Cerebral palsy (CP) is the leading cause of childhood disability and has a profound effect on physical function.1 About 1 in 303 children in the United States are diagnosed with CP, and half of these children have impaired upper extremity (UE) function (eg, reaching, grasping, and manipulating objects).2–4 Recently, virtual reality (VR) has been explored as a training tool that is used by therapists for improving motor performance in children with CP.5,6 Virtual reality is defined as “the use of interactive simulations created with computer hardware and software to present users with opportunities to engage in environments that appear to be and feel similar to real world objects and event.”7(p183) The use of VR by trained therapists enables the creation of an exercise environment in which children with CP can practice intensely and simultaneously receive positive visual and auditory feedback.5,6,8 Virtual reality also can provide a comparison between the degree of movement performed by children in the real world and the degree of movement they observed in the virtual environment.7 Such a spatial representation allows for real-time visual feedback relating to knowledge of performance, which is crucial for children with CP in learning motor skills.7 All these features make VR a potentially viable tool to train motor skills in children with CP.
Few review articles recently published examine the effect of VR on improving motor function in children with disabilities (Table 1).5,6,9–11 Two studies,6,10 although focused on children with CP, did not focus on UE function nor report effect size. Snider et al6 examined the effect of VR on upper limb, lower limb, visual-spatial abilities, playfulness, and social acceptance in children with CP and found that 12 of the 13 studies included in their systematic review reported at least 1 outcome measure improved after VR intervention. Galvin et al10 systematically reviewed 5 studies (1 randomized controlled trial [RCT] design and 4 case series) on UE function in children with CP and showed that the use of VR is at an “emerging” stage because all research, per Gavin et al's conclusions, had methodological limitations, such as small sample sizes and inconsistencies in outcome measurements. However, they did not report effect sizes for VR, nor were outcome measurements classified per the International Classification of Functioning, Disability, and Health (ICF). In addition, reports of 2 studies using an RCT design and 6 case series were published after the databases were accessed by Gavin et al, which might influence any subsequent conclusion. In this article, we strengthen the systematic review by (1) adding more research studies, (2) applying the ICF model to classify outcome variables, (3) quantifying the effect size (Cohen's d) between pre-VR and post-VR treatments as well as between VR and conventional therapy; (4) focusing only on children with CP, and (5) identifying the association between the VR effect and key characteristics of the child as well as the intervention protocol by examining the effect of moderators on VR effect size.
Data Sources and Study Selection
The inclusion criteria for this review were the following: (1) participants were children with CP; (2) intervention used in the study involved VR technology with a focus on UE function; (3) outcome measurements included UE movements such as reaching, grasping, or UE function measured by a general fine motor assessment scale (such as the Quality of Upper Extremities Skills Test); and (4) articles were written in English. In June 2013, we searched PubMed, CINAHL, Cochrane, and PsycINFO (via EBSCO portal), using the following key words or mesh terms (as applicable): arm, upper extremity, upper arm, reach, grasp or grip, VR, computer game, virtual environment, Wii, Kinect, PlayStation, EyeToy, video game, CP, and hemip* (see Appendix 1 for our search strategy for PubMed). A total of 72 published articles were first found. After carefully reviewing the titles and abstracts, we determined that 14 articles met the criteria, and these were selected for further data extraction. We then used “virtual reality” and “children” to expand the search, and 231 published articles were found. After reading abstracts and titles, we selected 23 articles for further screening. Nine of the 23 articles retrieved from the second key word search were included in the 14 articles from the first key word search. Thus, a total of 28 articles were retrieved; 8 were review articles and 20 were research articles.
After reading the 8 review articles, 4 additional research articles were retrieved and included; with review articles excluded, the total number of research articles included for review was 24 (Figure 1).12–32 Among the 24 articles, 6 articles were excluded because of the following: (1) participants were not children with CP or had mixed diagnoses26,29; (2) no experimental data were included14,20; (3) intervention was not focused on UE function12; and (4) outcome measures were not related to UE function.22 Therefore, a total of 18 articles were included for final data extraction.13,15–19,21,23–25,27,28,30–32
A meta-analysis coding template was created and used to code the demographic, methodological, and miscellaneous variables extracted from each report included in the review, following the method suggested by Cooper and Hedges.33 Demographic variables included age, ethnicity, gender, diagnosis, severity, cognitive status, and other disabilities of the participants in each study. Sample size, sampling method, research design, type of UE movements investigated, VR type, VR dosing (duration, intensity, length, and total treatment duration), and instruments used to measure outcome variables were coded as methodological characteristics. Year and type of publication (journal, book chapter, dissertation, and conference proceeding), name of the authors, country, and affiliation of the authors were included as miscellaneous variables.
