Effectiveness of Virtual Reality in Children With Cerebral Palsy: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

Probably not a substantial claim: may not pass S3 (load-bearing).

The database searches identified 418 nonoverlapping published articles.

“A total of 418 nonoverlapping published articles were found from the database searches.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The nineteen included studies had 504 participants in total.

“There were 504 participants in the 19 studies.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

When outcomes were classified by the ICF model, nine studies had activity-only outcomes, seven had body structure plus activity outcomes, two had activity plus participation outcomes, and one had outcomes in all components.

“When classifying outcome measures based on the ICF model (Tab.2), the majority (9 studies) included outcome in the activity component only, 7 studies contained body structure and activity components, 2 studies contained activity and participation components, and 1 contained all components.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Across the included studies, total VR intervention duration ranged from 8 to 80 hours, daily dosage from 20 to 90 minutes, intervention length from 4 to 20 weeks, and frequency from once weekly to seven days weekly.

“VR intervention over the entire study ranged from 8 to 80 hours, and the daily dosage ranged from 20 minutes to 90 minutes.Intervention length ranged from 4 weeks to 20 weeks, and treatment frequency ranged from once per week to 7 d/wk.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The average age of participants across the included studies ranged from 4.6 to 12.1 years.

“The average age of participants ranged from 4.6 to 12.1 years.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

VR intervention was delivered mostly in clinics (nine studies), followed by laboratories (six studies) and homes (four studies).

“VR intervention was offered mostly in clinics (9 studies), followed by in the laboratory (6 studies) and home (4 studies).”

May combine two claims.

PEDro scores of the included randomized trials ranged from 4 to 8, indicating fair to good methodological quality.

“The PEDro quality scores for the studies ranged from 4 to 8 (4 studies with a score of 4, 4 studies with a score of 5, 3 studies with a score of 6, 3 studies with a score of 7, and 5 studies with a score of 8), indicating that the quality of the RCTs included in this meta-analysis was fair to good.”

The quote only partly states this claim.

Across all included studies, VR intervention had a strong effect on motor function in children with CP compared with conventional therapy or controls, with an effect size of 0.861 and a 95% confidence interval of 0.51 to 1.22.

“Across all studies, there was a strong effect (d = 0.861; 95% CI = 0.51-1.22) of VR intervention in children with CP.”

• Also stated in: § Discussion · boosted

  “In general, when combining all outcome measures of all studies, virtual reality intervention showed a strong effect in improving motor function in children with CP when comparing with conventional therapy or controls (d = 0.861).”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Nine included studies had a matched-dose comparison group, while ten did not.

“Nine studies had a matched-dose comparison group, whereas 10 studies did not.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Nintendo Wii was the most commonly used VR system, used in nine of the nineteen studies.

“The mostly commonly used VR system was Nintendo Wii (9/19 studies).”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The included studies focused on arm function only in eight studies, postural control only in two, postural control plus ambulation in four, postural control plus arm function in one, arm function plus ambulation in one, and all three domains in four.

“The focus of the outcome measures included arm function only (8 studies); postural control only (2 studies); combination of postural control and ambulation (4 studies); combination of postural control and arm function (1 study); combination of arm function and ambulation (1 study); and combination of arm function, postural control, and ambulation (4 studies).”

May combine two claims.

The overall meta-analysis showed large heterogeneity, with an I-squared value of 70% and Q of 60.36.

“Not surprisingly, we also found large heterogeneity among the studies as the value of I 2 was 70% (Q = 60.36;P <.000001).”

May combine two claims.

The postural control meta-analysis showed large heterogeneity, with an I-squared value of 67% and Q of 27.50.

“We found large heterogeneity among the studies as the value of I 2 was 67% (Q = 27.50;P = .001).”

May combine two claims.

For ambulation, the overall effect size was 0.755 with a 95% confidence interval of 0.348 to 1.161, indicating a medium to large effect.

“The overall effect size was 0.755 (95% CI = 0.348-1.161),indicating a medium to large effect size (Fig. 3b).”

