PHYSIOTHERAPY THEORY AND PRACTICE

2022, VOL. 38, NO. 13, 2402–2416

https://doi.org/10.1080/09593985.2021.1941458

Effectiveness of virtual reality-based rehabilitation versus conventional therapy

on upper limb motor function of chronic stroke patients: a systematic review and
meta-analysis of randomized controlled trials
Reem M. Al-Whaibi PhD, PTa, Maher S. Al-Jadid PhD, MDb, Hager R. ElSerougy PhD, PTc,
and Wanees M. Badawy PhD, PT d
a
Rehabilitation Sciences Department, Health and Rehabilitation Sciences College, Princess Nourah Bint Abdulrahman University, Riyadh, Saudi
Arabia; bRehabilitation Medicine Department, Prince Sultan Military Medical School, Riyadh, Saudi Arabia; cDepartment of Physical Therapy for
Neuromuscular Disorders and its Surgery, Faculty of Physical Therapy, Misr University for Science and Technology, Giza, Egypt; dDepartment of
Physical Therapy for Neuromuscular Disorders and its Surgery, Faculty of Physical Therapy, Cairo University, Giza, Egypt

ABSTRACT ARTICLE HISTORY

Objective: To systematically review the available randomized controlled trials in the literature Received 14 August 2020
concerning the application of virtual reality (VR) rehabilitation interventions compared to conven­ Revised 19 February 2021
tional physical therapy, in regaining the upper limb motor function among patients with chronic Accepted 9 May 2021
stroke. KEYWORDS
Methods: A systematic electronic database search was conducted for related studies published Stroke; virtual reality;
from inauguration and until June 25, 2020 in nine databases. Another new search was done on conventional therapy; upper
February 1, 2021 and no new studies were identified. limb
Results: Six studies were included in the analysis. Significant improvement was seen following the
VR therapy in patients with chronic stroke, compared to their scores prior to it (SMD = 0.28; 95%
CI = 0.03–0.53; p = .03). There was neither heterogeneity (I2 = 0% and P = .5) nor a risk of bias (P = .8)
among the included studies. VR interventions produced a comparable effectiveness to that of the
conventional rehabilitation, with no statistically significant difference (SMD = 0.15; 95% CI = −0.14–
0.44; P = .3). There was neither heterogeneity (I2 = 40% and P = .1) nor a risk of bias (P = .5) among
the included studies.
Conclusions: The upper limb motor function of patients with chronic stroke who underwent VR-
based rehabilitative intervention showed significant improvement as compared to the pre-
treatment state. Our analysis also revealed no superiority of VR interventions over conservative
therapies; however, the difference observed did not accomplish statistical significance.

Introduction
in turn, will significantly impact the ability to self-care as
Despite the novel advancements in stroke management well as their ability to participate in social activities.
in a clinical setting, which had considerably limited the Since the use of upper limbs is not dispensable during
number mortality cases attributable to stroke, it is still our social and daily life activities, it is of great impor­
identified as highest contributor to mortality and motor tance to regain their function in patients who recently
disability worldwide (Global of Burden Disease, recovered from stroke (Bailey, Klaesner, and Lang,
2016Stroke Collaborators, 2019; Simpkins et al., 2020). 2015).
Meanwhile, the number of neurologically-impaired In this context, many motor rehabilitative programs,
patients with significant disabilities has increased involving motor learning paradigms, have been pro­
remarkably (Baumann, Lurbe-Puerto, Alzahouri, and posed in order to promote the functional recovery of
Aïach, 2011). That being said, only a limited number an impaired limb of patients who recently recovered
of cases could witness a functional improvement in their from stroke (Arya, Pandian, Verma, and Garg, 2011;
upper limbs (ULs) (Choo et al., 2015). Even though Brewer, Horgan, Hickey, and Williams, 2013;
physical training combined with aerobic exercise is con­ Langhorne, Bernhardt, and Kwakkel, 2011). Based on
sidered the gold standard of motor rehabilitation follow­ the witnessed advances in technologies exploring the
ing stroke recovery, approximately 15 to 30% of patients human brain, which has improved our understanding
with stroke are disabled for life (Roger et al., 2012). This, of neuroplasticity and its correlation with stroke

CONTACT Wanees M. Badawy wanees.alamir@pt.cu.edu.eg Assistant Professor of Physical Therapy for Neuromuscular Disorders and its Surgery,
Faculty of Physical Therapy, Cairo University, Giza 12613, Egypt
© 2021 Taylor & Francis Group, LLC
 PHYSIOTHERAPY THEORY AND PRACTICE 2403