The quality of RCTs was evaluated using the scoring protocol developed by Kwakkel et al34 and Cambach et al,35 which was used by Huang et al36 to evaluate the validity on RCT research included in their systematic review investigating the effect of constraint-induced movement therapy on children with hemiplegic CP.34–36 This scoring system consisted of 16 items with categories on randomization, matching, blinding procedure, dropouts and intention-to-treat analysis, characteristics of measurement instruments, control of cointerventions, comparability of group characteristics, and control for dose of therapy.
Because this scoring system was mainly designed for RCTs, we then adopted a similar adapted scoring protocol, which was used by Huang et al36 to evaluate case series or single-subject design articles. This scoring system consisted of 11 items with categories on study population, design, blinding procedure, measurement instruments, control of cointervention, control for dose of therapy, and appropriate statistical analysis.
The level of evidence for each study also was rated on the basis of the categorization proposed by the Oxford Centre for Evidence Based Medicine (http://www.cebm.net/index.aspx?o=1025).
Relevant information from the included studies was extracted and coded by the same author (Y.C.). A second reviewer checked the extracted data. Any discrepancies were resolved by discussion to reach consensus.
Because most of the reviewed articles were either case series or single-subject studies, we used the preintervention score as the control and compared this with the immediate postintervention score (see Tables 2–4). For the 3 RCTs included in our review, the effect size between the VR group and the conventional therapy group was also computed. The effect size was calculated with the Comprehensive Meta-analysis software (Version 2.2; Biostat Inc, Englewood, New Jersey) to compute Cohen's d between the pre- and postinterventions and between the VR group and the conventional therapy group. In this study, effect sizes were computed from the means, standard deviations, and sample sizes for all outcome measures. In some case series or single-subject design articles, individual data were offered, and then means and standard deviations were calculated to obtain the effect size. If the research contained more than 1 effect size, an average effect size was used to represent the research. After computing the effect size for each report, all the effect sizes were combined to form a common estimate of effect size. Heterogeneity tests and z scores were also calculated with the Comprehensive Meta-analysis software. These effect sizes were interpreted with Cohen's convention of small (0.2), medium (0.5), and large (0.8).37 Meta-analyses were run using a random effect model that accounts for true variation in effects that vary from study to study and also accounts for random error within each study. Examples of true variation that could potentially affect study effect sizes were children's characteristics (eg, age), intervention protocols (eg, intervention length), and outcome measures.
We then sought to determine the role of experimental factors in explaining the considerable interstudy variation observed in effect sizes. These experimental factors can be treated as moderator variables in a meta-analysis. Meta-regressions (using a method-of-moments model) or subgroup meta-analyses were used to examine the following potential moderator variables: (1) children's age, (2) CP type, (3) intervention weekly dosage, (4) intervention length, (5) total intervention duration, (6) intervention setting, (7) VR system, (8) research design, and (9) use of kinematics as the outcome measure. All these analyses were done using the Comprehensive Meta-analysis software.
The data extracted from the 18 eligible studies are summarized in Tables 2 to 3. Among the 18 studies, 4 reported only 1 participant17,30–32; therefore, these studies were excluded from the effect size calculation but remained in the summary tables. Among these 4 excluded articles, Burdea et al31 and Huber et al32 reported the same participants as those reported in Golomb et al.16 Thus, the meta-analytic data in this review included 14 studies (Figure 1).13,15,16,18,19,21,23–25,27,28,38–40
Level of Evidence and Quality of Study
Among the 18 articles, only 3 RCTs were rated as 2b level of evidence, with the rest of the studies ranked as level 4 evidence. For the 3 RCTs, the quality score for Jannink et al,19 Reid and Campbell,25 and Rostami et al39 was 5/16, 9/16, and 11/16, respectively (Table 4). Among the level 4 evidence, 2 studies using single-subject design had better validity (8/11).13,18 In general, the methodological quality of articles that investigated VR effect was poor to fair.
