Clinical evaluation of a non-immersive virtual environment in stroke rehabilitation

Background and purpose: We describe our attempts to evaluate the effectiveness of a virtual environment developed to rehabilitate stroke patients in the task of making a hot drink. Methods: Single case studies were performed in 13/138 (9%) stroke patients undergoing rehabilitation in a UK stroke unit. Participants in AB/BA (n = 5) and ABA (n = 2) design studies received 5 one-hour sessions of attention control training (A phase) and 5 one-hour sessions of virtual environment training (B phase). An AB design with random duration of A and B phases (minimum duration of A and B phases 3 and 5 days respectively, with total duration of 3 weeks) was used in 6 participants. Results: Visual inspection of scores across all cases showed a trend towards improvement over time in both real and virtual hot drink making ability in both control and intervention phases. There was no significant difference (Wilcoxon, p > 0.05) in the improvements in real and virtual hot drink making ability during all control and intervention phases in the 13 cases. Ceiling effects limited the evaluation of effectiveness in 5 of the 8 cases in which daily performance measures were used. Conclusions: Few people in this setting were suitable for this intervention. The case studies showed no evidence of a strong effect of this intervention, but we had great difficulty in performing single case studies. We conclude that more testing and development of this system is required before it is subjected to rigorous testing of clinical effectiveness.