recovery (Johansson, 2011; Kleim and Jones, 2008; experience as well as the ability of the medical person­
Talelli, Greenwood, and Rothwell, 2006), neural plasti­ nel. On the other hand, it has been reported that the
city has been proposed to have the ability to alter the trainings involved in conventional therapy, despite
structure as well as the function of the central nervous their repetition patterns, are still not sufficient to
system (Hosp and Luft, 2011; Takeuchi and Izumi, induce neural-plasticity-based motor improvement of
2012a). In turn, the rehabilitation of impaired limbs in the disabled limb (Saposnik and Levin, 2011). The
stroke survivors has been noted to depend mainly on aforementioned limitations drove researchers to
enhancing neural plasticity (da Silva Cameirão et al., develop and test other novel options that would be
2011; Saposnik et al., 2016). more efficient in regaining functional recovery of
In recent years, numerous stroke rehabilitation pro­ impaired limb in stroke patients, such as virtual reality
grams have been proposed to improve limb function (VR) (da Silva Cameirão et al., 2011; Laver et al., 2015;
following stroke, based on clinical research characteriz­ Lohse et al., 2014).
ing brain remodeling because of neuroplasticity Virtual reality is recognized as a novel, computer-
(Brewer, Horgan, Hickey, and Williams, 2013; Chen based simulation environment, which helps engage indi­
and Shaw, 2014; Dobkin and Dorsch, 2013; Johansson, viduals in various environments and tasks inside the
2011; Pollock, Baer, Langhorne, and Pomeroy, 2007; assigned virtual scenario (Saposnik and Levin, 2011),
Takeuchi and Izumi, 2012b; Veerbeek et al., 2014; even at advanced intensities than those in conventional
Zorowitz and Brainin, 2011). The efficacy of such pro­ rehabilitation programs for chronic stroke patients
grams in enhancing motor function has been examined (Chen et al., 2016). This approach has recently emerged
and verified by various meta-analyses studies (Hsu et al., as a potential therapeutic option for motor function
2012; Laver et al., 2015; Nijland, Kwakkel, Bakers, and rehabilitation in patients with UL disabilities after stroke
van Wegen, 2011), but still due to the complex nature of (Laver et al., 2015). VR allows stroke survivors to be
stroke recovery and the methodological heterogeneity in involved in a repetitive, intensive, and goal-oriented
various studies (Chen and Shaw, 2014; Pollock, Baer, program in order to promote and enhance their func­
Langhorne, and Pomeroy, 2007; Veerbeek et al., 2014; tional disabilities, activity restraints, and societal contri­
Zorowitz and Brainin, 2011), it is difficult to draw bution limitations (Merians et al., 2002). In addition, VR
a conclusion as to which programs are superior to others offers real-time visual feedback for movements, and
or which ones could be adopted for the entire rehabilita­ thus, it enhances patients’ engagement in such enjoyable
tion process. rehabilitation protocols (Levin, Weiss, and Keshner,
Task-based approach of sufficient intensity and gen­ 2015).
eral exercises have been proposed as methods that be Despite the fact that various VR interventions have
added to stroke rehabilitation program. (French et al., been directed to patients who recovered stroke, the
2007; Kita et al., 2013). Various passive stimulation efficacy of VR rehabilitation interventions is not yet
approaches that use proprioceptive, hepatic and visual confirmed by high-quality evidence. In particular, ran­
feedback are used to regain lost sensory and motor domized clinical trials on the qualitative as well as quan­
function, in stroke patients (Doyle, Bennett, Fasoli, and titative valuable effects of VR on upper extremities
McKenna, 2010; Kita et al., 2013). These include differ­ motor function and independence in carrying out
ent techniques of electrical stimulation such as: neuro­ ADLs among patients with chronic stroke are still
muscular stimulation; cutaneous electrical stimulation; scarce. Therefore, we conducted the current investiga­
transcutaneous electrical nerve stimulation (TENS) tion to systematically review the available high-quality
(Kita et al., 2013); intermittent pneumatic compression evidence (i.e. randomized controlled trials) in the litera­
(Cambier, De Corte, Danneels, and Witvrouw, 2003; ture concerning the use of VR rehabilitation interven­
Kita et al., 2013); thermal stimulation (TS) (Chen, tions, compared to conventional physical therapy, in
Liang, and Shaw, 2005; Wu et al., 2010); and peripheral regaining the UL motor function among patients with
magnetic stimulation (Kita et al., 2013). chronic stroke.
Conventional therapies, in the form of occupational
or physical therapy, are commonly approached to
further enhance the regain of motor functionality of Methods
upper extremities of stroke patients with significant
Search strategy and study selection
disabilities (Ackerley et al., 2016; Jarvis, Reid,
Edelstyn, and Hunter, 2014). That being said, this con­ We followed the recommendations of the Preferred
ventional approach is often time-consuming with low Reporting Items for Systematic Reviews and
compliance rate, and the outcomes often depend on the Meta-Analyses (PRISMA) checklist which were
2404 R. M. AL-WHAIBI ET AL.

built for conducting the systematic review and meta- Data extraction
analysis. We searched for eligible papers until
We extracted the data using an Excel sheet
June 25, 2020 in nine databases listed in order:
(Microsoft office 365, Microsoft, Redmond, WA). At
metaRegister of Controlled Trials (mRCT), Web of
least two investigators extracted the necessary infor­
science (ISI), Scopus, Google Scholar, PubMed,
mation from each paper and revised by a third inves­
System for Information on Grey Literature in
tigator for more accuracy. A senior author was
Europe (SIGLE), Clinical trials.gov, Virtual Health
consulted in case of disagreement between the three
Library (VHL) and the International Clinical Trials
investigators.
Registry Platform (ICTRP) operated by the World
Health Organization (WHO) databases using key­
words, medical subject (MeSH) terms. Another
search on the same databases was carried out on Risk of bias assessment
February 1, 2021. Publication types based on the Due to the inclusion of RCTs only, we decided to use the
participants, interventions, comparisons, outcomes, Cochrane risk-of-bias tool for randomized trials (RoB 2)
and study design (PICOS) framework: participants (Sterne et al., 2019). Three investigators rated the quality
were any patient with chronic stroke (chronic stroke of evidence according to several metrics: randomization
will be defined as ≥ 6 months after onset) (Wechsler bias, allocation bias, missing data bias, outcome bias,
et al., 2018), affected upper extremity, the interven­ and selection bias (Sterne et al., 2019). A senior author
tion was the VR-based therapy, the comparison was was consulted when there was a disagreement between
the conventional therapy, and the treatment effec­ the three investigators.
tiveness was the outcome of interest. “Conventional
therapy” could consist of usual conventional ther­
apy, physical therapy, occupational therapy or Statistical analysis
a combination of any of them.
“Effectiveness” was defined as the positive change All data were analyzed using R software version 4.0.1
from baseline values to the last follow up point. In (Team RC, R Foundation for Statistical Computing:
the case of multiple outcomes, effectiveness is mea­ Vienna, Austria). The standardized mean difference
sured by the change in the outcome that has been (SMD) effect size and its variance were calculated
reported as the primary outcome or most relevant to using the pre- and post-intervention data for both
other studies to maintain homogeneity (Molley, intervention and control groups (Hedges,
1990). We further did a manual search of references Pustejovsky, and Shadish, 2013). Using a “meta”
in our included papers to avoid missing relevant package, change from baseline for both intervention
studies (Vassar, Atakpo, and Kash, 2016). We and control were analyzed to calculate the pooled
included randomized-controlled trials study design SMD and the corresponding standard errors (SE).
that had a matched control group comparing the The SMD was used due to the difference in measure­
VR techniques to conventional therapy in chronic ment methodology among included studies plus it is
stroke patients. We did not restrict inclusion criteria more generalizable than mean difference (Takeshima
for publication data or language of the included et al., 2014). The related 95% confidence intervals
studies to prevent results bias by missing one rele­ (CI) of the computed effect size were calculated
vant paper. using a fixed- or random-effects model based on the
We excluded records or full texts that include extent of heterogeneity. To assess heterogeneity,
animal populations, non-randomized trials and Q statistics and I2 test considering were used, with
other study design such as cohorts, case-control, I2 value higher than 50% or P less than 0.05 consid­
cross-sectional and case series studies. We also ered significant (Higgins et al., 2019).
excluded studies that recruit the same patients We conducted Egger’s regression test to assess pub­
because they will lead to results bias. Three authors lication bias which was considered significant when P is
screened the title and abstracts of each record then less than 0.10 (Egger, Davey Smith, Schneider, and
performed full texts screening to for the inclusion of Minder, 1997; Peters et al., 2006). Moreover, using meta-
related papers. In each step, a fourth review revised regression, we investigated if any of the following factors
the screening results for ensuring more accuracy. induced response to VR interventions: 1) VR session
A senior author was consulted if disagreement duration (min); 2) post-stroke duration (months);
occurred. and 3) Age (years) (El-Qushayri et al., 2020).
 PHYSIOTHERAPY THEORY AND PRACTICE 2405

Results a multicenter study. The overall population size of the

included papers ranged from 12 individuals (Stockley
Search results
et al., 2017) up to 52 patients (Schuster-Amft et al., 2018)
Systematic search resulted in 526 records; of these 421 with equal sample sizes in both arms for most of the
records were screened using title and abstract after studies. Two studies reported a follow-up period for
exclusion of duplicates. Title and abstract screening 12 weeks (Kim, 2018; Stockley et al., 2017); another
yielded 84 papers eligible for full-text screening. Out of two for 4 weeks (Schuster-Amft et al., 2018;
these, we included five studies, in addition to another Subramanian et al., 2013); one for 8 weeks (Ahmad
one study from manual search (Figure 1). et al., 2019); and one for 3 weeks (Crosbie et al., 2012)
(Table 1).
Regarding the risk of bias, it was high in only a small
Characteristics and bias of the included studies
portion of all domains. The domain with the highest risk
Six studies compared VR versus controls. Two studies of bias was of the randomization process, while devia­
were conducted in the UK, one in Malaysia, one in tion from the intended interventions and the selective
Korea, one in Canada, and the last one was reporting had only some concerns. In contrast, there was

Figure 1. PRISMA flow diagram showing the process of the review.