Types of Outcome Measures
The following measures of UE motor function were used in the studies, including the Canadian Occupational Performance Measure, the Quality of Upper Extremity Test (QUEST), the Jebsen-Taylor Hand Function Test (J-L), the Melbourne Assessment of Unilateral Upper Limb Function (Melbourne), the Bruininks-Oseretsky Test of Motor Proficiency (BOTMP), the ABILHAND-Kids Questionnaire, the Children's Hand Use Experience Questionnaire, Pediatric Motor Activity Log, the Shriner's Hospital Upper Extremity Evaluation, reaching kinematics, range of motion, and grip and pinch strength. We used the ICF model to group the outcome variables (see Table 5). For example, the Canadian Occupational Performance Measure was categorized in the ICF's participation component; the QUEST, Melbourne, and BOTMP were in the activity component; and reaching kinematics was considered to represent body structure and function. Among the research used in this systematic review/meta-analysis, the majority of the articles included outcome variables in the activity and body structure components,13–18,21,23,30–32,37 1 article contained participation component measures only,24 1 included participation and activity components,25 3 included activity level only,19,38–39 and 1 included participation and body structure components27 (see Table 5).
The sample sizes included in this review ranged from 1 to 31 (considered very small sample sizes). Types of CP ranged from pure hemiplegia to a mix of diplegia, hemiplegia, and quadriplegia. The age of participants ranged from 3 to 18 years, with the majority of research investigating VR on children aged 8 years or older.
Type of Intervention
The VR systems used in these research efforts ranged from commercially available systems (eg, PlayStation) to expensive engineer-built systems (eg, IREX and NJIT-RAVR). The location of training also varied with more conducted in clinics or laboratory environments. The dosage of VR also varied from study to study with the majority of studies reporting a weekly dosage less than 120 minutes. The length of intervention also varied from 3 weeks to 14 months with the majority of studies having an intervention length around 3 to 4 weeks. The intervention frequency also varied from 1 day per week to as much as the children wanted.
Effectiveness of VR on UE Function
Comparisons Between VR and Conventional Therapy. Only 3 RCT reports provided data to compare VR with the conventional therapy sessions. The effectiveness of VR on UE function in the 3 RCTs showed a statistical significance of P = .08, with 2 studies showing a large effect. The average effect size was d = 1.97, 95% confidence interval (CI) = [−0.26, 4.20].
Comparisons Between Pre- and Post-VR Intervention (Immediately After VR)
The effectiveness of VR in the 3 RCTs was 0.30 (small), 1.31 (large), and 5.09 (large) with controversial results when the comparison was made between pre-VR and post-VR interventions. However, the efficacy of VR in the case series or single-subject design articles was more consistent: at least 1 positive change in 1 outcome measure was reported in all case series (see Table 5).
The results from the meta-analyses are summarized in Figure 2. When combining all the studies, the effect of VR is promising—children who received VR showed a strong effect (d = 1.00, 95% CI [0.45, 1.56]) when compared with their baseline scores (Figure 2). We also found moderate heterogeneity among studies because the value of I2 was 56%.
When further breaking down the effect size on the basis of outcome variables classified according to the ICF model, a large effect was found when measured through participation (d = 1.92, 95% CI = [1.19, 2.66]), a small effect on activity (d = 0.46, 95% CI = [−0.08, 1.16]), and a medium effect on body structure and function (d = 0.70, 95% CI = [0.10, 1.30]).
Subgroups and Meta-regression Analyses
Table 6 summarizes the findings of the subgroup meta-analyses probing possible roles that 8 experimental factors might have in explaining the effect size dispersion among the 14 studies. Intervention setting, VR system, and research design were significant factors (P < .05). Specifically, we found that VR in a home or laboratory setting seemed to be more effective than in a clinic setting (d = 1.30, I2 = 40%, for home; d = 1.47, I2 = 58%, for laboratory; and d = 0.28, I2 = 0%, for clinic). The VR system used in the studies varied; therefore, we classified the systems into engineer-built systems versus commercially available systems. Studies that used engineer-built systems had a strong VR effect (d = 1.27, I2 = 55%), whereas studies that used commercially available systems showed a small effect (d = 0.31, I2 = 0%). We also found that the VR effect in the RCT design or case series seemed to be more effective than in a cohort design (d = 1.95, I2 = 85%, for RCT; d = 1.16, I2 = 2%, for case series; and d = 0.21, I2 = 0%, for cohort design). A meta-regression analysis also showed a negative linear relationship between age and VR effect size (P = .07) (see Figure 3). The older the children were, the smaller the effect size.
