Robot-assisted training after stroke: arm and ADL gains, no clear dose-response link

Background

Robot-assisted training (RAT) is used extensively in rehabilitation after stroke, yet how training dose relates to functional recovery has not been established¹. A systematic review and meta-analysis assessed the efficacy of RAT and explored dose-effect relationships¹.

Methods

Searches covered Embase, PubMed, the Cochrane Library and Web of Science from database inception through October 2025¹. Eligible studies were randomised controlled trials in adults after stroke that compared RAT, or RAT plus conventional rehabilitation (CR), with CR¹. Study quality was judged with Cochrane's tool for assessing risk of bias, and certainty of evidence was rated with the GRADE framework¹. Effects were pooled as mean differences (MDs) or standardised MDs with 95% confidence intervals (CIs)¹.

Five training-dose parameters were examined through exploratory analyses by subgroup together with meta-regression using a random-effects model, with restricted cubic spline analyses testing for nonlinear relationships¹. The review was registered (CRD420251267710)¹.

Results

The analysis included 36 trials with 1,941 patients¹. Compared with CR, RAT led to significant gains in:

• Motor function of the upper limb: MD = 2.89, 95% CI: 1.49-4.28¹
• Activities of daily living (ADL): standardised mean difference = 0.28, 95% CI: 0.13-0.43¹

Balance did not improve significantly (MD = 1.27, 95% CI: -0.92 to 3.47)¹.

Subgroup analyses by dose level showed numerical differences, but none of the between-subgroup comparisons reached significance¹. No dose parameter was linearly associated with Fugl-Meyer Assessment-Upper Extremity (FMA-UE), Berg Balance Scale (BBS) or ADL scores, and spline analyses detected no nonlinear associations (all P > .05)¹.

Certainty of evidence was moderate for FMA-UE and ADL, whereas it was low for BBS¹.

Limitations

The authors advise caution with the dose findings, since several subgroups contained only a small number of studies and the training dose actually delivered was not fully reported¹.

Conclusion

RAT may enhance motor function of the upper limb and ADL following stroke, while no significant benefit for balance has been shown¹. The available evidence does not establish a clear linear or nonlinear link linking the dose of training with rehabilitation results¹.