 2406

Table 1. Characteristics of the included studies.

Outcome
Author/Year/ Sample Stroke Experimental Measurement Program
Country Size Description Male% Age Inclusion Criteria Exclusion Criteria Arm Control Arm Tools Outcomes Duration
Ahmad/2019/ 36 NR Exp= 17 Exp= 57 (1) stroke survivors at Those with severe A half hour of VR 2 hours of (1) FMA-UE, Significant time–group 8 weeks
Malaysia (94.4%), (8.20), least 6 months post- cognitive impairments games using routine (2) WMFT, interaction effect for
Con= 14 Con= stroke to include those (MMSE score less than Cy-Wee standard (3) IMI, (4) IMI (p = 0.001), Lawton
R. M. AL-WHAIBI ET AL.

(77.8%) 63 with chronic stroke, (2) 17) and taking any Z game exercise IADL, and IADL (p = 0.01) and SIS
(10.54) aged 18 years and prescribed drugs that controller plus therapy, 8 (5) SIS domain of
above, (3) the affected could potentially affect 1.5 hours of sessions (once communication (p =
arm scored at least 4 physical function and standard a week) 0.03). A significant
out of 6 according to balance (such as physiotherapy time effect was found
the motor assessment corticosteroids, exercises, 8 in FMA-UE (p = 0.001),
scale, and (4) able to antipsychotics, or sessions (once WMFT (p = 0.001),
participate in the VR antidepressants). a week) Lawton IADL (p =
games training 0.01), and SIS domains;
without limitation strength, ADL and
(presently with good stroke recovery (p <
health and no self- 0.05)
reported orthopaedic,
medical, or painful
conditions).
Crosbie/2012/ 18 First stroke Exp= 5 Exp= Medically stable; an adult Patients who presented VR (30-45 Physiotherapy (1) Upper limb Both groups 3 weeks
UK (55.6%), 56.1 aged 18–85 years; 6– with mental score test minutes), (duration and Motricity demonstrated small
Con= 5 (14.5), 24 months following of less than 7/10,15 three sessions number of Index and (7–8 points on upper
(55.6%) Con= a first stroke and able a star cancellation per day, for 3 sessions were (2) Action limb Motricity Index
64.6 to follow a two-step score of less than 48/ successive not reported) Research and 4 points on the
(7.4) command. 52,16 scored less than weeks. Arm Test. Action Research Arm
25 out of 100 on the Test), but non-
upper limb Motricity significant, changes to
Index,17 had800 their arm impairment
Clinical Rehabilitation and activity levels.
26(9) comorbid
conditions affecting
their rehabilitation
potential (e.g. cardiac,
respiratory or arthritic
problems) and
reported arm pain on
a visual analogue scale
of >6/10. Patients with
pacemakers were also
excluded.
(Continued)
Table 1. (Continued).
Outcome
Author/Year/ Sample Stroke Experimental Measurement Program
Country Size Description Male% Age Inclusion Criteria Exclusion Criteria Arm Control Arm Tools Outcomes Duration
Kim/2018/ 24 NR Total= 15 Exp= (1) Those who had NCS VR programs (1) FMA and The experimental group 12 weeks
Korea (62.5%) 50.91 passed at least six using video traditional (2) MFT participants’ daily
(9.57), months after the onset games plus rehabilitation living activities
Con= of a stroke, (2) Those traditional (12 weeks, 5 improved after
57.23 who scored 21 points rehabilitation times per training. In addition,
(14.63) or above in the Berg (12 weeks, 3 week, and 30 the experimental
Balance Scale, (3) times per min per time, group scored
Those who scored week, and 40 same exercise significantly higher on
Stage 4 or above in the min per time, program) all the tests than the
Brunnstrom motor same VR control group, but
recovery stages of arm exercise upper extremity
and hand, and (4) program) function between the
Those who understood groups was not
the therapist’s significantlydifferent.
instructions with 24
points or above in the
Mini-Mental State
Examination
Shuster-Amft 52 Ischemic stroke= Exp= 16 Exp= Patients had to be at Previous or current VR (four 45- Conventional (1) BBT, (2) Patients in the 4 weeks
/2018/ 43, (72.7%), 61.3 least six months after functional deficits of minute physiotherapy CAHAI-13, experimental and
Switzerland Hemorrhagic= Con= 23 (13.4), his or her first-ever the arm and hand training or and (3) SIS control group
11 (71.9%) Con= stroke (ischemic or motor function not sessions per occupational improved: Box and
61.2 hemorrhagic) with due to stroke, severe week for 4 therapy (four Block Test mean 21.5
(11.2) a persistent motor cognitive deficits weeks) 45-minute ±SD 16 baseline to
deficit of the arm and indicated by a Mini- training mean 24.1±SD 17
hand, indicated by Mental State sessions per follow-up; Chedoke-
a Chedoke-McMaster Examination score of week for 4 McMaster Arm and
Stroke Assessment 20 or lower, severe weeks) Hand Activity
(CMSA) score of three visual disorders, or Inventory mean 66.0
or greater on the arm a history of epileptic ±SD 21 baseline to
subscale and two or seizures triggered by mean 70.2±SD 19
greater on the hand visual stimuli within follow-up.
subscale. Patients also the past six months.
had to be able to sit in
a normal chair without
armrests or backrest
support and to score at
least one on the Box
and Block Test, which
was the primary
outcome measure.
PHYSIOTHERAPY THEORY AND PRACTICE

(Continued)
2407
Table 1. (Continued).
2408

Outcome
Author/Year/ Sample Stroke Experimental Measurement Program
Country Size Description Male% Age Inclusion Criteria Exclusion Criteria Arm Control Arm Tools Outcomes Duration
Stockley/ 12 NR Total= 6 Exp= Community dwelling NCS VR (30-minute Gym therapeutic (1)MAL, (2) There were no significant 12 weeks
2017/UK (50%) 70.8, stroke survivors who session), with exercise in the BBT, and (3) differences between
Con= had a confirmed first a total of 18 onsite gym FSS the experimental and
70.6* stroke more than 6 session over with a longer gym groups although
months ago. 12 weeks training there were significant
Participants were duration. within group
included if they improvements on the
complained of upper motor activity log
R. M. AL-WHAIBI ET AL.