Virtual reality intervention to improve UE function in children with CP still remains a relatively new method, and the evidence on its effectiveness is gradually emerging. Our meta-analysis combined 3 RCTs and other case series validated that VR had a potentially strong effect in improving the UE function of children with CP when comparing pre-VR data with post-VR data. Each study included in this review showed at least 1 positive outcome among the outcome variables except for the RCT done by Reid and Campbell.25 When comparing VR with conventional therapy, VR also showed a strong effect on the basis of the 3 RCTs. The possible mechanism for why VR worked remains uncertain. Recently, Levac et al,41 in a scoping review, suggested 9 potential active ingredients from VR therapy, which help children with CP improve their motor skills: (1) opportunities for practice because repetitive task practice advances functional abilities, (2) task specificity between the VR and real-world movements, (3) flexibility to individualize treatment parameters, (4) visual and/or auditory feedback, (5) social play equalization for participating in play situation, (6) neuroplastic changes, (7) problem solving through different virtual contexts, (8) motivation because children can select the games they like or compete with peers, and (9) support from the parents or therapist using verbal encouragement and feedback. All these components are essential when learning and improving a motor skill from a motor learning perspective. Therefore, all these components may be part of the underlying mechanism explaining why VR works.
In our review, when classifying the outcome variables on the basis of the ICF model, studies using measurements of the participation component and body structure and function component showed medium-to-large effect sizes, whereas studies using measurements of the activity component showed small effects. The outcome measurements of the activity component varied with different purposes and sensitivities.42 For example, the QUEST and the Melbourne Assessment are evaluation tools that measure quality of UE movement by testing various unimanual items—the QUEST is more focused on dissociated movements, and the Melbourne assesses items involving reach, grasp, release, and manipulation. These latter 2 tools were scored using dichotomous or ordinal scales. The BOTMP and the Peabody Developmental Motor Scale-Fine Motor Domain (PDMS-2) were used to measure fine motor skills using an ordinal scale. The raw scores from these 2 tools can be converted into a standard score with an age equivalent. The Jebsen-Taylor test measures movement speed in 7 unimanual tasks and can be compared with reference values. Consequently, different assessment tools, which were designed for different purposes and mainly used an ordinal scale to evaluate the child's performance in the ICF's activity component, may not be sensitive enough to capture change in the child's performance and may result in a small effect size when examining the activity component.
Our findings also showed that training children in laboratories and their homes had a better effect than training in clinics. It is not surprising to have a better effect in a well-controlled laboratory environment. For the better effect found in a home-based training, several advantages can be gained through this natural environment. We anticipate less distress during VR practice for both children with CP and their parents because they are in their own familiar environment. Also, the training can be tailored to fit into the family's daily routine, which potentially allows children to receive therapy for a longer period of time. A home-based intervention can also save the family time (in terms of travel time back and forth to the clinic) and money. Parents would be involved more throughout the process, which will increase the opportunities for parent-child interaction. This home-based intervention may also increase opportunities for the children to interact socially with their siblings and other family members.
It is also worth noting that the use of the engineer-built system was more effective than using the commercial system. Our explanation is that the engineer-built system can meet the children's needs through better adjustment of game difficulty and training goals. Commercial systems, on the contrary, are restricted to predetermined task difficulties, which are typically too difficult for children with CP. However, the cost to build an engineer-built VR system is generally much higher (eg, IREX system $16 249, www.flaghouse.ca) than that of a commercially available system (eg, Xbox Kinect system $250-$400, www.amazon.com). How to decrease the cost of an engineer-built system to be affordable to a family with a child with CP is something to consider in the future.
In our review, VR effect size had a negative linear association with children's age—the older the children, the smaller the effect. This was consistent with other findings reflecting the importance of early intervention because younger children may have more potential for brain plasticity and adaptability to move away from stereotyped reaching patterns than do children at older ages.43,44
Interestingly, no association between VR effect size and treatment dosing (treatment length, duration, and weekly dose) was found. Possible explanations for this finding include the following: (1) the number of participants in each study was low, and (2) high heterogeneity was evident among studies included in this review, including age range, severity, and type of diagnosis.
