Performing the Action Research Arm Test in Immersive Virtual Reality to Assess Post-Stroke Upper-Limb Activity

Burton, Quentin [UCL] Lejeune, Thierry [UCL] Dehem, Stéphanie [UCL] Lebrun, Noémie [UCL] Ajana, Khawla [UCL] Edwards, Martin [UCL] Everard, Gauthier [UCL]

Background: To plan treatment and measure patients’ recovery, frequent and time-bounded functional assessments are recommended after a stroke. With the increasing needs for neurorehabilitation, new technological technics such as virtual reality have emerged. To answer these needs, we developed an immersive virtual reality version of the Action Research Arm Test (ARAT-VR). Objective: This study aimed to assess the validity, test-retest reliability and usability of the ARAT-VR among individuals with stroke, healthcare professionals and healthy control subjects (HCS). Methods: Thirty individuals with stroke, ten healthcare professionals and 25 HCS were recruited. Content validity was assessed by asking healthcare professionals to compare the difficulty of the movements performed during the ARAT-VR in comparison to the classical Action Research Arm Test (ARAT). Concurrent validity was measured through correlation between the ARAT-VR and ARAT scores. A third of individuals with stroke and HCS were reseen in a convenient delay to measure test-retest reliability. Usability was measured thanks to the System Usability Scale. Results: On average, healthcare professionals estimated that virtual tasks movements’ difficulty was comparable to those of the classical ARAT expect for the items involving marbles. When assessing paretic hands, strong correlations were found (r = 0.84) between classical ARAT and ARAT-VR scores. Test-retest reliability (ICC = 0.99; p < 0.001) and usability (SUS = 82.5 [75 - 90]) were found to be excellent. Conclusion: The ARAT-VR is a valid, reliable and usable tool holding potential to increase assessment’s frequency and remote evaluation among individuals with stroke.