limb difficulties (median change: 0.59,

without pain, could range: 0.2–1.25; p<
move at least one 0.05) within the
block within 60 experimental group.
seconds on the box
and block test, and
were not severely
cognitively impaired,
scoring over 17 on the
Montreal Cognitive
Assessment.
Subramanian/ 32 Ischemic stroke= Exp= 11 Exp= Inclusion Criteria: (1) (1) brainstem/cerebellar VR (45-minute Physical therapy (1) FMA, (2) Data were divided based 4 weeks
2013/ 22, (68.8%), 60.0 were between 40 and lesions, (2) sessions), 3 (45-minute RPSS, (3) on the affection level
Canada Hemorrhagic= Con= 12 (11.0), 80 years, (2) sustained comprehension times a week sessions), 3 WMFT, and (mild and moderate to
7, No (75%) Con= a single ischemic or difficulties, and (3) for 4 weeks. times a week (4) MAL-AS severe): Impairment
information= 3 62.0 hemorrhagic stroke 6 marked apraxia, for 4 weeks. level affected
(9.7) to 60 months attention, or visual outcomes. After VE
previously, (3) scored 3 field deficits. training, the mild
to 6/7 on the Chedoke- group increased elbow
McMaster Stroke extension (RET, 25.5°).
Assessment arm The moderate-to-
subscale, and (4) had severe group in VE
no other neurologic or increased arm use at
neuromuscular/ POST (0.5 points) and
orthopedic problems reaching ability at RET
affecting the upper (2.2 points). The
limb and trunk. moderate-to-severe
group training in PE
increased reaching
ability earlier (POST,
1.7 points) and both
elbow extension
(10.7°) and arm use
(0.4 points) at RET, but
these changes were
accompanied by
increased
compensatory trunk
displacement (RET,
30.2 mm).
*Median; FMA-UE: Fugl-Meyer assessment for upper extremities; WMFT: Wolf motor function test; IMI: intrinsic motivation inventory; IADL: Lawton of instrumental activities of daily living; SIS: stroke impact scale; MFT:
Manual Function Test; NCS: Not Clearly Specified; NR: Not Reported; BBT: Box and Block Test; CAHAI-13: Chedoke McMaster Arm and Hand Activity Inventory; Exp: Experimental arm; Con: Control arm; FSS: Fatigue Severity
Scale; MAL: Motor Activity Log; RPSS: Reaching Performance Scale in Stroke; MAL-AS: Motor Activity Log-Amount Scale
 PHYSIOTHERAPY THEORY AND PRACTICE 2409

Figure 2. Quality of the included studies. A: Risk of bias graph: review authors’ judgments about each risk of bias item presented as
percentages across all included studies; B: Risk of bias summary: review authors’ judgments about each risk of bias item for each
included study (D1: bias arising from the randomization process; D2: bias due to deviations from intended interventions, D3: bias due
to missing outcome data, D4: bias in measurement of the outcome, and D5: bias in selection of the reported result).

Figure 2. Continued.

Figure 2. Quality of the included studies. A: Risk of bias graph: review authors’ judgments about each risk of bias item presented as
percentages across all included studies; B: Risk of bias summary: review authors’ judgments about each risk of bias item for each
included study (D1: bias arising from the randomization process; D2: bias due to deviations from intended interventions, D3: bias due
to missing outcome data, D4: bias in measurement of the outcome, and D5: bias in selection of the reported result).
2410 R. M. AL-WHAIBI ET AL.

(SMD = 0.15; 95% CI = −0.14–0.44; P = .30). There

was neither heterogeneity (I2 = 40% and P = .1) nor
a risk of bias (P = .56) among the included studies. Most
of the studies revealed a minimal significant difference
of VR interventions over conservative therapies
(Figure 4). The meta-regression analysis did not show
any significant effect of the VR session duration (min),
post-stroke duration (months), or age (years) factors on
the comparative treatment effects (Supplementary
Table 2).

Discussion
The present systematic review and meta-analysis are
aimed to summarize the comparative effectiveness of
VR-based interventions compared to conventional
therapies in regaining the functional state of ULs in
patients with chronic stroke (after ≥ 6 months of stroke
onset).
VR-based programs have been proposed to have sev­
eral benefits, including being economical, providing
clear motivation, improving treatment effects, and pro­
viding opportunities for stroke patients get engaged in
Figure 2. Continued.
a realistic environment resembling real objects and
events through the integration of multiple sensory sti­
a low risk of bias in domains of missing data and out­ muli such as visual, auditory, tactile, and somatosensory
come measures among all of the included studies systems (da Silva Ribeiro et al., 2015; Gibbons,
(Figure 2). Thomson, de Noronha, and Joseph, 2016;
Iruthayarajah et al., 2017; Tinga et al., 2016). In light of
the previous reasons, a VR-based rehabilitation program
Effectiveness of VR therapy
has recently emerged as an approach of great value in
Six studies (seven datasets) evaluating the effectiveness promoting functional recovery of upper extremities in
of VR therapy (pre- versus post-intervention) were patients after stroke.
included in the analysis. There was a significant In this regard, a total of six randomized controlled
improvement in following the VR therapy in stroke trials (RCTs, 174 patients) that included stroke survivors
survivors, compared to their scores prior to it and who were assigned to receive one of two interven­
(SMD = 0.28; 95% CI = 0.03–0.53; P = .03). tions (i.e. VR-based intervention vs. conventional ther­
Noteworthy, only one study (Subramanian et al., 2013) apy) were included in our analysis. We analyzed the pre-
did show a slight deterioration in the motor functions of post effectiveness of VR-based interventions regarding
the upper extremity; however, this difference was not the recovery of UL motor function (FMA-UE), activity
significant. There was neither heterogeneity (I2 = 0% and (BBT), and participation (MAL-AS) in patients with
P = .51) nor a risk of bias (P = .8) among the included chronic stroke and significant impairment of UL motor
studies (Figure 3). The meta-regression analysis did not function. We noted a significant improvement in UL
show any significant effect of the VR session duration function following VR-based intervention compared to
(min), post-stroke duration (months), or age (years) the pretest score, with an overall effect size of 0.28 (95%
factors on the treatment effects (Supplementary CI = 0.03–0.53). Furthermore, in terms of between-
Table 1). group differences (VR-based program vs. conventional
In the same context, six studies (seven datasets) eval­ therapy), the effectiveness of VR was slighter higher than
uating the comparative effectiveness of VR therapy and conventional therapy in improving limb performance;
conventional therapy were included in the analysis. however, this difference did not reach statistical signifi­
Overall, VR interventions produced a comparable pro­ cance. Our analyses did not reveal any significant het­
portional recovery to that of the conventional rehabilita­ erogeneity among included studies nor a significant
tion, with no statistically significant difference publication bias.
 PHYSIOTHERAPY THEORY AND PRACTICE 2411

Figure 3. Forest plot for the effectiveness of virtual reality-based rehabilitation on upper limb motor function of chronic stroke patients
(pre- versus post-intervention).

