Table 1.

Characteristics and outcomes of randomized controlled trials of studies included the systematic review.

Author, year, country	Objectives	Intervention	Telerehabilitation Technology device	Population	Deficits and Outcome category measured	Results
Chen et al, 2017 China [18]	1. To evaluate the effects of home-based tele-supervised rehabilitation on physical function for stroke survivors 2. To determine if rehabilitation therapy can relieve the burden on caregivers	Physical exercises and electromyography-triggered neuromuscular stimulation (ETNS) versus Conventional therapy 12 weeks intervention	Phone-based	Intervention Group (I.G.): n = 27 Control Group (C.G.): n=27	HEMIPLEGIA (Motor) Modified Barthel Index (MBI), Berg Balance Scale (BBS), Modified Rankin Scale (mRS), Caregiver Strain Index, Root mean square (RMS) of extensor carpi radialis longus and tibialis anterior muscle	Both groups demonstrated significant effects within groups over the 3 time points in increasing MBI, BBS, RMS value of extensor carpi radialis longus and tibialis anterior, and decreasing Caregiver Strain Index (P < .001). No significant between-group differences. Percentage of participants with MRS grades 0 and 1 in 2 groups increased over time without significant difference between the groups
da Fonseca, et al, 2017 Brazil [19]	Assess the therapeutic effect of virtual reality associated with conventional physiotherapy on gait balance and the occurrence of falls after a stroke	Virtual Reality therapy program versus Conventional treatment 10 weeks intervention	Computer-based	I.G. n=14 C.G. n=13	HEMIPARESIS (Motor) Dynamic Gait Index, Number of falls	Improved gait balance and reduced occurrence of falls in both groups. Improved gait balance in the I.G. vs C.G (P = .047) Reduction in falls in the I.G. vs C.G. (P = .049) Intergroup analysis: no difference in the two outcomes
Choi et al., 2016 South Korea [20]	To develop and evaluate the feasibility and effectiveness of a mobile game-based upper extremity virtual reality program for stroke patients	Virtual Reality Rehab program on Smartphone and a tablet PC (MoU-Rehab) vs Conventional OT 2 weeks intervention	Phone-based	I.G. n=12 C.G: n=12	UPPER LIMB DYSFUNCTION (Motor) Fugl–Meyer Assessment of the upper extremity [FMA-UE], Brunnström stage [B-stage] for the arm and the hand, Manual muscle testing [MMT], Modified Barthel Index (MBI), EuroQol-5 Dimension [EQ-5D], Beck Depression Inventory [BDI], Satisfaction questionnaire	A greater improvement in the FMA-UE, B-stage, and MMT found after treatment with the MoU-Rehab than with conventional therapy Improvements in the MBI, EQ-5D, and BDI was not significantly different between the two groups Experimental group completed treatment without adverse effects, and were generally satisfied with MoU-Rehab
Lorna Paul, 2016 UK [21]	To evaluate the potential effectiveness of STARFISH in stroke survivors	Mobile phone app-based behavioral change intervention (STARFISH) vs Usual care 6 weeks intervention	Phone-based	I.G. n=16 C.G. n=8	PHYSICAL ACTIVITY (Motor) Physical activity, Sedentary time, Fatigue Severity Scale, Instrumental Activity of Daily Living Scale, Ten-Meter Walk Test, Stroke Specific Quality of Life Scale, and Psychological General Well-Being Index	Average daily step count increased by 39.3% in the I.G. and reduced by 20.2% in the C.G. (P = 0.005 for group-time interaction) Similar patterns of data and group-time interaction were seen for walking time (P = 0.002) and fatigue (P = 0.003) No significant group–time interactions for other outcome measures
