benefit
| Endpoint | P | Patient-reported treatment goal - the symptom is the goal |
| Replication | R1 | Single confirmatory trial |
| Independence | I2 | Decisive evidence is publicly or non-profit funded |
| Effect size | E0 | Null |
| Precision | C0 | The confidence interval leaves room for benefit |
The scoring table and the verdict agree (D).
Sub-claim grades by effect
This ingredient is marketed for several effects. A single overall grade blends strong and weak claims together, so each effect is graded separately here. The overall grade reflects the strongest disconfirming or core claim.
| Effect (sub-claim) | Grade | Basis |
|---|---|---|
| Increased three-month arm-function success versus usual care | D | Action Research Arm Test success was 44% versus 42%, adjusted odds ratio 1.17, and failed. |
| Superior arm function versus enhanced upper-limb therapy | D | The adjusted odds ratio was 0.78 and did not show superiority. |
| Improved activities of daily living | C | An earlier small-trial meta-analysis found SMD 0.31, but this was not aligned with the large trial. |
Cross-check — Codex and Claude
Evidence Table
| Study | Design | Sample | Funding | Endpoint | Result | Weight |
|---|---|---|---|---|---|---|
| Rodgers H et al. 2019 (RATULS) | Multicenter pragmatic randomized assessor-masked three-group trial | 669 | Public funding from the United Kingdom NIHR Health Technology Assessment Programme | Three-month Action Research Arm Test upper-limb function success | The primary outcome failed: 44% versus 42%, adjusted odds ratio 1.17 (98.3% CI 0.70 to 1.96). | Decisive large independent null trial |
| Mehrholz J et al. 2018 | Cochrane systematic review and meta-analysis | 957 | Academic Cochrane review; funding varied across included trials | Activities of daily living, arm function, and strength | Arm-function SMD 0.32 (95% CI 0.18 to 0.46), I-squared 36%, P<0.0001; activities-of-daily-living SMD 0.31 (95% CI 0.09 to 0.52), I-squared 59%. | Earlier positive pooled signal from small trials |
Receipt — 2 References
All 2 cited sources were verified for existence at the original page (as of 2026-07-24).
Reviewed and approved: Chamgap Editorial Team · Approval date: 2026-07-24 · Corrections: none
Cite this verdict
[Chamgap] Robot-assisted upper-limb training x superior arm-function recovery after stroke — Evidence Grade D·30. 2 cited sources checked. Source: https://chamgap.com/en/verdicts/cognition/robot-assisted-upper-limb-training-stroke-functional-recovery/ · CC BY 4.0CC BY 4.0 — free to use with attribution; do not distort grades, numbers, or verdict meaning.
What this document does and does not do
Chamgap is an information source. It reports what research has and has not confirmed; it does not tell readers what to take or buy. That decision belongs to readers and, when needed, medical or legal professionals. This verdict reflects literature available up to the search date and may change as new research appears. Nothing here is medical advice.