http://cre.sagepub.com/Clinical Rehabilitation

http://cre.sagepub.com/content/26/9/798The online version of this article can be found at:

 DOI: 10.1177/0269215511434575

2012 26: 798 originally published online 24 January 2012Clin RehabilJH Crosbie, S Lennon, MC McGoldrick, MDJ McNeill and SM McDonough

pilot studyVirtual reality in the rehabilitation of the arm after hemiplegic stroke: a randomized controlled

  

Published by:

http://www.sagepublications.com

can be found at:Clinical RehabilitationAdditional services and information for    

  http://cre.sagepub.com/cgi/alertsEmail Alerts:

 

http://cre.sagepub.com/subscriptionsSubscriptions:  

http://www.sagepub.com/journalsReprints.navReprints:  

http://www.sagepub.com/journalsPermissions.navPermissions:  

What is This? 

- Jan 24, 2012OnlineFirst Version of Record  

- Aug 7, 2012Version of Record >>

at UZH Hauptbibliothek / Zentralbibliothek Zürich on June 30, 2014cre.sagepub.comDownloaded from at UZH Hauptbibliothek / Zentralbibliothek Zürich on June 30, 2014cre.sagepub.comDownloaded from

Clinical Rehabilitation26(9) 798 –806© The Author(s) 2012Reprints and permission: sagepub.co.uk/journalsPermissions.navDOI: 10.1177/0269215511434575cre.sagepub.com

CLINICALREHABILITATION

434575 CRE26910.1177/0269215511434575Crosbie et al.Clinical Rehabilitation2012

1Neurological Physiotherapist, Belfast, UK2Health and Rehabilitation Sciences Research Institute and School of Health Sciences, University of Ulster, Newtownabbey, UK3Northern Ireland Stroke Research Network, Northern Health and Social Care Trust, UK4School of Computing and Information Engineering, University of Ulster, Coleraine, UK

Virtual reality in the rehabilitation of the arm after hemiplegic stroke: a randomized controlled pilot study

JH Crosbie1,2, S Lennon2, MC McGoldrick3, MDJ McNeill4 and SM McDonough2

AbstractObjective: To assess the feasibility of a trial to investigate the effectiveness of virtual reality-mediated therapy compared to conventional physiotherapy in the motor rehabilitation of the arm following stroke, and to provide data for a power analysis to determine numbers for a future main trial.Design: Pilot randomized controlled trial.Setting: Clinical research facility.Participants: Eighteen people with a first stroke, 10 males and 8 females, 7 right and 2 left side most affected. Mean time since stroke 10.8 months.Interventions: Participants were randomized to a virtual reality group or a conventional arm therapy group for nine sessions over three weeks.Main measures: The upper limb Motricity Index and the Action Research Arm Test were completed at baseline, post intervention and six weeks follow-up.Results: Outcome data were obtained from 95% of participants at the end of treatment and at follow-up: one participant withdrew. Compliance was high; only two people reported side-effects from virtual reality exposure. Both groups demonstrated small (7–8 points on upper limb Motricity Index and 4 points on the Action Research Arm Test), but non-significant, changes to their arm impairment and activity levels.Conclusion: A randomized controlled trial of virtual reality-mediated therapy comparable to conventional therapy would be feasible, with some suggested improvements in recruitment and outcome measures. Seventy-eight participants (39 per group) would be required for a main trial.

KeywordsVirtual reality, stroke rehabilitation, arm

Received: 29 September 2010; accepted: 15 November 2011

Article

Corresponding author:Suzanne McDonough, Health and Rehabilitation Sciences Research Institute, University of Ulster, Newtownabbey, BT37 0QB, UKEmail: s.mcdonough@ulster.ac.uk

at UZH Hauptbibliothek / Zentralbibliothek Zürich on June 30, 2014cre.sagepub.comDownloaded from

Crosbie et al. 799

Introduction

Paresis of the arm post stroke remains an important clinical problem with fewer than 50% of people post stroke recovering some degree of movement or function.1 Underutilization of the affected limb can occur after stroke2 and people tend to compensate with the intact limb.3 Virtual reality is a computing technology that presents simulated information to allow individuals to experience and interact with or within three-dimensional environments.4 Virtual reality may hold some solutions to these problems. It has been shown to be an interactive and enjoyable medium that, with sufficient use, may improve upper limb motor function in adults with stroke.5 Virtual technologies can be used to produce envi-ronments in which intensity of practice and feed-back on performance can be manipulated to provide tailored motor training.6,7 However, reviews have indicated that although very promising there are still problems relating to the limited quality of methodologies used.8,9 The majority of evidence has been based on single-case or uncontrolled designs. Recent studies continue to report the vari-ety of set-up and design of the technology. Examples of such are a novel upper limb system consisting of three-dimensional tracker, custom forearm support, workstation and library of three-dimensional exer-cises10 and a haptic master–slave system.11