It is worth noting the limitations of using meta-regression to examine the association between effect size and subject-level measures (eg, children's age).45–47 The subject-level measures reported in the published literature were typically the averages of patient characteristics. The relationship with patient averages across different studies may not reflect the true relationship for all individual patients within all studies. For example, Petkova et al45 and Berlin et al47 compared the relationship of medicine effect size with grouped data and again with individual data and found that the group-level analysis failed to detect the true relationship found in the individual-level analysis. Thus, although we did find that children's age was negatively related to VR effect size in our review and this finding was consistent with other literature, the interpretation of this finding required caution because we used group-level data (ie, averaged age from each study) rather than individual age data from all children participating in these studies.
The findings of our review were quite consistent with others that show that VR seems to be a viable tool to improve UE function in children with CP. However, the majority of the published research using a VR intervention used case series or single-subject design. The quality of the research was poor with high heterogeneity among studies and various intervention protocols. A larger-scale RCT with a more homogenous participant group of similar age and diagnosis is needed. Moreover, this large-scale RCT should use a home-based VR intervention with younger children with CP using an engineer-built system to maximize the effect on UE function in children with CP.
CLINICAL INTERPRETATION OF EFFECT SIZE
Effect sizes have been proposed to be viewed as benchmarks for understanding clinically meaningful change in health status—a medium effect size (d = 0.5) corresponds to an amount of change that is noticeable to a careful observer.48–50 That is, a medium effect size can represent improvement after receiving the treatment that could be perceived by the patients as beneficial and important. For example, in our review, Rostami et al39 used Pediatric Motor Activity Log to evaluate children's improvement before and after VR intervention and reported a change score for the amount of hand use at 1.71 points. This change score far exceeded the minimal clinically important difference, which was 0.39 as reported by Lin et al,51 showing a beneficial and important improvement. Similar conclusions could be made by simply calculating the effect size (d = 3.92), which showed a large effect for VR intervention. In summary, a medium effect size could be used as a parameter to determine clinically meaningful change.
Virtual reality is a viable tool to improve UE function in children with CP. Implementing a home-based or laboratory-based VR intervention program using an engineer-built VR system for young children with CP seems to yield the most improvement in their UE function. However, more vigorous research designs are needed to make a conclusive recommendation.
1. Jones MW, Morgan E, Shelton JE, Thorogood C. Cerebral palsy
: introduction and diagnosis (part I). J Pediatr Health Care. 2007;21(3):146–152.
2. Fedrizzi E, Pagliano E, Andreucci E, Oleari G. Hand function in children with hemiplegic cerebral palsy
: prospective follow-up and functional outcome in adolescence. Dev Med Child
3. Krigger KW. Cerebral palsy
: an overview. Am Fam Phys. 2006;73(1):91–100.
4. Centers for Disease Control and Prevention. Cerebral palsy
. Accessed August 25, 2012.
5. Parsons TD, Rizzo AA, Rogers S, York P. Virtual reality
in paediatric rehabilitation: a review. Dev Neurorehabil. 2009;12(4):224–238.
6. Snider L, Majnemer A, Darsaklis V. Virtual reality
as a therapeutic modality for children with cerebral palsy
. Dev Neurorehabil. 2010;13(2):120–128.
7. Weiss PL, Rand D, Katz N, Kizony R. Video capture virtual reality
as a flexible and effective rehabilitation tool. J NeuroEng Rehabil. 2004;1:12–24.
8. Wilson P, Foreman N, Stanton D. Virtual reality
, disability and rehabilitation. Disabil Rehabil. 1997;19:213–220.
9. Galvin J, Levac D. Facilitating clinical decision-making about the use of virtual reality
within paediatric motor rehabilitation: describing and classifying virtual reality
systems. Dev Neurorehabil. 2011;14(2):112–122.
10. Galvin J, McDonald R, Catroppa C, Anderson V. Does intervention using virtual reality
improve upper limb function in children with neurological impairment: a systematic review
of the evidence. Brain Inj. 2011;25(5):435–442.
11. Wang M, Reid D. Virtual reality
in pediatric neurorehabilitation: attention deficit hyperactivity disorder, autism and cerebral palsy
. Neuroepidemiology. 2011;36:2–18.
12. Bryanton C, Bosse J, Brien M, McLean J, McCormick A, Sveistrup H. Feasibility, motivation, and selective motor control: virtual reality
compared to conventional home exercise in children with cerebral palsy
. Cyberpsychol Behav. 2006;9(2):123–128.
13. Chen YP, Kang LJ, Chuang TY, et al. Use of virtual reality
to improve upper-extremity control in children with cerebral palsy
: a single-subject design. Phys Ther. 2007;87(11):1441–1457.