Several factors should be carefully considered The protocols used in the investigated studies where
when interpreting our results. For instance, although different. Ahmad et al. (2019) conducted bilateral ULs
all the involved studies included patients who had VR games inform of Mosquito Swat, Music Catch,
a chronic stroke, there was a noticeable difference ReBounce, Bejeweled, and Balloon Popping, 10-Pin
in stroke severity and characters among the included Bowling, Air Hockey, Mah-Jongg, and Solitaire.
populations. For example, one study recruited Crosbie et al. (2012) used virtual tasks simulate a range
patients with persistent hypertonia (score 4 or more of upper limb tasks related to reach to target, reach and
in the Motor Assessment Scale) (Ahmad et al., 2019); grasp and game tasks within a 3D environment. In the
one study recruited patients after their first experi­ study conducted by Kim (2018) the study group under­
ence of stroke without much detail on the degree of went a range of exercise programs including tennis,
their disability of the severity of the stroke (Crosbie baseball, and golf in Wii Sports. Schuster-Amft et al.
et al., 2012); one study included patients with mild (2018) and Stockley et al. (2017) utilized a game-based
disability (stage 4 or more in the Brunnstorm Motor virtual reality system (YouGrabber) for upper-limb
Recovery Staging) (Kim, 2018); two studies included rehabilitation. While participated patients in
moderate-to-severely impaired patients (Chedoke- Subramanian et al. (2013) study were involved in
McMaster Stroke Assessment Stage of 3 or more) a simulated supermarket scene using 3D Computer
(Schuster-Amft et al., 2018; Stockley et al., 2017); Assisted Rehabilitation Environment (CAREN) soft­
and one study included patients who could move ware. These variations between the introduced protocols
a least 1 block within 60 seconds in the Box and had an impact on the results of these studies.
Block Test (BBT) (Subramanian et al., 2013). In the A critical point should also be considered, the dura­
same context, two randomized trials did not report tion of the VR-based intervention varied greatly among
their exclusion criteria clearly (Kim, 2018; Stockley included trials, for example, the program duration was
et al., 2017). 3 weeks in one trial (Crosbie et al., 2012), 4 weeks in two

Figure 4. Forest plot for the comparative effectiveness of virtual reality-based rehabilitation versus conventional therapy on upper limb
motor function of chronic stroke patient.
2412 R. M. AL-WHAIBI ET AL.

trials (Schuster-Amft et al., 2018; Subramanian et al., which included 27 trials/ 1094 patients concluded that
2013), 8 weeks in one trial (Ahmad et al., 2019), and VR-based intervention in stroke survivors was signifi­
12 weeks in two trials (Kim, 2018; Stockley et al., 2017). cantly better than conventional therapy for the recovery
The intensity of the intervention was low in some trials of upper extremity function. Our finding is inconsistent
(30 minutes, one session per week, for 8 weeks) (Ahmad with these observations; however, it should be noted that
et al., 2019) and was very intense in other studies (30– these meta-analyses analyzed stroke patients collectively,
45-minute session, 3 sessions per day, for 3 weeks) of acute, subacute, and chronic stroke, while we investi­
(Crosbie et al., 2012). Also, the intervention arm of the gated those with chronic stroke only. Going in line with
two trials included patients who underwent adjuvant our finding, when Mekbib et al. (2020) conducted
VR-based and conventional interventions. This variabil­ a subgroup analysis based on patients with chronic
ity in the protocols of VR-based and conventional ther­ stroke only, the author noted no statistically significant
apy interventions could potentially explain the reason difference between VR-based interventions and conven­
we could not reach a statistically significant result in the tional therapy (control group) in terms of improving
efficacy of VR interventions in improving UL motor upper limb functions (FMA-UE).
function compared to conventional therapies. Moreover, Lee, Park, and Park (2019) conducted
Another point to consider is that the outcome mea­ a systematic review of 21 studies and 562 patients to
surement tools of UL function were variable among examine the efficacy of VR rehabilitation training com­
included studies in our meta-analysis, and the most pared to conventional therapy in chronic stroke popula­
commonly used tools were: Fugl-Meyer Assessment for tion in regaining UL function, and the authors reported
Upper Extremities (FMA-UE) (Ahmad et al., 2019; Kim, an overall moderate effect size of 0.431. However, the
2018); Motricity Index (MI) (Crosbie et al., 2012); Box authors did not report if this finding was related to a pre-
and Block Test (BBT) (Schuster-Amft et al., 2018; post analysis or a comparison between VR and conven­
Stockley et al., 2017); and Motor Activity Log-Amount tional therapy interventions.
Scale (MAL-AS) (Subramanian et al., 2013). Given the fact that physical adaptation to physical
Furthermore, the majority of included studies had practice usually requires 6 to 8 weeks of training
some concerns (Kim, 2018; Schuster-Amft et al., 2018; (Kraemer et al., 2002), we assumed that adaptation
Stockley et al., 2017; Subramanian et al., 2013) in terms to VR rehabilitation interventions requires around
of risk of bias, while one study had a low risk of bias the same duration. However, half of our included
(Crosbie et al., 2012) and another had a high risk of bias trials included patients who underwent VR rehabili­
(Ahmad et al., 2019). There were major concerns related tation training for a period of 3 to 4 weeks (Crosbie
to the selection of reported results in most studies since et al., 2012; Schuster-Amft et al., 2018; Subramanian
most studies did not clarify if they had a previously et al., 2013). We think that this difference, in part,
published protocol and whether or not they complied would be the reason why no statistical significance
with these protocols. Moreover, there was a major con­ was noted between both interventions. The charac­
cern related to the deviation from the intended inter­ teristics of chronic stroke patients would make it
vention since only one study reported that they excluded difficult to maintain long-term VR therapy. The dis­
drugs that could have affected the physical functionality advantage of VR rehabilitation is that it focuses on
of included patients (i.e. corticosteroids, antipsychotic, accomplishing the rather than the quality of task
and antidepressants) (Ahmad et al., 2019), and only two performance (Molier, Van Asseldonk, Hermens, and
studies monitored if the patients secretly carried out self- Jannink, 2010). That being said, research should put
implemented physical activity outside of the VR or the into consideration that a period of at least 8 weeks of
conventional therapy setting (Crosbie et al., 2012; training is essential for adaptation (Lee, Park, and
Schuster-Amft et al., 2018). Park, 2019). It looks like there was not a significant
In our meta-analysis we noted a significant improve­ difference between both arms because chronic stroke
ment in UL motor function following VR-based rehabi­ patients took some time to adapt to VR programs.
litative intervention in patients with chronic stroke; For this the affected side needs training from three to
however, when compared to conventional therapy, the five times per week and lasts for 11 to 13 weeks for
difference was less significant. Several meta-analyses implying positive effects with VR as a rehabilitation
have been conducted to determine the effectiveness of approach (Levin, Weiss, and Keshner, 2015).
VR rehabilitation in stroke patients (Ahn and Hwang, Nevertheless, it is of critical importance to conduct
2019; Lee, Park, and Park, 2019; Mekbib et al., 2020). more studies with longer training periods with stan­
Both meta-analyses of Ahn and Hwang (2019) which dardized treatment protocols in order to further elu­
included 9 studies/698 patients and Mekbib et al. (2020) cidate this point.
 PHYSIOTHERAPY THEORY AND PRACTICE 2413

