Embodied robots in rehabilitation-oriented interventions for older adults with disability or cognitive impairment: a scoping review.

Only 11 studies met the strict rehabilitation-oriented criteria, and they are heterogeneous in design, population, robot type, dose, comparator, and outcome measures. No formal risk-of-bias or certainty-of-evidence appraisal was performed; the authors state findings should be read as an evidence map, not a comparative effectiveness assessment. Four of five physical rehabilitation studies were conducted in Chinese hospital settings, so service models, staffing patterns, devices, and outcome reporting may not generalize to other health systems. Safety, adverse events, adherence, withdrawals, and intervention dose were inconsistently reported; the authors stress that absence of safety reporting does not mean no adverse events occurred. Engineering databases, trial registries, and grey literature were not searched, so early feasibility studies, conference-only reports, and registry-only protocols may be missing. Intervention periods ranged from 4 to 12 weeks where reported, but dose-response interpretation was not possible due to heterogeneity in robot type, clinical target, session frequency, comparator, and outcome measures.

The authors declared that no financial support was received for the work or its publication.

Do not read this review as evidence that rehabilitation robots work for older adults. It is a scoping review that maps what studies exist; it performed no formal quality appraisal, included only 11 heterogeneous studies, and the authors explicitly state the findings 'preclude conclusions about comparative effectiveness, long-term impact, or clinical scalability.' The 'short-term signals' noted are a description of the landscape, not a treatment recommendation.