van den Berg et al, 2016, Australia [22]	To investigate the effects of Caregiver-Mediated Exercises program commenced in hospital plus e-health support on self-reported mobility.	Caregiver-mediated training program with e-health support versus Usual care 8 weeks intervention	Tablet-based	I.G., n=31 C.G., n=32	MOBILITY CHALLENGES (Motor) Stroke Impact Scale mobility domain, Length of stay, Other Stroke Impact Scale domains, readmissions, motor impairment, strength, walking ability, balance, mobility, (extended) activities of daily living, Psychosocial functioning, Self-efficacy, Quality of life, and fatigue	ITT: No between-group difference in Stroke Impact Scale mobility (P = 0.6); Careers reported less fatigue (4.6, confidence interval [CI] 95% 0.3–8.8; P = 0.04) and higher self-efficacy (−3.3, CI 95% −5.7 to −0.9; P = 0.01) at week 12 Per-protocol analysis: I.G. demonstrated a trend toward improved mobility (−9.8, CI 95% −20.1 to 0.4; P=0.06), significantly improved extended activities of daily living scores at week 8 (−3.6, CI 95% −6.3 to −0.8; P=0.01) and week 12 (3.0, CI 95% −5.8 to −0.3; P=0.03), a 9-day shorter length of stay (P=0.046), and fewer readmissions over 12 months (P < 0.05)
Wolf, et al. 2015 USA [23]	To determine the efficacy of a home-based telemonitored robotic-assisted therapy as part of a home exercise program (HEP) compared with a dose-matched HEP-only intervention among individuals underserved stroke survivors < 6 months	Robotic-Assisted Therapy (The Hand Mentor Pro robotic device) + HEP versus HEP 8 weeks intervention	Computer-based	I.G., n= 47 C.G., n= 45	PERSISTENT HEMIPARESIS (Motor) Action Research Arm Test (ARAT), Wolf Motor Function Test (WMFT), Fugl-Meyer Assessment	Both groups demonstrated improvement across all upper extremity outcomes
Aparicio- López, et al. 2015, Spain [24]	To evaluate whether combination of computerized cognitive rehabilitation with right hemifield eye-patching (RHEP) in patients with left spatial neglect after stroke is better than computerized cognitive rehabilitation applied in isolation	Computerized cognitive rehabilitation (Guttmann, NeuroPersonalTrainer® telerehabilitation platform) + RHEP vs Cognitive rehabilitation program 15 one-hour interventions	Computer-based	Combination treatment (TC), N= 13 Single treatment (ST), N= 15	VISUO-SPATIAL NEGLECT (Higher cortical function) Bell Cancelation Test, Figure Copying of Ogden (FCO), Line Bisection, Baking Tray Task (BTT), Reading Task, Catherine Bergego Scale (CBS)	Both the ST and the TC group showed improvements in neuropsychological examination protocol although there were no differences pre- and post-treatment on the functional scale in either group No statistically significant differences were observed in intergroup comparison
Chumbler et al, 2015, USA [25]	To determine the effect of a multifactorial stroke telerehabilitation (STeleR) in-home intervention on falls-related self-efficacy and patient satisfaction	Functionality based exercises and adaptive strategies (STeleR) Intervention vs Usual care 3 months intervention	Phone-based	STeleR Group: n=25 Usual Care Group: n=23	FALLS (Motor) Falls Self Efficacy Falls Efficacy Scale Stroke-Specific Patient Satisfaction with Care (SSPSC) scale,	STeleR group showed significant improvements in their satisfaction with hospital care, P =.029) compared with control group. No improvements in fall-related self-efficacy.
Linder et al, 2015, USA [26]	To determine the effects of home-based robot-assisted rehabilitation coupled with a Home Exercise Program (HEP) compared with a HEP alone on depression and quality of life in people after stroke	Robot-assisted therapy plus HEP versus HEP 8 weeks intervention	Phone-based monitoring of intervention	Robot-assisted therapy and HEP, n= 51 HEP, n= 48 participants	DEPRESSION (Neuropsychiatric) Stroke Impact Scale (SIS), Center for Epidemiologic Studies Depression Scale (CES–D)	Significant changes in all but one domain on the SIS and the Center for CES-D for both groups.