In our study, we conducted a regression analysis to based interventions in regaining UL motor function in
determine if some factors, such as age, the duration of chronic stroke patients.
treatment training, and post-stroke duration would have
affected our outcomes; however, we noted no significant
change. Consistent with our findings, Lee, Park, and Declaration of interest
Park (2019) analyzed the effect of VR rehabilitation The authors declare no conflict of interest.
intervention by treatment duration, and the authors
noted that longer durations were associated with stron­
ger effect, but such change did not achieve statistical Funding
significance. Since the number of included trials in our
meta-analysis was limited and the findings were not This work was supported by the Deanship of Scientific
Research at Princess Nourah bint Abdulrahman University,
statistically significant, we suggest that more clinical
Saudi Arabia, through the Research Groups Program [RGP-
trials are still warranted in order to reach a final conclu­ 1440-0012].
sion in this regard.

ORCID
Limitations
Wanees M. Badawy PhD, PT http://orcid.org/0000-0002-
Our study had several limitations, the most important of 9814-4162
which is the small number of included trials besides the
small sample size in each study ranging from 12–52
patients per study. Secondly, the variability in the VR- References
based rehabilitation protocols could have made it diffi­ Ackerley SJ, Byblow WD, Barber PA, MacDonald H,
cult to reach a statistically significant result and half of McIntyre-Robinson A, Stinear CM 2016 Primed physical
the included trials had treatment duration of 4 weeks or therapy enhances recovery of upper limb function in
less, which we think is not enough to reach adaptation. chronic stroke patients. Neurorehabilitation and Neural
Repair 30: 339–348. doi:10.1177/1545968315595285.
Thirdly, there were considerable concerns related to the Ahmad MA, Singh DK, Mohd Nordin NA, Hooi Nee K,
quality of included studies, in particular, deviation from Ibrahim N 2019 Virtual reality games as an adjunct in
intended intervention and selective reporting of results. improving upper limb function and general health among
Fourthly, the measurement tools of UL function were stroke survivors. International Journal of Environmental
widely variable among included studies, and thus, there Research and Public Health 16: 5144. doi:10.3390/
ijerph16245144.
should be a standardized criterion to determine
Ahn S, Hwang S 2019 Virtual rehabilitation of upper extremity
improvement in UL in future studies. Finally, we cannot function and independence for stroke: A meta-analysis.
conclude that VR-based intervention has superior effi­ Journal of Exercise Rehabilitation 15: 358–369.
cacy over conventional therapy in terms of regaining the doi:10.12965/jer.1938174.087.
functional state of the upper limbs of patients with Arya KN, Pandian S, Verma R, Garg RK 2011 Movement
chronic stroke based on the aforementioned limitations. therapy induced neural reorganization and motor recovery
in stroke: A review. Journal of Bodywork and Movement
Therefore, we well-designed RCTs with larger sample Therapies 15: 528–537. doi:10.1016/j.jbmt.2011.01.023.
sizes, longer training duration, and standardized proto­ Bailey RR, Klaesner JW, Lang CE 2015 Quantifying real-world
col are still warranted. upper-limb activity in nondisabled adults and adults with
chronic stroke. Neurorehabilitation and Neural Repair 29:
969–978. doi:10.1177/1545968315583720.
Baumann M, Lurbe-Puerto K, Alzahouri K, Aïach P 2011
Conclusion
Increased residual disability among poststroke survivors
In this systematic review and meta-analysis, we noted and the repercussions for the lives of informal caregivers.
a significant improvement in upper limb function of Topics in Stroke Rehabilitation 18: 162–171. doi:10.1310/
tsr1802-162.
patients with chronic stroke who underwent VR-based Brewer L, Horgan F, Hickey A, Williams D 2013 Stroke reha­
rehabilitative intervention compared to the pre- bilitation: Recent advances and future therapies. Monthly
treatment state. Our analysis also revealed no superiority Journal of the Association of Physicians 106: 11–25.
of VR interventions over conservative therapies. Based doi:10.1093/qjmed/hcs174.
on the limitations in the current evidence, we recom­ Cambier DC, De Corte E, Danneels LA, Witvrouw EE 2003
Treating sensory impairments in the post-stroke upper limb
mend conducting more robust randomized controlled
with intermittent pneumatic compression. Results of
trials of larger sample sizes and longer training duration a preliminary trial. Clinical Rehabilitation 17: 14–20.
to reach a final conclusion about the effectiveness of VR- doi:10.1191/0269215503cr580oa.
2414 R. M. AL-WHAIBI ET AL.

