Virtual reality for stroke rehabilitation: characteristics of protocols, pilot and feasibility studies

Introduction: Virtual reality (VR) for stroke rehabilitation is a therapeutic intervention expected to follow the randomized control trials (RCTs) requirements. This study aimed to identify the characteristics of protocols, pilot and feasibility studies reporting stroke rehabilitation with VR methods. Materials and methods: A systematic study was conducted regarding publications reporting on the use of VR for stroke rehabilitation. PubMed, Web of Science, and Institute of Electrical and Electronics Engineers bibliographic databases were searched on March 2019. The keywords were ("stroke" or "stroke rehabilitation" or "neurological rehabilitation") and ("virtual reality" or "virtual reality game" or "computer-aided therapy" or "assisted therapy") and ("quality of life" or "activities of daily living"). All eligible studies published in English were included. The following were collected: experimental design, inclusion criteria for participants, age range, VR intervention, comparative intervention, the primary and secondary outcome. Results: Title and abstract screening stage had 326 studies, 60 entered the full-text screening stage. Five study protocols of RCTs, 1 protocol for feasibility study, 3 pilot studies and 2 feasibility studies were fully evaluated. All articles provided a structured abstract, 7 were registered in a RCT registry. All RCTs were assessor-blinded, with one exception. The upper extremity in adults was the target of the VR rehabilitation in 9/10 cases, only 2 provided the diagnostic criteria. The settings of intervention were community-dwelling (3 papers), hospital (2) or patient's home (1). Data were collected at least twice (pre- and post-treatment). The lack of details on randomization and the VR intervention did not allow for study reproducibility, despite 9/10 papers presenting randomization procedure. Four study protocols provided information regarding the sample size calculation, sample size varying between 26 and 59. Conclusion: Not all VR for stroke interventions were registered in a trial registry, insufficient details were provided regarding randomization and/or VR intervention.