The experimental group received the VirTele program. VirTele is an 8-week home rehabilitation program that includes Jintronix exergames [31] for UE rehabilitation and the Reacts app (Technologies innovatrices d’imagerie and Reacts) [35] to conduct videoconference sessions with clinicians. The experimental group received the VirTele equipment at home which included a computer, a Kinect camera, the Reacts app, the Jintronix software, and a USB internet key (if needed). Before starting the intervention, participants, including clinicians and survivors of chronic stroke, received a 1-hour training session to familiarize themselves with the Jintronix exergames and the Reacts app [32].

The Jintronix exergames included 5 games for UE training (Space Race, Fish Frenzy, Pop Clap, Catch and Carry an apple, and Kitchen clean-up) [31]. The clinician adjusted the difficulty parameters of each game remotely (eg, speed, duration, number of repetitions, and direction of the trajectory) according to the participant’s preference and functional abilities. An automated log system of the participant’s performance during exergames was available on the Jintronix portal (eg, active time spent on exergames, scores, number of tasks completed, and amount of trunk compensation), allowing the clinician to monitor the participant’s progression. The training protocol included five 30-minute sessions of Jintronix exergames per week for 8 weeks, targeting 20 hours of training overall.

The Reacts app [35], a videoconferencing platform, was used by the clinician to schedule videoconference meetings synchronized with sessions when the survivor of stroke was playing exergames to, for example, supervise the participant’s performance, correct their posture, grade the difficulty based on performance, and match games to the participant’s preferences and needs. Furthermore, the Reacts app was also used by the clinician to administer motivational interviewing [36].

Motivational interviewing is a person-centered approach used in behavioral interventions, which comprises behavior change techniques (BCTs) [37] and relational techniques [36]. Motivational interviewing has also been associated with the self-determination theory (SDT) [38]. The SDT is an approach that highlights the importance of autonomy and engaging individuals in their decision-making processes [39]. According to the SDT, clinicians can create a social environment that fosters autonomy (volition in one’s actions), competence (belief in one’s actions), and relatedness (a sense of belonging), which are 3 dimensions that are essential for promoting autonomous motivation and well-being [39]. In line with the SDT, survivors of stroke were given greater autonomy in determining their program by being able to choose from a range of exercises, being involved in grading the difficulty level of the games, and identifying strategies to increase the use of their affected UE in the long term through self-directed exercises and daily activities (eg, using the affected UE for dressing). In addition to exergames, supplementary exercises targeting hand fine motor skills and UE were suggested by the clinician to meet the individual goals of the survivors of stroke.

The videoconferencing sessions were scheduled as follows: 3 times a week for the first 2 weeks, twice a week for the following 2 weeks, and then once a week for the remaining 4 weeks to maintain motivation, ensure that the exercises are adequately tailored, and identify strategies to maintain the activity level of the UE after the study ended.

The training of the VirTele group was conducted at the participant’s home after the installation of the equipment and lasted approximately 30 minutes to 1 hour. The training included a practical workshop on the use of the exergames and videoconferencing system. At the end of the training, a VirTele user manual (developed by the research team) was provided to the participants. Clinicians were trained in motivational interviewing [36] before the start of the study. A motivational interviewing guide (discussion plan) based on BCTs [37] and motivational techniques [36] was conceived by the research team and provided to the clinicians as a support tool that can help them choose strategies adapted to the client’s needs.

For further information regarding the Reacts app, Jintronix exergames, and motivational interviewing, refer to the published study protocol [32] or previous case studies exploring VirTele use among survivors of stroke [33,40].

In Canada, survivors of chronic stroke receive the Graded Repetitive Arm Supplementary Program (GRASP) [41] as a home rehabilitation training program to exercise the affected UE and use it in activities of daily living [2]. Therefore, the control group received the GRASP, which included exercises for the arm and hand (strengthening and range of motion) and functional activities targeting the UE [41]. The GRASP equipment included various sizes of Lego and wooden blocks, poker chips, clothes pegs, popsicle sticks and toothpicks, paper clips of various sizes, various jars, a weight of 0.45 kg, tennis ball, foam ball, plastic cup, modeling clay, knife and fork, and a target board [41]. The control group was invited to perform the GRASP exercises for 8 weeks, 5 days per week (30-minute sessions), targeting 20 hours of exercise overall (same as the experimental group) [32]. The time spent on the GRASP program, the number of sessions, and events such as fatigue and pain were reported at T2 after the intervention was terminated. No follow-up was provided during the 8-week intervention period, similar to conventional therapy. However, at the end of the study, the participants were offered one session with the clinician to discuss strategies for improving the use of UE in activities of daily living [32]. All participants received a 30-minute training to familiarize themselves with the GRASP equipment and exercises [32].