Chen JC, Liang CC, Shaw FZ 2005 Facilitation of sensory and review with meta-analysis. Topics in Stroke Rehabilitation
motor recovery by thermal intervention for the hemiplegic 23: 440–457. doi:10.1080/10749357.2016.1183349.
upper limb in acute stroke patients: A single-blind rando­ Global of Burden Disease 2016 Stroke Collaborators 2019
mized clinical trial. Stroke 36: 2665–2669. doi:10.1161/01. Global, regional, and national burden of stroke,
STR.0000189992.06654.ab. 1990-2016: A systematic analysis for the Global Burden of
Chen JC, Shaw FZ 2014 Progress in sensorimotor rehabilita­ Disease Study 2016. Lancet Neurology 18: 439–458.
tive physical therapy programs for stroke patients. World Hedges LV, Pustejovsky JE, Shadish WR 2013 A standardized
Journal of Clinical Cases 16: 316–326. doi:10.12998/wjcc.v2. mean difference effect size for multiple baseline designs
i8.316. across individuals. Research Synthesis Methods 4:
Chen L, Lo WL, Mao YR, Ding MH, Lin Q, Li H, Zhao JL, 324–341. doi:10.1002/jrsm.1086.
Xu ZQ, Bian RH, Huang DF 2016 Effect of virtual reality on Higgins J, Thomas J, Chandler J, Cumpston M, Li T, Page M,
postural and balance control in patients with stroke: Welch V (Eds) 2019 Cochrane Handbook for Systematic
A systematic literature review. BioMed Research Reviews of Interventions version 6.0. www.training.
International 2016: 7309272. doi:10.1155/2016/7309272. cochrane.org/handbook.
Choo PL, Gallagher HL, Morris J, Pomeroy VM, van Wijck F Hosp JA, Luft AR 2011 Cortical plasticity during motor learn­
2015 Correlations between arm motor behavior and brain ing and recovery after ischemic stroke. Neural Plasticity
function following bilateral arm training after stroke: 2011: 871296. doi:10.1155/2011/871296.
A systematic review. Brain and Behavior 5: e00411. Hsu WY, Cheng CH, Liao KK, Lee IH, Lin YY 2012 Effects of
doi:10.1002/brb3.411. repetitive transcranial magnetic stimulation on motor func­
Crosbie JH, Lennon S, McGoldrick MC, McNeill MD, tions in patients with stroke: A meta-analysis. Stroke 43:
McDonough SM 2012 Virtual reality in the rehabilitation 1849–1857. doi:10.1161/STROKEAHA.111.649756.
of the arm after hemiplegic stroke: A randomized con­ Iruthayarajah J, McIntyre A, Cotoi A, Macaluso S, Teasell R
trolled pilot study. Clinical Rehabilitation 26: 798–806. 2017 The use of virtual reality for balance among indivi­
doi:10.1177/0269215511434575. duals with chronic stroke: A systematic review and
da Silva Cameirão M, Bermúdez I, Badia S, Duarte E, meta-analysis. Topic in Stroke Rehabilitation 24: 68–79.
Verschure PF 2011 Virtual reality based rehabilitation doi:10.1080/10749357.2016.1192361.
speeds up functional recovery of the upper extremities Jarvis K, Reid G, Edelstyn N, Hunter S 2014 Development of
after stroke: A randomized controlled pilot study in the the occupational therapy stroke arm and hand record: An
acute phase of stroke using the rehabilitation gaming upper limb treatment schedule. British Journal of
system. Restorative Neurology and Neuroscience 29: Occupational Therapy 77: 126–133. doi:10.4276/
287–298. doi:10.3233/RNN-2011-0599. 030802214X13941036266469.
da Silva Ribeiro NM, Ferraz DD, Pedreira É, Pinheiro Í, Da Johansson BB 2011 Current trends in stroke rehabilitation.
Silva Pinto AC, Neto MG, Dos Santos LR, Pozzato MG, A review with focus on brain plasticity. Acta Neurologica
Pinho RS, Masruha MR 2015 Virtual rehabilitation via Scandinavica 123: 147–159. doi:10.1111/j.1600-
Nintendo Wii® and conventional physical therapy effec­ 0404.2010.01417.x.
tively treat post-stroke hemiparetic patients. Topics in Kim JH 2018 Effects of a virtual reality video game exercise
Stroke Rehabilitation 22: 299–305. doi:10.1179/ program on upper extremity function and daily living activ­
1074935714Z.0000000017. ities in stroke patients. Journal of Physical Therapy Science
Dobkin BH, Dorsch A 2013 New evidence for therapies in 30: 1408–1411. doi:10.1589/jpts.30.1408.
stroke rehabilitation. Current Atherosclerosis Reports 15: Kita K, Otaka Y, Takeda K, Sakata S, Ushiba J, Kondo K,
331. doi:10.1007/s11883-013-0331-y. Liu M, Osu R 2013 A pilot study of sensory feedback by
Doyle S, Bennett S, Fasoli SE, McKenna KT 2010 Interventions transcutaneous electrical nerve stimulation to improve
for sensory impairment in the upper limb after stroke. manipulation deficit caused by severe sensory loss after
Cochrane Database of Systematic Reviews 6: CD006331. stroke. Journal of NeuroEngineering and Rehabilitation
Egger M, Davey Smith G, Schneider M, Minder C 1997 Bias in 10: 55. doi:10.1186/1743-0003-10-55.
meta-analysis detected by a simple, graphical test. British Kleim JA, Jones TA 2008 Principles of
Medical Journal 315: 629–634. doi:10.1136/ experience-dependent neural plasticity: Implications for
bmj.315.7109.629. rehabilitation after brain damage. Journal of Speech,
El-Qushayri AE, Khalaf KM, Dahy A, Mahmoud AR, Language, and Hearing Research 51: S225–S239.
Benmelouka AY, Ghozy S, Mahmoud MU, Bin-Jumah M, doi:10.1044/1092-4388(2008/018).
Alkahtani S, Abdel-Daim MM 2020 Fournier’s gangrene Kraemer WJ, Koziris LP, Ratamess NA, Hakkinen K, Triplett-
mortality: A 17-year systematic review and meta-analysis. McBride NT, Fry AC, Gordon SE, Volek JS, French DN,
International Journal of Infectious Diseases 92: 218–225. Rubin MR, et al. 2002 Detraining produces minimal
doi:10.1016/j.ijid.2019.12.030. changes in physical performance and hormonal variables
French B, Thomas LH, Leathley MJ, Sutton CJ, McAdam J, in recreationally strength-trained men. Journal of Strength
Forster A, Langhorne P, Price CI, Walker A, Watkins CL and Conditioning Research 16: 373–382.
2007 Repetitive task training for improving functional abil­ Langhorne P, Bernhardt J, Kwakkel G 2011 Stroke
ity after stroke. Cochrane Database Systematic Review 4: rehabilitation. Lancet 377: 1693–1702. doi:10.1016/S0140-
CD006073. 6736(11)60325-5.
Gibbons EM, Thomson AN, de Noronha M, Joseph S 2016 Laver KE, George S, Thomas S, Deutsch JE, Crotty M 2015
Are virtual reality technologies effective in improving lower Virtual reality for stroke rehabilitation. Cochrane Database
limb outcomes for patients following stroke - A systematic of Systematic Reviews 2015: CD008349.
 PHYSIOTHERAPY THEORY AND PRACTICE 2415

