Effect of High-Speed Robot-Assisted Gait Training on Gait Speed in Ambulatory Patients with Stroke who Have Moderate-to-Severe Hemiplegia: A Preliminary Multiple Single-Case Experimental Design

Abstract

Background

Robot-assisted gait training (RAGT) is a promising intervention for patients with stroke. However, its effectiveness in improving gait speed in ambulatory patients with stroke who have moderate-to-severe motor paralysis remains unclear.

Objective

To determine whether high-speed RAGT, which facilitates faster gait training by increasing step length via swing assistance to the paretic lower limb, improves gait speed more effectively than conventional treadmill training in ambulatory patients with stroke who have moderate-to-severe hemiplegia.

Methods

Five subacute patients with stroke (Functional Ambulation Categories ≥ 2; Fugl–Meyer Assessment lower limb motor score ≤ 22) participated in a multiple AB single-case experimental design. Period A involved conventional treadmill training, while Period B incorporated high-speed RAGT using the Welwalk WW-1000, which combines a treadmill and a robotic device attached to the paretic lower limb. In high-speed RAGT, maximum swing assistance was applied to increase step length and facilitate faster gait. Each participant completed two 5-min training sessions across eight days per period. The primary outcomes were comfortable and maximum gait speeds.

Results

No training-related adverse events occurred. During training, high-speed RAGT enabled gait speeds approximately 1.6 times higher than conventional treadmill training. Tau-U effect sizes ranged from −0.16 to 0.98 for comfortable gait speed and 0.08–0.97 for maximum gait speed. Meta-analysis showed Tau values of 0.67 (comfortable) and 0.77 (maximum), indicating large changes.

Conclusions

This preliminary study suggested that high-speed RAGT is feasible and may improve gait speed. Given the study's nonrandomized, unblinded design and small sample size, larger randomized trials are warranted.