At the start of the study, participant evaluations were conducted at the Center for Interdisciplinary Research in Rehabilitation of Greater Montreal in the presence of an evaluator. At the start of the COVID-19 pandemic, all research activities at the research site were suspended, and all evaluations were conducted remotely. For the experimental group, the evaluations were conducted using the Reacts videoconferencing system. For the control group, the evaluations were conducted either by phone or by a videoconferencing system available at the participant’s home.

Given the novelty of VirTele, the feasibility data collected for the experimental group included the number and active time spent on exergame sessions, frequency and time spent by the clinician during videoconferencing sessions, exercise adherence, and resource use (equipment and technical support). These were obtained directly from the Jintronix and Reacts systems, as well as from intervention logs completed by the clinician at the end of each session. Safety indicators, such as the occurrence of adverse events (eg, pain, fatigue, and dizziness), were documented by the clinician and technical team [32]. Information about technical difficulties was obtained from a log completed by the clinicians and technical team [32]. Satisfaction with the technology and the interaction between the clinician and the survivor of stroke were assessed using the Modified Short Feedback Questionnaire (adapted from Davis [42]) and the Health Care Climate Questionnaire (Perceived Autonomy Support) [43]. The Modified Short Feedback Questionnaire includes 2 subscales with 6 items, including a 7-point Likert-type scale ranging from extremely likely (score 1) to extremely unlikely (score 7). The first subscale evaluates the perceived usefulness (total score range 6-42; a lower score indicates that the technology is extremely useful) and the second subscale evaluates the perceived ease of use (total score range 6-42; a lower score indicates that the technology is extremely easy to use). The Health Care Climate Questionnaire includes 6 items with a 7-point Likert-type scale ranging from strongly agree (score 1) to strongly disagree (score 7), evaluating the need for support from the clinician, as perceived by the survivor of stroke (total score range 6-42; a high score indicates a higher perceived need for support from the clinician).

The process indicators were also documented to inform the validity of the study protocol and included data on recruitment rate (rate of participants per month and duration of recruitment) and retention rate (percentage of participants who completed the VirTele program) [32].

The Fugl-Meyer Assessment-UE (FMA-UE) [44,45] motor function score was used as the primary outcome to evaluate UE motor function impairment. The FMA-UE motor function score [44,45] captures synergy, coordination, and sensorimotor functions (UE, wrist, and hand). The FMA-UE score has been shown to be valid in participants with stroke [46] and reliable for administration at a distance (video observation of an evaluator administering the FMA-UE on site) [47]. Given the COVID-19 pandemic, we tested the feasibility of administering the FMA-UE motor function at a distance with no on-site evaluator (video observation of the participant’s performance), which was pretested in a previous study with a first survivor of stroke [33]. With respect to the various collection methods (by videoconferencing, by telephone, or on site), the FMA-UE motor function score was adjusted to 60 in all participants (experimental and control groups) by eliminating the parts of the scale that could not be evaluated remotely (reflex activity component).

The secondary outcomes included the Motor Activity Log-30 (MAL-30) [48,49], the Stroke Impact Scale-16 (SIS-16) [50,51], and the Treatment Self-Regulation Questionnaire-15 (TSRQ-15) [52].

The MAL-30 captures both quality (MAL-30 quality of use) and quantity of use (MAL-30 amount of use) of the affected UE in 30 daily activities (eg, writing on paper, brushing teeth, and using a fork or spoon for eating) [48,49]. The MAL-30 is reliable and valid for the poststroke population [53].

The SIS-16 is a 16-item questionnaire that captures the impact of stroke on the quality of life regarding hand function, activities of daily living, and mobility [50,51]. The SIS-16 has demonstrated good reliability and validity [54].

The TSRQ-15, a 15-item questionnaire, captures different processes of motivation consistent with the SDT, including autonomous motivation, “where a person accepts changes and behaves autonomously”; amotivation or the “lack of motivation”; external regulation, “where a person behaves to obtain a reward, or avoid punishment”; and introjected regulation, “where a person behaves for pride or to avoid feeling guilty” [52]. The TSRQ-15 has demonstrated good reliability and validity across health care and rehabilitation contexts [52,55].