Woolf et al., 2016, UK [27]	To test the feasibility of a face to face versus remotely delivered word finding therapy for people with aphasia	Word finding therapy (I) remote therapy delivered from a University lab (Mainstream video conferencing) (II) clinical site (Mainstream video conferencing)/ (III) face to face therapy/ (IV) an attention control condition. 4 weeks intervention	Video-conferencing	I group: 5 patients II group: 5 patients III group: 5 patients IV group: 5 patients	APHASIA (Higher cortical function) Recruitment and attrition rates, Participant observations and interviews, Treatment fidelity checking Tests of picture naming and naming in conversation	Compliance and satisfaction with the intervention was good Treatment fidelity was high for both remote and face to face delivery (1251/1421 therapist behaviors were compliant with the protocol) Participants who received therapy improved on picture naming significantly more than controls (mean numerical gains: 20.2 (remote from University); 41 (remote from clinical site); 30.8 (face to face); 5.8 (attention control); P <.001) There were no significant differences between groups in the assessment of conversation.
Benvenuti et al., 2014, Italy [28]	To evaluate the safety, acceptance, adherence, and effectiveness of a community-based exercise program for upper limb paresis in patients with chronic stroke and the effects of telerehabilitation monitoring in kiosks distributed through the community	Home Exercise Program (Carr and Shepherd “Motor Learning Program” + kiosk) versus Usual care 3 month intervention	Tele-monitoring at kiosks	Treatment Group, n= 45 Usual Care Group, n=143	UPPER LIMB PARESIS (Motor) Upper extremity impairments Motricity Index, Nine-Hole Peg Test, Wolf Motor Function Test (WMFT), Nottingham Extended ADLs (NEADL), Barthel Index, Short Physical Performance Battery, Stroke Impact Scale (SIS)	Patients in the experimental group demonstrated significant gains in arm function as measured by the WMFT (P <.0001), 9-Hole Peg Test (P <.009), Motricity Index (P <.0001), and NEADL (P <.002) The intervention received high satisfaction ratings and produced no adverse events Only 30% of the subjects attended kiosks regularly Outcomes for this group did not differ significantly from those who only practiced at home
Lin et al., 2014, Republic of China [29]	To evaluate the effect of a bidirectional and multiuser telerehabilitation system on balance and satisfaction in patients with chronic stroke living in long-term care facilities (LTCFs)	Physical exercise Telerehabilitation Program (The 3D animation exercise videos + 3D interactive games) versus Conventional therapy 4 weeks intervention	Computer-based	Telerehabilitation (Tele) group, n= 12 patients Conventional therapy group, n=12 patients	DYSFUNCTION OF BALANCE AND FUNCTIONAL ACTIVITY (Motor) Berg Balance Scale (BBS), Barthel Index (BI), Telerehabilitation satisfaction	Training programs conducted for both the Tele and Conventional groups showed significant effects within groups on the participant BBS as well as the total and self-care scores of BI No significant difference between groups could be demonstrated The satisfaction of participants between the Tele and the Conventional groups also did not show significant difference
Lloréns R et al, 2014, Spain [30]	To evaluate the clinical effectiveness of a Virtual reality -based telerehabilitation program in the balance recovery of hemiparetic individuals post-stroke in comparison to an inclinic program; to compare the subjective experiences; to contrast the costs	Virtual reality based Telerehabilitation system at home versus Telerehabilitation system at clinic - twenty 45-minute training sessions 3× a week intervention	Computer-based	I.G., n= 15 C.G., n= 16	RESIDUAL HEMIPARESIS (Motor) Berg Balance Scale (BBS), Balance and gait subscales of the Performance- Oriented Mobility Assessment, Brunel Balance Assessment (BBA), The System Usability Scale and the Intrinsic Motivation Inventory, Expenses in dollars for cost	Significant improvement in both groups from the initial to the final assessment in the BBS (P=0.001), in the balance (P=0.006) and gait subscales (P=0.001) of the Tinetti Performance- Oriented Mobility Assessment, and in the BBA (P=0.002; and P=0.001). No significant differences between groups in any balance scale, nor in the feedback questionnaires. With regards to subjective experiences, both groups considered the VR system similarly usable and motivating. The in-clinic intervention resulted in more expenses than the telerehabilitation program (654.72 $ per person).
