Reliability and validity of virtual reality box & block test in healthy adults and patients with stroke: a prospective, multi-center, exploratory, cross-sectional study

Background: Virtual reality (VR) has emerged as a promising tool, offering immersive environments, real-time feedback, and kinematic data to support both evaluation and training in rehabilitation. The Box and Block Test (BBT) is a simple yet well-validated clinical tool commonly used to assess upper extremity function. This study aimed to develop and validate two versions of the virtual reality Box & Block Test (VR-BBT) in healthy adults and patients with stroke. Methods: Participants completed the conventional BBT as well as two versions of a VR-BBT developed for this study: a physical interaction version (VR-PI) and a non-physical interaction version (VR-N). Primary outcome measures included the number of blocks transferred in the BBT, VR-PI, and VR-N. In patients with stroke, additional kinematic parameters (e.g., movement speed and distance) from the VR-BBT were analyzed. Results: Twenty-four healthy adults and 24 patients with stroke were enrolled. Patients with stroke scored 55.65 ± 12.04 and 27.04 ± 16.88 in BBT, 28.40 ± 12.44 and 14.19 ± 10.24 in VR-PI, and 30.69 ± 11.29 and 16.21 ± 11.92 in VR-N for the unaffected and affected hands, respectively. The VR-BBT showed strong correlations with the BBT (r = 0.841 for VR-PI, r = 0.827 for VR-N). Intraclass correlation coefficients (ICC) indicated excellent reliability (BBT = 0.982, VR-PI = 0.940, VR-N = 0.943). Correlations with FMA-UE were 0.839, 0.657, and 0.676 for BBT, VR-PI, and VR-N, respectively. The affected hand exhibited statistically significantly lower movement speed than the unaffected hand. Movement distance was greater in the affected hand than in the unaffected hand for VR-N (p = 0.046), with a similar but non-significant trend for VR-PI (p = 0.062). Conclusions: The VR-BBT demonstrated strong reliability and validity despite differences in performance counts compared to BBT. The affected hand showed greater movement distance but lower speed, indicating inefficient motor control. These parameters were significantly associated with FMA-UE, suggesting their potential as objective markers of upper limb motor impairment. These findings suggest that the VR-BBT could serve as a complementary tool for motor function assessment, with potential applications in tele-rehabilitation and virtual reality-based rehabilitation. Trial registration KCT0009584 (Clinical Research Information Service, Republic of Korea).