Lee HS, Park YJ, Park SW 2019 The effects of virtual reality controlled trial. Lancet Neurology 15: 1019–1027.
training on function in chronic stroke patients: doi:10.1016/S1474-4422(16)30121-1.
A systematic review and meta-analysis. BioMed Research Saposnik G, Levin M, Outcome Research Canada (SORCan)
International 2019: 7595639. doi:10.1155/2019/7595639. Working Group 2011 Virtual reality in stroke rehabilita­
Levin MF, Weiss PL, Keshner EA 2015 Emergence of virtual tion: A meta-analysis and implications for clinicians. Stroke
reality as a tool for upper limb rehabilitation: Incorporation 42: 1380–1386. doi:10.1161/STROKEAHA.110.605451.
of motor control and motor learning principles. Physical Schuster-Amft C, Eng K, Suica Z, Thaler I, Signer S,
Therapy 95: 415–425. doi:10.2522/ptj.20130579. Lehmann I, Schmid L, McCaskey MA, Hawkins M,
Lohse KR, Hilderman CG, Cheung KL, Tatla S, Van Der Verra ML, et al. 2018 Effect of a four-week virtual
Loos HF 2014 Virtual reality therapy for adults reality-based training versus conventional therapy on
post-stroke: A systematic review and meta-analysis explor­ upper limb motor function after stroke: A multicenter par­
ing virtual environments and commercial games in therapy. allel group randomized trial. Public Library of Science One
Public Library of Science One 9: e93318. doi:10.1371/jour­ 13: e0204455.
nal.pone.0093318. Simpkins AN, Janowski M, Oz HS, Roberts J, Bix G, Doré S,
Mekbib DB, Han J, Zhang L, Fang S, Jiang H, Zhu J, Roe AW, Stowe AM 2020 Biomarker application for precision med­
Xu D 2020 Virtual reality therapy for upper limb rehabilita­ icine in stroke. Translational Stroke Research 11: 615–627.
tion in patients with stroke: A meta-analysis of randomized doi:10.1007/s12975-019-00762-3.
clinical trials. Brain Injury 34: 456–465. doi:10.1080/ Sterne JA, Savović J, Page MJ, Elbers RG, Blencowe NS,
02699052.2020.1725126. Boutron I, Cates CJ, Cheng HY, Corbett MS, Eldridge SM,
Merians AS, Jack D, Boian R, Tremaine M, Burdea GC, et al. 2019 RoB 2: A revised tool for assessing risk of bias in
Adamovich SV, Recce M, Poizner H 2002 Virtual randomised trials. British Medical Journal 366: l4898.
reality-augmented rehabilitation for patients following doi:10.1136/bmj.l4898.
stroke. Physical Therapy 82: 898–915. doi:10.1093/ptj/ Stockley RC, O’Connor DA, Smith P, Moss S, Allsop L,
82.9.898. Edge W 2017 A mixed methods small pilot study to describe
Molier BI, Van Asseldonk EH, Hermens HJ, Jannink MJ 2010 the effects of upper limb training using a virtual reality
Nature, timing, frequency and type of augmented feedback; gaming system in people with chronic stroke.
does it influence motor relearning of the hemiparetic arm Rehabilitation Research and Practice 2017: 9569178.
after stroke: A systematic review. Disability and doi:10.1155/2017/9569178.
Rehabilitation 32: 1799–1809. doi:10.3109/ Subramanian SK, Lourenço CB, Chilingaryan G, Sveistrup H,
09638281003734359. Levin MF 2013 Arm motor recovery using a virtual reality
Molley AG 1990 Applying effectiveness and outcomes intervention in chronic stroke: Randomized control trial.
research to clinical practice. In: Heithoff KA, Lohr K (Eds) Neurorehabililitaion and Neural Repair 27: 13–23.
Institute of Medicine (US) Division of Health Care Services doi:10.1177/1545968312449695.
- Effectiveness and Outcomes in Health Care: Proceedings Takeshima N, Sozu T, Tajika A, Ogawa Y, Hayasaka Y,
of an Invitational Conference, p. 179–189. Washington, Furukawa TA 2014 Which is more generalizable, powerful
DC: National Academies Press. and interpretable in meta-analyses, mean difference or stan­
Nijland R, Kwakkel G, Bakers J, van Wegen E 2011 dardized mean difference? BioMed Central Medical Research
Constraint-induced movement therapy for the upper pare­ Methodology 14: 30. doi:10.1186/1471-2288-14-30.
tic limb in acute or sub-acute stroke: A systematic review. Takeuchi N, Izumi S 2012a Maladaptive plasticity for motor
International Journal of Stroke 6: 425–433. doi:10.1111/ recovery after stroke: Mechanisms and approaches. Neural
j.1747-4949.2011.00646.x. Plasticity 2012: 359728. doi:10.1155/2012/359728.
Peters JL, Sutton AJ, Jones DR, Abrams KR, Rushton L 2006 Takeuchi N, Izumi S 2012b Noninvasive brain stimulation for
Comparison of two methods to detect publication bias in motor recovery after stroke: Mechanisms and future views.
meta-analysis. JAMA 295: 676–680. doi:10.1001/ Stroke Research and Treatment 2012: 584727. doi:10.1155/
jama.295.6.676. 2012/584727.
Pollock A, Baer G, Langhorne P, Pomeroy V 2007 Talelli P, Greenwood RJ, Rothwell JC 2006 Arm function after
Physiotherapy treatment approaches for the recovery of stroke: Neurophysiological correlates and recovery
postural control and lower limb function following stroke: mechanisms assessed by transcranial magnetic
A systematic review. Clinical Rehabilitation 21: 395–410. stimulation. Clinical Neurophysiology 117: 1641–1659.
Roger VL, Go AS, Lloyd-Jones DM, Benjamin EJ, Berry JD, doi:10.1016/j.clinph.2006.01.016.
Borden WB, Bravata DM, Dai S, Ford ES, Fox CS, et al. 2012 Tinga AM, Visser-Meily JM, van der Smagt MJ, Van Der
Heart disease and stroke statistics - 2012 update: A report Stigchel S, van Ee R, Nijboer TC 2016 Multisensory stimu­
from the American Heart Association. Circulation 125: e2– lation to improve low- and higher-level sensory deficits
e220. doi:10.1161/CIR.0b013e31823ac046. after stroke: A systematic review. Neuropsychology
Saposnik G, Cohen LG, Mamdani M, Pooyania S, Review 26: 73–91. doi:10.1007/s11065-015-9301-1.
Ploughman M, Cheung D, Shaw J, Hall J, Nord P, Vassar M, Atakpo P, Kash MJ 2016 Manual search approaches
Dukelow S, et al. 2016 Efficacy and safety of used by systematic reviewers in dermatology. Journal of the
non-immersive virtual reality exercising in stroke rehabili­ Medical Library Association 104: 302–304. doi:10.3163/
tation (EVREST): A randomised, multicenter, single-blind, 1536-5050.104.4.009.
2416 R. M. AL-WHAIBI ET AL.

Veerbeek JM, van Wegen E, van Peppen R, van der Wees PJ, Wu HC, Lin YC, Hsu MJ, Liu SM, Hsieh CL, Lin JH 2010
Hendriks E, Rietberg M, Kwakkel G 2014 What is the Effect of thermal stimulation on upper extremity motor
evidence for physical therapy post stroke? A systematic recovery 3 months after stroke. Stroke 41: 2378–2380.
review and meta-analysis. PLoS One 9: e87987. doi:10.1161/STROKEAHA.110.593673.
Wechsler LR, Bates D, Stroemer P, Andrews-Zwilling YS, Zorowitz R, Brainin M 2011 Advances in brain recovery and
Aizman I 2018 Cell therapy for chronic stroke. Stroke 49: rehabilitation 2010. Stroke 42: 294–297. doi:10.1161/
1066–1074. doi:10.1161/STROKEAHA.117.018290. STROKEAHA.110.605063.
Copyright of Physiotherapy Theory & Practice is the property of Taylor & Francis Ltd and its
content may not be copied or emailed to multiple sites or posted to a listserv without the
copyright holder's express written permission. However, users may print, download, or email
articles for individual use.