Smith et al., 2012, USA [31]	To develop and test the efficacy of a web-based intervention for alleviating depression in male stroke survivors and their spousal caregivers (CG) that blends both peer and professional support.	Web-based intervention for alleviating depression Intervention condition at home (Professional Guide, Educational Videos, Online Chat Sessions, E- mail and Message Board, and Resource Room) versus Information-only control condition at home (minimal support with individualized access to relevant online information) 11 weeks intervention	Videos and messaging	I.G., n=15 C.G., n=17	DEPRESSION (Neuropsychiatrie) 20-item CESD, PHQ-9, 9 items from the Mastery Scale, 10-item Self-Esteem Scale, 11 items from the MOS Social Support Survey, Credibility/Expectancy Questionnaire	At posttest and one month later, CGs in the intervention group reported significantly lower depression than CGs in the control group with baseline depression controlled There was no significant effect on depression among SSs Although no significant treatment effects for either SSs or CGs were found on the secondary outcomes, post-treatment changes on some constructs were significantly correlated with change in depression
Redzuan et al., 2012, Malaysia [32]	To evaluate the effectiveness of an intervention using video to deliver therapy at home for patients with stroke	Video-Based Physical Therapy Program (Digital videodisk containing therapy techniques) versus Weekly outpatient therapy sessions 3 month intervention	Videos	I.G., n= 44 C.G., n= 46	HEMIPLEGIC STROKE (Motor) Modified Barthel Index (MBI) score, Incidence of poststroke complications Caregiver Strain Index	No significant differences with regard to the number of patients with improved MBI score, complication rate, or Caregiver Strain Index score between the 2 groups Both groups had significant increases in the MBI score at 3 months (P <.001 for both groups) Regression analysis revealed that only stroke severity significantly influenced the MBI score (P <.001), complication rate (P <.01), and caregiver stress level (P <.05).
Chaiyawat, et al. 2012, Thailand [33]	To develop and examine the effectiveness of individual 6-month home rehabilitation program in ischemic stroke patients upon disability and quality of life at 2 years.	Home-based physical exercise program Interventional care (Audiovisual materials) versus Usual care 6 month intervention	Videos	I.G., n=30 C.G., n= 30	POST-STROKE DISABILITY (Motor) Barthel’s Index (BI), Modified Rankin Score (mRS) and Utility index (EQ-5D), Disability and quality of life	At 2 years, the BI was significantly improved in the I.G. more than C.G. (97.2 ± 2.8 vs. 76.4 ± 9.4, P < 0.001) The good outcome, defined as BI 95–100, or mRS 0 or 1. For BI, there were 29 patients (96.7%) in I.G. vs 12 patients (42.9%) in C.G. (95% CI, 42.0, 85.0, P = 0.03) For mRS, there were 28 patients (93.3%) in intervention group vs 9 patients (32.1%) in usual care group (95% CI, 38.2, 87.0, P = 0.02) Number needed to treat for good outcome in mRS was 2.0 (95% CI: 1.0, 1.3) The mean (SD) of utility index in intervention group and control group were 0.9 ± 0.02 and 0.7 ± 0.04 respectively (P = 0.03) There was no significant interaction in baseline characteristics and treatment outcome.