14. Dunne A, Do-Lenh S, O' Laighin G, Shen C, Bonato P. Upper extremity rehabilitation of children with cerebral palsy
using accelerometer feedback on a multitouch display. Conf Proc IEEE Eng Med Biol Soc. 2010;2010:1751–1754.
15. Fluet GG, Qiu Q, Kelly D, et al. Interfacing a haptic robotic system with complex virtual environments to treat impaired upper extremity motor function in children with cerebral palsy
. Dev Neurorehabil. 2010;13(5):335–345.
16. Golomb MR, McDonald BC, Warden SJ, et al. In-home virtual reality
videogame telerehabilitation in adolescents with hemiplegic cerebral palsy
. Arch of Phys Med Rehabil. 2010;91(1):1–8.
17. Golomb MR, Warden SJ, Fess E, et al. Maintained hand function and forearm bone health 14 months after an in-home virtual-reality videogame hand telerehabilitation intervention in an adolescent with hemiplegic cerebral palsy
. J Child
18. Green D, Wilson PH. Use of virtual reality
in rehabilitation of movement in children with hemiplegia—a multiple case study evaluation. Disabil Rehabil. 2012;34(7):593–604.
19. Jannink MJ, van der Wilden GJ, Navis DW, Visser G, Gussinklo J, Ijzerman M. A low-cost video game applied for training of upper extremity function in children with cerebral palsy
: a pilot study. Cyberpsychol Behav. 2008;11(1):27–32.
20. Li W, Lam-Damji S, Chau T, Fehlings D. The development of a home-based virtual reality
therapy system to promote upper extremity movement for children with hemiplegic cerebral palsy
. Technol Disabil. 2009;21:107–113.
21. Qiu Q, Ramirez DA, Saleh S, et al. The New Jersey Institute of Technology Robot-Assisted Virtual Rehabilitation (NJIT-RAVR) system for children with cerebral palsy
: a feasibility study. J Neuroeng Rehabil. 2009;6:40.
22. Reid D. The influence of virtual reality
on playfulness in children with cerebral palsy
: a pilot study. Occup Ther Int. 2004;11(3):131–144.
23. Reid DT. The use of virtual reality
to improve upper-extremity efficiency skills in children with cerebral palsy
: a pilot study. Technol Disabil. 2002;14:53–61.
24. Reid DT. Benefits of a virtual play rehabilitation environment for children with cerebral palsy
on perceptions of self-efficacy: a pilot study. Pediatr Rehabil. 2002;5(3):141–148.
25. Reid DT, Campbell K. The use of virtual reality
with children with cerebral palsy
: a pilot randomized trial. Ther Recreation J. 2006;40(4):255–268.
26. Sveistrup H, Thornton M, Bryanton C, et al. Outcomes of intervention programs using flatscreen virtual reality
. Conf Proc IEEE Eng Med Biol Soc. 2004;7:4856–4858.
27. Wann JP, Turnbull JD. Motor skill learning in cerebral palsy
: movement, action and computer-enhanced therapy. Baillieres Clin Neurol. 1993;2(1):15–28.
28. Weightman A, Preston N, Levesley M, et al. Home based computer-assisted upper limb exercise for young children with cerebral palsy
: a feasibility study investigating impact on motor control and functional outcome. J Rehabil Med. 2011;43(4):359–363.
29. Wille D, Eng K, Holper L, et al. Virtual reality
-based paediatric interactive therapy system (PITS) for improvement of arm and hand function in children with motor impairment—a pilot study. Dev Neurorehabil. 2009;12(1):44–52.
30. You SH, Jang SH, Kim YH, Kwon YH, Barrow I, Hallett M. Cortical reorganization induced by virtual reality
therapy in a child
with hemiparetic cerebral palsy
. Dev Med Child
31. Burdea GC, Jain A, Rabin B, Pellosie R, Golomb M. Long-term hand tele-rehabilitation on the PlayStation 3: benefits and challenges. Conf Proc IEEE Eng Med Biol Soc. 2011:1835–1838.
32. Huber M, Rabin B, Docan C, Burdea GC, AbdelBaky M, Golomb MR. Feasibility of modified remotely monitored in-home gaming technology for improving hand function in adolescents with cerebral palsy
. IEEE Trans Inf Technol Biomed. 2010;14(2):526–534.
