AbstractBackgroundRobot-assisted therapy can increase therapy dose after stroke, which is often considered insufficient in clinical practice and after discharge, especially with respect to hand function. Thus far, there has been a focus on rather complex systems that require therapist supervision. To better exploit the potential of robot-assisted therapy, we propose a platform designed for minimal therapist supervision, and present the preliminary evaluation of its immediate usability, one of the main and frequently neglected challenges for real-world application. Such an approach could help increase therapy dose by allowing the training of multiple patients in parallel by a single therapist, as well as independent training in the clinic or at home.MethodsWe implemented design changes on a hand rehabilitation robot, considering aspects relevant to enabling minimally-supervised therapy, such as new physical/graphical user interfaces and two functional therapy exercises to train hand motor coordination, somatosensation and memory. Ten participants with chronic stroke assessed the usability of the platform and reported the perceived workload during a single therapy session with minimal supervision. The ability to independently use the platform was evaluated with a checklist.ResultsParticipants were able to independently perform the therapy session after a short familiarization period, requiring assistance in only 13.46(7.69-19.23)% of the tasks. They assigned good-to-excellent scores on the SUS to the user-interface and the exercises (85.00(75.63-86.88) and 73.75(63.13-83.75) out of 100, respectively). Nine participants stated that they would use the platform frequently. Perceived workloads lay within desired workload bands. Object grasping with simultaneous control of forearm pronosupination and stiffness discrimination were identified as the most difficult tasks.DiscussionOur findings demonstrate that a robot-assisted therapy device can be rendered safely and intuitively usable upon first exposure with minimal supervision through compliance with usability and perceived workload requirements. The preliminary usability evaluation identified usability challenges that should be solved to allow real-world minimally-supervised use. Such a platformcould complement conventional therapy, allowing to provide increased dose with the available resources, and establish a continuum of care that progressively increases therapy lead of the patient from the clinic to the home.
Minimally Supervised Categorization of Text with Metadata