Of the 42 inquiries by phone, 11 (26%) participants met the eligibility criteria and accepted to participate in the study. The rate of participant recruitment per month ranged from 0 to 6. In the VirTele group, 85% (5/6) of the participants completed the 8-week intervention (or the 3-month intervention during the COVID-19 pandemic). One of the participants discontinued the intervention because of persistent technical difficulties in accessing the VR system despite training (Figure 1).

The active time spent on exergames, the number of exergame sessions, the total number of repetitions, and activities performed in parallel to VirTele (which implies the use of UEs) of each participant receiving VirTele intervention are reported in Table 2.

The frequency of videoconference sessions varied between 9 and 11 sessions during the first 4 weeks (mean 9.8, SD 1.0), followed by 3 to 7 sessions during the second month (mean 5. SD 1.826) and 4 to 6 sessions during the third month (mean 5.0, SD 1.4).

The frequency of use and time spent on GRASP, as well as activities performed in parallel with GRASP (which implied the use of UEs), are reported in Table 3. Participant ID4, who did not use the GRASP during the 8-week intervention, reported that he was discouraged by the program as it mainly focused on his hand and wrist, which he could not move anymore since the stroke. Although participant ID9 received GRASP for 3 months, he only used the program for 6 weeks.

Technical issues were reported from the clinicians’ logs and included loss of password (to access the Reacts app) by the participant, internet issues, update of the system, sound or video cut off, and problems with the avatar (did not follow the movements of the UE). Technical issues were mainly managed by the clinician, the participant, or the participant’s caregiver. The technical team intervened once on site to deliver a 3G key, as the participant had no more internet access, and once by telephone with a participant to help them recover their passwords.

The clinicians’ logs showed that BCTs and motivational techniques were applied during the VirTele intervention with each participant in the experimental group. Among the 4 participants who completed the VirTele intervention, 3 (75%) participants (ID1, ID10, and ID11) reported more frequent use of the affected UE in activities of daily life and self-directed exercises (during the intervention), and 3 (75%) participants (ID1, ID5, and ID10) maintained the use of the affected UE after the intervention was terminated.

The 4 participants in the experimental group reported fatigue of the affected UE, which was managed by the clinician (by suggesting rest and stretching postures). Participant ID10 reported an increase in pain in the less-affected UE during the third month of VirTele; however, it did not seem to affect his adherence to the intervention, as recorded in the automatic logs accessible in the Jintronix portal (executed 59 sessions of exergames and spent 14 hours playing during the third month; Table 2).

The Health Care Climate Questionnaire showed a high score for perceived autonomy support in the experimental group (mean score 41.0, SD 1.7). Regarding the results of the perceived usefulness and perceived ease of use, most participants (3/4, 75%) found the technology extremely easy to use (mean score 11.0, SD 6.6) and extremely or quite useful (mean score 13.8, SD 15.5). Participant ID5 found the technology extremely or quite difficult to use (score 37/42) and slightly useful (score 20/42).

Regarding the MAL-30 quantity of use, 100 % (4/4) of all participants in the experimental group exhibited improvement from baseline (T1) to postintervention (T2), with the maintenance of benefits over time from 1 (T3) to 2 months (T4) after the intervention, whereas the control group showed improvement in 80% (4/5) of the participants from baseline (T1) to postintervention (T2), with maintained gains over time from 1 (T3) to 2 months (T4) after the intervention (Table 5). The MCID of the MAL-30 quantity of use was not available at that time.

For the MAL-30 quality of use, all participants in the experimental (4/4, 100%) and the control (5/5, 100%) groups demonstrated improvement from baseline (T1) to postintervention (T2), maintained over time from 1 (T3) to 2 months (T4) after the intervention, 2 of which reached the MCID (between 1.0 and 1.1; Table 5) [49].

For the SIS-16 hand function, only one of the participants in the experimental group exhibited improvement for the item “carry heavy objects (eg, bag of groceries),” with a score higher than the MCID (between 9.4 and 14.1) [51]. All participants in the control group (5/5, 100%) demonstrated improvement from baseline (T1) to postintervention (T2), maintained over time from 1 (T3) to 2 months (T4) after the intervention, with all scores higher than the MCID (Table 6).

