- Design
- Systematic review and meta-analysis of 36 randomised trials
- Population
- 1,941 adults after stroke
- Primary outcome
- Upper-limb motor function, balance and activities of daily living
- Effect
- Fugl-Meyer UE MD 2.89 (1.49–4.28); ADL SMD 0.28 (0.13–0.43); balance not improved
A meta-analysis in Medicine pooled 36 randomised trials (1,941 patients) of robot-assisted training, alone or added to conventional rehabilitation, after stroke.
Robot-assisted training improved upper-limb Fugl-Meyer scores by 2.89 points and activities of daily living modestly (standardised mean difference 0.28). Balance did not improve. Analysis of five dose measures — session length, frequency, total hours and others — found no relationship between dose and benefit. Certainty was moderate for arm function and daily activities.
The arm benefit is small: a 2.9-point Fugl-Meyer gain is below what many consider clinically important. Robots allow more repetitions with fewer therapists, which may matter where staff are short, but they do not replace task-specific practice.
Where robots are available, they are a reasonable adjunct for arm rehabilitation. They are not a reason to cut conventional therapy time.
- Consider robot-assisted training as an adjunct for upper-limb recovery after stroke where available
- Do not use robotic devices in place of balance and gait training
- Keep conventional, task-specific therapy as the core of rehabilitation
- Do not prescribe extra robot hours expecting more benefit; no dose–response was found
Why it matters
It tempers the expectation that more robot time means more recovery.
The statistics, in plain English
A standardised mean difference of 0.28 is a small effect. The balance result's interval (−0.92 to 3.47) includes no effect.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for neurology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free