May combine two claims.

For postural control, the overall effect size was 1.003 with a 95% confidence interval of 0.503 to 1.502, indicating a large effect.

“The overall effect size was 1.003 (95% CI = 0.503-1.502),indicating a large effect size (Fig. 3c).”

May combine two claims.

For ambulation, studies including children with diplegia or mixed CP type had better effects than studies including only children with hemiplegia, with effect sizes of 1.064, 0.883, and 0.249 respectively.

“Studies that included children with diplegia or mixed type seemed to have a better effect than studies that included only children with hemiplegia (d = 1.064 for diplegia, d = 0.883 for mixed type, and d = 0.249 for hemiplegia).”

May combine two claims.

The overall funnel plot showed minor asymmetry, and the trim-and-fill correction shifted the overall effect size from 0.861 to 0.724.

“Minor asymmetry was noted in the plot, and the Duval-Tweedie trim and fill correction was used to correct the overall effect size.This correction shifted the overall effect size from 0.861 to 0.724.”

May combine two claims.

For ambulation, publication bias was found, and trim-and-fill correction shifted the effect size from 0.755 to 0.378.

“Publication bias was found, and the Duval-Tweedie trim and fill correction was used.The effect size on ambulation shifted from 0.755 to 0.378.”

May combine two claims.

For arm function, VR system, children's age, and daily VR dosage were significant moderators.

“For the subgroups and meta-regression analyses, VR system, children's age, and VR daily dosage were shown to be significant factors (P = .03for VR system, P = .01for children's age, and P = .04for daily dosage).”

May combine two claims.

For arm function, minor publication bias asymmetry was noted, and trim-and-fill correction shifted the effect size from 0.835 to 0.691.

“Publication bias was also assessed, and minor asymmetry was noted.After using the Duval-Tweedie trim and fill correction, we found that the overall effect size shifted from 0.835 to 0.691.”

May combine two claims.

The arm function meta-analysis showed large heterogeneity, with an I-squared value of 75% and Q of 47.82.

“We found large heterogeneity among the studies, as the value of I 2 was 75% (Q = 47.82;P < .000001).”

May combine two claims. The quote only partly states this claim.

For arm function, there was a statistically significant negative linear relationship between age and effect size, with a slope of -0.1307, meaning younger children had larger effects.

“Meta-regression analyses also showed a statistically significant negative linear relationship between age and study effect size: the younger the children were, the larger the effect”

May combine two claims.

For ambulation, there were statistically significant negative linear relationships between age and effect size and between PEDro score and effect size: younger children and lower PEDro scores were associated with larger effects.

“Meta-regression analyses showed a statistically significant negative linear relationship between age and study effect size, and between PEDro score and study effect size: the younger the children's age was, the larger the effect size; and the lower the PEDro score was, the larger the effect size.”

May combine two claims.

For ambulation, CP type, children's age, and PEDro score were significant moderators.

“For the subgroups and meta-regression analyses, CP type, children's age, and PEDro score were shown to be significant factors (P = .02for CP type, P = .037for children's age, and P = .0002for PEDro score).”

May combine two claims.

The ambulation meta-analysis showed medium heterogeneity, with an I-squared value of 59% and Q of 17.01.

“We found medium heterogeneity among the studies, as the value of I 2 was 59% (Q = 17.01;P = .017).”

May combine two claims.

In the overall analysis, only VR system was a significant moderator, with engineer-built systems appearing more effective than commercially available systems (effect sizes 1.572 versus 0.628).

“Only”

May combine two claims.

Most VR systems used in the included studies provided more than one game with similar training goals, and children with CP in these studies were encouraged to choose which games to play or at least to discuss game choice with their therapists.

“Most of the VR system can provide more than 1 game with similar training goals and children with CP in these studies were encouraged to choose which games they wanted to play (or at least had the chance to discuss with their therapists which games to choose).”

May combine two claims.

The association between VR effect and treatment dosage was significant only for arm function, with larger daily dosage producing larger effects.

