Robot assisted upper limb therapy combined with upper limb rehabilitation was at least as effective on a range of outcomes, and cost less to deliver, as an equal dose of upper limb rehabilitation alone for people with stroke

Objective: Evaluate effectiveness of group based robotassisted therapy (RAGT) with individual arm therapy (IT) for people with moderate to severely affected upper extremities following stroke. Design: Single-blind randomised trial. PEDro scale: Criteria met for random allocation, allocation concealment, baseline comparability, blind assessors, adequate followup, intention to treat, reporting between-group comparisons, point estimates/variability. Settings: Two in-patient stroke rehabilitation settings in Germany. Participants: Adults with first-occasion stroke (supratentorial) < 8 weeks with nonor minimally functional upper extremity, in an in-patient programme ≥ 6 weeks. Fifty randomised, 23 analysed per group (reasons given); adequate power estimated at N = 40. RAGT n = 25, 13 male, mean age 71.4 years; IT n = 25, 15 male, mean age 69.7 years. Interventions: Both groups participated in comprehensive rehabilitation programmes. In addition IT completed 2 9 30 minutes of individual arm therapy per workday for four weeks which included task-oriented motor re-learning provided by centre-based therapists. The RAGT group completed 30 minutes of individual arm therapy per day plus arm studio group therapy assisted by one supervised therapy assistant for 30 minutes per day. Bi-ManuTrack, Reha-Digit, Reha-Slide, and Reha-Slide duo provided passive mobilisation and repetitive training of isolated shoulder, elbow, wrist and finger movements. Outcome measures: Blinded assessment of Fugl-Meyer Assessment (FMA) upper extremity score (range 0–66; primary outcome) at baseline, post-intervention and three months. Secondary outcomes (un-blinded): Action Research Arm Test, Box and Blocks, Modified Ashworth Scale (MAS), Medical Research Council scale (muscle strength), and Barthel Index (activities of daily living). Costs per treatment were calculated. Main findings: No significant between group differences for any measure at any time point. Within group differences demonstrated significant improvements at each time point on each measure except MAS. Clinically significant mean differences at four weeks on FMA: RAGT +11.1 (SD: 10.6); IT +14.6 (SD: 11.2); three months FMA: RAGT +16.8 (SD: 16.0); IT +20.6 (SD: 14.6). No major adverse events; fingertip blisters in two RAGT participants. 17/24 RAGT participants provided favourable feedback for use as adjunct to individual arm therapy. Costs per treatment session were 4.15€ RAGT and 10.00€ IT. Authors’ conclusions: When combined with individual arm therapy, robot assisted group training or an equal dose of individual arm therapy alone were equally effective to improve upper extremity outcomes. Robot assisted therapy cost less to deliver. Contact details of original author: s.hesse@medicalpark. de