33. Cooper H, Hedges LV, eds. The Handbook of Research Synthesis. New York City: Russell Sage Foundation; 1994.
34. Kwakkel G, Wagenaar RC, Koelman TW, Lankhorst GJ, Koetsier JC. Effects of intensity of rehabilitation after stroke: a research synthesis. Stroke. 1997;28:1550–1556.
35. Cambach W, Wagenaar RC, Koelman TW. The long-term effects of pulmonary rehabilitation in patients with asthma and chronic obstructive pulmonary disease: a research synthesis. Arch Phys Med Rehabil. 1999;80:103–111.
36. Huang HH, Fetters L, Hale J, McBride A. Bound for success: a systematic review
of constraint-induced movement therapy in children with cerebral palsy
supports improved arm and hand use. Phys Ther. 2009;89(11):1126–1141.
37. Cohen J. Statistical Power Analysis for the Behavioral Sciences. Hillsdale, NJ: Erlbaum; 1988.
38. Rios DC, Gilbertson T, McCoy SW, et al. NeuroGame Therapy to improve wrist control in children with cerebral palsy
: a case series. Dev Neurorehabil. 2013;16(6):398–409.
39. Rostami HR, Arastoo AA, Nejad SJ, Mahany MK, Malamiri RA, Goharpey S. Effects of modified constraint-induced movement therapy in virtual environment on upper-limb function in children with spastic hemiparetic cerebral palsy
: a randomised controlled trial. NeuroRehabil. 2012;31(4):357–365.
40. Winkels DG, Kottink AI, Temmink RA, Nijlant JM, Buurke JH. Wii™-habilitation of upper extremity function in children with cerebral palsy
. An explorative study. Dev Neurorehabil. 2013;16(1):44–51.
41. Levac D, Rivard L, Missiuna C. Defining the active ingredients of interactive computer play interventions for children with neuromotor impairments: a scoping review. Res Dev Disabil. 2012;33:214–223.
42. Klingels K, Jaspers E, Van de Winckel A, De Cock P, Molenaers G, Feys H. A systematic review
of arm activity measures for children with hemiplegic cerebral palsy
. Clin Rehabil. 2010;24:887–900.
43. Morgan C, Novak I, Badawi N. Enriched environments and motor outcomes in cerebral palsy
: systematic review
. Pediatrics. 2013;132(3):e735–e746.
44. Spittle A, Orton J, Anderson P, Boyd R, Doyle LW. Early developmental intervention programmes post-hospital discharge to prevent motor and cognitive impairments in preterm infants. Cochrane Database Syst Rev. 2012;12:CD005495.
45. Petkova E, Tarpey T, Huang L, Deng L. Interpreting meta-regression: application to recent controversies in antidepressants’ efficacy. Stat Med. 2013;32(17):2875–2892.
46. Schmid CH, Stark PC, Berlin JA, Landais P, Lau J. Meta-regression detected associations between heterogeneous treatment effects and study-level, but not patient-level, factors. J Clin Epidemiol. 2004;57(7):683–697.
47. Berlin JA, Santanna J, Schmid CH, Szczech LA, Feldman HI. Individual patient- versus group-level data meta-regressions for the investigation of treatment effect modifiers: ecological bias rears its ugly head. Stat Med. 2002;21(3):371–387.
48. Mann BJ, Gosens T, Lyman S. Quantifying clinically significant change: a brief review of methods and presentation of a hybrid approach. Am J Sports Med. 2012;40(10):2385–2392.
49. Eisen SV, Ranganathan G, Seal P, Spiro A. Measuring clinically meaningful change following mental health treatment. J Behav Health Serv Res. 2007;34(3):272–289.
50. Norman GR, Sloan JA, Wyrwich KW. Interpretation of change in health-related quality of life: the remarkable universality of half a standard deviation. Med Care. 2003;41(5):582–592.
51. Lin KC, Chen HF, Chen CL, et al. Validity, responsiveness, minimal detectable change, and minimal clinically important change of the Pediatric Motor Activity Log in children with cerebral palsy
. Res Dev Disabil. 2012;33:570–577.
APPENDIX Details of Search Strategy, With PubMed as an Example
- Arm or upper extremity or upper arm or reach or grasp or grip = 313 983 hit
- Virtual reality or computer game or virtual environment or Wii or Kinect or PlayStation or EyeToy or video game = 10 787 hit
- Cerebral palsy or hemip* = 46 984 hit
- 1 and 2 and 3 = 72 hit
- We then reviewed the title and abstract and selected 14 from the 72 articles.