“Interestingly, the association between VR effect and treatment dosage was only significant in arm function: the larger the daily dosage, the larger the VR effect.”

Classified as probably not a substantial claim.

The meta-analyses used a random-effects model that accounted for between-study variation and within-study random error.

“Meta-analyses were run with a random-effects model that accounted for true interstudy variation in effects as well as for random error within studies.”

May combine two claims.

Dewar et al. found conflicting evidence on whether VR improves postural control in children with CP: two studies showed improvement in standing balance and one showed no improvement in functional standing balance.

“This study showed a conflicting result regarding whether VR could enhance postural control in children with CP: 2 studies showed improvement in standing balance and 1 showed no improvement in functional standing balance.18”

• cites [18] Dewar, 2015, Exercise interventions improve postural control in children with cerebral palsy: a systematic review, Dev Med Child Neurol sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Bonnechère et al. concluded that comparison across studies was difficult because rehabilitation strategies were not standardized and clinical assessment tools differed.

“The authors concluded it was difficult to compare the different studies because of the lack of standardized rehabilitation strategies and different clinical assessment tools.19”

• cites [19] Bonnechère, 2014, Can serious games be incorporated with conventional treatment of children with cerebral palsy? A review, Res Dev Disabil sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The methodological quality of the included randomized trials was assessed with the PEDro scale.

“The quality of RCTs was evaluated using the Physiotherapy Evidence Data-base (PEDro scale).25”

• cites [25] Macedo, 2010, There was evidence of convergent and construct validity of Physiotherapy Evidence Database quality scale for physiotherapy trials, J Clin Epidemiol sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The authors conducted the systematic review in accordance with the PRISMA statement.

“We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement to conduct this review.23,24”

• cites [23] Moher, 2009, Preferred Reporting Items for Systematic Reviews and Meta-Analyses: the PRISMA statement, PLoS Med sourceNot checked yet

  Not checked in the free run.

• cites [24] Liberati, 2009, The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration, PLoS Med sourceNot checked yet

  Not checked in the free run.

May combine two claims.

The authors' earlier subgroup analyses found better effects for younger children, for home- or laboratory-based VR, and for engineer-built VR systems.

“In addition, the subgroup analyses showed younger children receiving home or laboratory-based VR and using an engineer-built VR system had a better effect.16”

• cites [16] Chen, 2014, Effect of virtual reality on imporving upper-extremity function in children with cerebral palsy: a meta-analysis., Pediatr Phys Ther sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

One author extracted and coded the data, a second reviewer checked all extracted data, and disagreements were resolved by discussion to consensus.

“Relevant information from the included studies was extracted and coded by the same author (Y.C.).A second reviewer (H.D.F.) checked all the extracted data.Any discrepancies were resolved by discussion in order to reach a consensus.”

May combine two claims.

In the authors' earlier meta-analysis, ICF-based effects were large for participation, small for activity, and medium for body structure and function.

“When the outcome variables were further broken down according to the ICF model, a large effect was reported in participation (d = 1.92), a small effect on activity (d = 0.46), and a medium effect on body structure and function (d = 0.70).16”

• cites [16] Chen, 2014, Effect of virtual reality on imporving upper-extremity function in children with cerebral palsy: a meta-analysis., Pediatr Phys Ther sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

A medium effect size of 0.5 corresponds to a change that a careful observer would notice.

“Clinically, a medium effect size (d = 0.5) corresponds to an amount of change that is noticeable to a careful observer.[28][29][30]”

• cites [28] Mann, 2012, Quantifying clinically significant change: a brief review of methods and presentation of a hybrid approach, Am J Sports Med sourceNot checked yet

  Not checked in the free run.

• cites [29] Eisen, 2007, Measuring clinically meaningful change following mental health treatment, J Behav Health Serv Res sourceNot checked yet

  Not checked in the free run.