For the SIS-16 activities of daily life, 50% (2/4) of the participants in the experimental group demonstrated improvement from baseline (T1) to postintervention (T2), maintained over time from 1 (T3) to 2 months (T4) after the intervention in only 1 participant (MCID was not detected). In the control group, 60% (3/5) of the participants exhibited improvement higher or within the MCID from baseline (T1) to 2 months after the intervention (T4; Table 6).

Regarding the SIS-16 mobility, 50% (2/4) of participants in the experimental group exhibited improvement from baseline (T1) to postintervention (T2), maintained over time from 1 (T3) to 2 months (T4) after the intervention, with a score higher than the MCID in only 1 participant. In the control group, 40% (2/5) of the participants exhibited improvement from baseline (T1) to postintervention (T2), maintained over time from 1 (T3) to 2 months (T4) after the intervention, with scores within or higher than the MCID (Table 6).

In the experimental group, 75% (3/4) of the participants demonstrated an increase in their autonomous motivation score from baseline (T1) to 2 months after the intervention (T4). Further examination of the regulations that define the controlled motivation in the experimental group showed an increase in introjected regulation from baseline (T1) to postintervention (T2) in 75% (3/4) of the participants, maintained over time from 1 (T3) to 2 months (T4) after the intervention in only 1 participant. In parallel, the external regulation showed an increase of 75% (3/4) in the participants from baseline (T1) to postintervention (T2), with a tendency to decrease in the follow-up period from 1 (T3) to 2 months (T4) after the intervention. The motivation score was substantially low in all participants at all times (Table 7).

In the control group, only one of the participants demonstrated an increase in autonomous motivation from baseline (T1) to postintervention (T2), maintained over time from 1 (T3) to 2 months (T4) after the intervention. The examination of the introjected regulation showed substantially no change from baseline (T1) to postintervention (T2) in 80% (4/5) of the participants. The external regulation scores showed a tendency of increase in 40% (2/5) of the participants, maintained over time from 1 (T3) to 2 months (T4) after the intervention in only 1 participant. One of the participants showed a decrease from baseline (T1) to 2 months after the intervention (T4) in both introjected and external regulations. Amotivation scores tended to increase in 80% (4/5) of participants (Table 7).

Regarding the primary outcome (FMA-UE), the experimental group reached the MCID from baseline (T1) to 2 months after the intervention (T4). This result is particularly relevant as the MCID was detected even if the total score of the scale was adjusted to 60, which supports the feasibility of administering the FMA-UE motor function remotely (without an evaluator on site). This result also supports the findings of a previous study [47] that examined the measurement properties of FMA-UE when administered remotely.

Regarding the secondary outcomes, both groups demonstrated improvement in the MAL-30 quantity and quality of use, which may suggest that the VirTele intervention is comparable with conventional therapy in terms of somatosensory information feedback, affecting the UE quality of movement. The supplementary exercises provided in VirTele (in addition to exergames) may have played a role in the integration of somatosensory information by manipulating real-life objects with force and tactile feedback, which are important for motor learning [14].

Regarding the quality of life (SIS-16 scores), the control group reported improvement in activities of daily life and hand function in 60% (3/5) and 100% (5/5) of the participants, respectively. In contrast, the experimental group reported varied and inconclusive results in terms of activities of daily life and hand function, despite the increased use of the UE (MAL-30 quantity) and improvement in the quality of use (MAL-30 quality). Participant ID1 reported a score of 100% (from T1 to T4) in SIS-16 hand function and activities of daily life, which indicates that no further gains can be achieved. Participant ID10 reported the appearance of pain in the less-affected UE (during the third month of VirTele intervention), which may have affected his performance during activities of daily life and the score of the SIS-16 hand function for the item “carry heavy objects (eg, bag of groceries)” as survivors of stroke often use compensatory strategies by the less-affected UE to help or assist the performance of the affected UE [59].

Further explanation of the difference between the two groups regarding the SIS-16 scores may be associated with the training paradigm; the GRASP mainly targeted the hand and wrist, with little focus on gross motor skills, whereas the VirTele intervention mainly targeted gross motor skills, with supplementary exercises for the hand. Thus, training with the GRASP might better meet individual needs when it comes to performance in activities that require fine motor skills, although both groups demonstrated improvements in the quality and quantity of use of the UE. This also suggests that combining VirTele with conventional therapy such as the GRASP may maximize the recovery potential, which echoes the findings of Laver et al [60] who determined that the use of VR combined with conventional therapy had a significant effect on UE outcomes compared with when it was used alone (not significant).