Deng, et al. 2012, USA [34]	To explore the feasibility of using telerehabilitation to improve ankle dorsiflexion during the swing phase of gait in people with stroke and to compare complex versus simple movements of the ankle in promoting behavioral change and brain reorganization	Complex/Simple movement training 20 days intervention	Computer-based	Track Group: 8 patients Move Group: 8 patients	PARETIC ANKLE (Motor) 10-m walk test and motion capture system, Ankle tracking during fMRI	Dorsiflexion during gait was significantly larger in the track group compared with the move group (P=.017) For fMRI, although the volume, percent volume, and intensity of cortical activation failed to show significant changes, the frequency count of the number of participants showing an increase versus a decrease in these values from pretest to posttest measurements was significantly different between the 2 groups, with the track group decreasing and the move group increasing (P=.0014)
Chumbler, et al. 2012, USA [35]	To examine a telerehabilitation intervention that uses telehealth technology to improve outcomes of patients post-stroke after discharge home	Physical exercise program Stroke telerehabilitation (STeleR) [home televisits, in-home messaging devices (IHMD) and telephone intervention] versus Usual care 3 months intervention	Phone-based	STeleR Group: 25 participants Usual Care Group: 23 participants	PHYSICAL DYSFUNCTION AND DISABILITY (Motor) Motor subscale of the Telephone Version of Functional Independence Measure (FONEFIM) Function scales of the Late-Life Function and Disability Instrument (LLFDI), 3 subscales of the LLFDI Function Component: upper extremity function, basic lower extremity function, and advanced lower extremity function	FEONFEIM and LLFDI improved for the STeleR group and declined for the usual care group (not statistically significant) Secondary Outcomes: At 6 months, compared with the usual care group, the STeleR group showed statistically significant improvements in 4 of the 5 LLFDI disability component subscales (P 0.05), and approached significance in 1 of the 3 Function component subscales (P 0.06)
Piron, et al. 2009, Italy [36]	To compare the effects of a traditional rehabilitation therapy with an innovative rehabilitative Virtual Reality-based technique provided at distance by telemedicine	Telerehabilitation system (motor tasks and video consulting system) versus Conventional physiotherapy 4 week intervention	Video-based	TeleRehab group: 18 patients Control group: 18 patients	PARETIC UPPER LIMB (Motor) Fugl-Meyer Upper Extremity scale, ABILHAND scale, Ashworth scale	Both rehabilitative therapies significantly improved all outcome scores after treatment, but only the Fugl-Meyer Upper Extremity scale showed differences in the comparison between groups.
Huijgen, et al. 2008, Netherlan ds [37]	To investigate the feasibility of the Home Care Activity Desk (HCAD) training system for clinical use by comparing its clinical outcomes with usual care	Home Care Activity Desk training system-exercise versus Usual care 4 week intervention	Not clear	I.G., n= 11 C.G., n= 5	ARM/HAND DYSFUNCTION (Motor) Action Research Arm Test and the Nine-Hole Peg Test, Visual Analogue Scale	There were no significant differences between the two groups on the outcome measures (ARAT and NHPT); in both groups, patients maintained or even improved their arm/ hand function The HCAD training was found to be as feasible as usual care in terms of clinical outcomes, and both therapists and patients were satisfied with the HCAD intervention
Piron et al., 2008, Italy [38]	To assess how the presence or absence of the therapist affected patient satisfaction in the VR sessions as compared to patient satisfaction experienced in the Tele-VR sessions.	Virtual reality (VR) therapy program (3D motion tracking system) at home versus hospital based therapy - 4 weeks intervention	Video-based	Tele-VR group: 5 patients VR group: 5 patients	ARM MOTOR IMPAIRMENT (Motor) Multidimensional, disease and treatment specific satisfaction questionnaire, Fugl-Meyer scale, Upper Extremity (Fugl-Meyer UE) score	Tele-VR treated patients showed median values equal to or higher than the VR group patients in all 12 items investigated, except one In motor performance, the Tele-VR group improved significantly (P 0.05), while the VR group showed no significant change Patients assigned to the Tele-VR group were able to engage in therapy at home and the videoconferencing system ensured a good relationship between the patient and the physical therapist whose physical proximity was not required
Carey et al., 2007, USA [39]	To compare 2 telerehabilitation training strategies, repetitive tracking movements versus repetitive simple movements, to promote brain reorganization and recovery of hand function	Computerized tracking training versus Movement training 2 week intervention	Computer-based	Track group (TG): 10 patients Move group (MG):10 patients	PARETIC HAND - FINGER EXTENSION DISABILITY (Motor) Box and Block test, Jebsen Taylor test, and Finger range of motion, along with a finger-tracking activation paradigm during fMRI.	The TG showed significant improvement in all 4 behavioral tests; the MG improved in the Box and Block and Jebsen Taylor tests A consistent group pattern of brain reorganization was not evident The move group, after crossing over, did not show further significant improvements

VR=Virtual Reality; OT=Occupational Therapy; ITT=Intention to treat analysis; I.G.=Intervention Group; C.G.= Control Group; HEP=Home Exercise Program.