• cites [30] Norman, 2003, Interpretation of change in health-related quality of life: the remarkable universality of half a standard deviation, Med Care sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Publication bias in the primary meta-analyses was assessed with a funnel plot and the Duval-Tweedie trim-and-fill correction.

“The effect of publication bias on the primary meta-analyses was addressed by combining a funnel plot assessment with the Duval-Tweedie trim and fill correction.22”

• cites [22] Borenstein, 2009 sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Two authors independently searched six electronic databases and reference lists in December 2016.

“Two authors (Y.C., H.D.F.) independently conducted systematic literature searches in December 2016, using the electronic databases: PubMed, CINAHL, Cochrane Central Register of Controlled Trials, ERIC, PsycINFO, and Web of Science, as well as a manual search of the reference lists of each article.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Levac et al.'s scoping review suggested that VR helps children with CP improve motor skills through user-side factors, VR system or game properties, and personal support.

“Recently, Levac et al 11 in a scoping review suggested some possible reasons from VR therapy that help children with CP improve their motor skills including from the user's side (VR enhanced problem-solving and cognitive engagement during play and increased motivation and neuroplasticity changes);”

• cites [11] Levac, 2012, Defining the active ingredients of interactive computer play interventions for children with neuromotor impairments: a scoping review, Res Dev Disabil sourceNot checked yet

  Not checked in the free run.

May combine two claims.

The active elements of VR can change personal and environmental barriers, and by reducing barriers or enhancing enablers, children with CP may reduce impairments and activity limitations and improve participation.

“All these active elements from VR help to change personal and environmental barriers that a child with CP may face.By decreasing these barriers or enhancing the enablers, a child with CP may gradually decrease impairments of his/her body structure and function and activity limitations, and gradually improve participation in school, communities, and society.11”

• cites [11] Levac, 2012, Defining the active ingredients of interactive computer play interventions for children with neuromotor impairments: a scoping review, Res Dev Disabil sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Learning is optimized when learners repeat tasks and also cognitively engage in solving the motor task.

“Learning would be optimized if learners not only repeat the tasks alone but also cognitively engage in problem solving the motor task.12”

• cites [12] Levac, 2014, Motor learning and virtual reality sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Children work hard to achieve higher game scores during VR training, which facilitates cognitive engagement in solving motor tasks and enhances motor learning.

“For example, children will work hard to achieve a higher game score when receiving training in VR, which facilitate children's cognitive engagement in solving the motor task by 'thinking through' how to make their movements more efficiently and faster, which enhance their motor learning.12”

• cites [12] Levac, 2014, Motor learning and virtual reality sourceNot checked yet

  Not checked in the free run.

May combine two claims.

VR games usually have attainable goals that can be progressed and that enhance children's attention in therapy.

“VR games usually have goals to attain that can be progressed and enhance children's attention in therapy.[11][12][13]”

• cites [11] Levac, 2012, Defining the active ingredients of interactive computer play interventions for children with neuromotor impairments: a scoping review, Res Dev Disabil sourceNot checked yet

  Not checked in the free run.

• cites [12] Levac, 2014, Motor learning and virtual reality sourceNot checked yet

  Not checked in the free run.

• cites [13] Weiss, 2006, Virtual reality applications in neurorehabilitation sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The mechanism by which VR produced its effects remained uncertain.

“The possible mechanism for why VR worked remained uncertain.”

May combine two claims.

The flexibility of many VR applications can make training ecologically meaningful in an enriched environment and give children opportunities to cognitively engage in solving motor tasks.

“The flexibility of many VR applications offers the opportunities to ensure that training is ecologically meaningful in an enriched environment, which provide the opportunity for children to cognitively engage in problem solving their motor task.12,13”

• cites [12] Levac, 2014, Motor learning and virtual reality sourceNot checked yet

  Not checked in the free run.

• cites [13] Weiss, 2006, Virtual reality applications in neurorehabilitation sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Children with CP performed better with VR than with conventional intervention, which may focus more on internal attention cues.

“Consequently, children with CP performed better with VR than with the conventional intervention in which therapists might focus more on internal attention cues by requesting the children to move their arms faster or to open their hands wider.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Lohse et al found a small-to-medium VR effect on activity in adults with stroke (G=0.58).

“G = 0.58 for activity level”

• cites Lohse et al 67Not checked yet

  Not checked in the free run.

May combine two claims.

Game selection supports children's autonomy, enhances motivation, increases task interest, and leads to better learning by giving learners control over practice conditions.

“This kind of game selection supports children's need for autonomy, enhances their motivation, increases task interest, and leads to a better learning as this grants learners control over their practice conditions.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Wulf and Lewthwaite suggested that supporting a learner's autonomy through choices and enhanced expectancies can optimize learning by increasing motivation.

“Wulf and Lewthwaite 64 recently suggested supporting a learner's autonomy by providing choices and enhanced expectancies (eg, positive feedback, reduced perceived task difficulty) could optimize learning through increased motivation as autonomy and enhanced expectancies allow learners to feel”

• cites [64] Wulf, 2016, Optimizing performance through intrinsic motivation and attention for learning: the OPTIMAL theory of motor learning, Psychon Bull Rev sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Lohse et al found a small-to-medium VR effect on participation in adults with stroke (G=0.56).

“G = 0.56 for participation level”

• cites Lohse et al 67Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

In VR, a child with CP does not need full finger flexion to pick up a virtual object; a small range such as 5 degrees can allow the pick-up and reduce perceived task difficulty.

“For example, in order for a child with CP to pick up a virtual object in the virtual environment, the child does not need to have a full range of finger flexion.Instead, a small range (eg, 5°) can easily allow the child to pick up an object in the virtual world, which reduces perceived task difficulty.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Lohse et al found a small-to-medium VR effect on body structure/function in adults with stroke (G=0.48).

“G = 0.48 for body structure and function”

• cites Lohse et al 67Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Wulf and Lewthwaite proposed that learners given external attention cues learn faster and perform better than learners given internal attention cues.

“This was consistent with the proposal of Wulf and Lewthwaite 64 : when the instructions to learners prompted them to direct their attention to an external focus (eg, minimize movements of a balance platform) rather than to direct their attention to an internal focus (eg, try not to move your feet), learners in the external attention cue”

• cites [64] Wulf, 2016, Optimizing performance through intrinsic motivation and attention for learning: the OPTIMAL theory of motor learning, Psychon Bull Rev sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Children are more motivated and engaged in video games than adults.

“have stronger motivation and are more engaged in playing video games”

• cites [69] The NPD Group, Inc, 2017Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Included studies used various intervention protocols.

“and various intervention protocols.”

May combine two claims.

Commercial systems are not designed to train children with CP because their preset game difficulties are restricted and require adaptation.

“Commercial systems, on the other hand, are not designed to train children (eg, gradually increase task difficulty, changing task demands and contexts) with CP because these systems have more restrictions on their preset game difficulties and require some adaptations and careful game selections prior to use.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The age association is consistent with early intervention because younger children may have more brain plasticity and adaptability.

“This was consistent with the idea of early intervention as younger children may have more brain plasticity and adaptability to improve their motor function than children at older ages.71,72”

• cites [71] Morgan, 2013, Enriched environments and motor outcomes in cerebral palsy: systematic review and meta-analysis., Pediatrics sourceNot checked yet

  Not checked in the free run.

• cites [72] Spittle, 2012, Early developmental intervention programmes post-hospital discharge to prevent motor and cognitive impairments in preterm infants, Cochrane Database Syst RevNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

Weiss et al's model describes VR rehabilitation as three nested spaces: interaction space, transfer phase, and real world.

“In this model, there were 3 nested circles, the inner”

• cites Weiss et al 13Not checked yet

  Not checked in the free run.

May combine two claims.

VR is a viable intervention to improve arm function, ambulation, and postural control in children with CP.

“In conclusion, VR is a viable intervention to improve arm function, ambulation, and postural control in children with CP.”