Two randomized controlled trials have compared virtual reality-mediated therapy for the arm to no intervention12 and to alternative digital applica-tions.13 In the trial reported by Jang et al.12 signifi-cant improvements were found in the Box and Block Test; the Fugl-Meyer Assessment and manual func-tion test. This study also reported a novel demon-stration of virtual reality-induced neuroplastic changes associated with motor recovery as indicated through the use of functional magnetic resonance imaging. Fischer et al.13 reported significant changes on the Box and Block Test and the Fugl-Meyer Assessment, along with a significant change in Wolf Motor Function Test score. However, the authors indicated that overall gains were small. One further clinical trial has compared conventional therapy to virtual reality-based therapy with little detail as to the content of the conventional therapy sessions.14

The aim of this study was to conduct a pilot ran-domized controlled trial to investigate the effective-ness of virtual reality-mediated therapy compared to conventional physiotherapy in the motor rehabilita-tion of the arm following stroke. Objectives were (1) to identify the pattern of recruitment and retention; (2) to pilot the methodological procedures; (3) to provide data to inform a power analysis to determine sample size for a future trial; and (4) to assess the effects of virtual reality-mediated therapy on the arm impairment and activity levels of people with stroke.

Method

Data were collected over an 18-month period. In this pilot randomized controlled trial participants were assigned into groups by an independent col-league, using a computer-generated randomization table. Group allocation cards were concealed in opaque envelopes to be opened by the treating research therapist, following assessment and screen-ing by the blinded independent outcome assessor. Ethical approval was obtained from the Queens University Belfast Research Ethics Committee and from the Office of Research Ethics Committees in Northern Ireland.

The records of those who had been admitted to two local hospital stroke units for rehabilitation in the preceding 6–12 months were examined for potential participants who met the selection criteria. Volunteers were also sought from local stroke sup-port groups. People were contacted by letter and telephone follow-up and then screened for recruit-ment into the trial. History taking and screening were completed on first contact by the independent outcome assessor to ensure the participant fulfilled the inclusion criteria.

All participants provided written informed con-sent. Inclusion criteria were that the potential par-ticipant was medically stable; an adult aged 18–85 years; 6–24 months following a first stroke and able to follow a two-step command. Potential partici-pants were excluded if they presented with a mental score test of less than 7/10,15 a star cancellation score of less than 48/52,16 scored less than 25 out of 100 on the upper limb Motricity Index,17 had

at UZH Hauptbibliothek / Zentralbibliothek Zürich on June 30, 2014cre.sagepub.comDownloaded from

800 Clinical Rehabilitation 26(9)

comorbid conditions affecting their rehabilitation potential (e.g. cardiac, respiratory or arthritic prob-lems) and reported arm pain on a visual analogue scale of >6/10. A score of 0–25 on the Motricity Index is deemed to indicate very severe motor loss.18

Following medical advice it was agreed to also exclude anyone with a cardiac pacemaker as the electromagnetic motion tracker used within the vir-tual reality system might interfere with such devices. The length and timing of the virtual reality interven-tion was derived from a detailed review of existing evidence published in this area.8 The system com-prised a desktop computer, a head-mounted display unit, a motion tracking system and sensors. The therapist could navigate through the data input and graphics by means of drop-down menus. The head-mounted display facilitated the participant’s immer-sion in the virtual environment. Three sensors were applied to the shoulder, elbow and hand of the par-ticipant, who could then manipulate objects and carry out upper limb tasks within a 3D environment. The virtual tasks were designed to simulate a range of upper limb tasks related to reach to target, reach and grasp and game tasks.19 As the participant’s per-formance progressed the tasks could be made easier or more difficult by means of changing the distance or height of objects or the speed of stimulus. The experimental group underwent a three-week train-ing period using virtual reality-mediated upper limb therapy. This was delivered over three sessions per week of 30–45 minutes duration, which consisted of participant set-up and physical practice focusing on specific upper limb tasks. Each person in the vir-tual reality group received nine treatment sessions (three per week for three weeks).

The control group received therapy, focusing on the upper limb, to control for any dose effect in terms of the amount of intervention between groups. Conventional therapy followed an evidence-based approach. This was delivered by a physiotherapist, experienced in stroke rehabilitation, and followed a programme of techniques, which included muscle facilitation, stretching exercises, strengthening activities and the inclusion of the more affected upper limb in functional tasks.20 The content of therapy was recorded for each session, using a

treatment checklist. A sample of these sessions was video-recorded to allow the content to be checked by an independent physiotherapist, also experi-enced in stroke rehabilitation.

Given the nature of the treatments it was not pos-sible for the treating therapist and the participants to be blinded. Therefore only the outcome assessor was blinded to group allocation. Non-compliance was defined as receipt of less than three out of a total of nine treatments, but such participants were included in subsequent follow-ups for the purposes of inten-tion-to treat analysis. Participants were requested to continue with their normal activities. If they included any additional upper limb exercise or activities dur-ing the duration of the study they were asked to record this in a treatment diary. Both the virtual reality-mediated therapy and the conventional ther-apy were administered by the same physiotherapist.

The upper limb Motricity Index17 and the Action Research Arm Test21 were used to assess motor impairment and function. These have been vali-dated for use in the stroke population.22,23 These were completed at baseline, post intervention and six weeks follow-up. An exit questionnaire was completed by each participant at the end of inter-vention. This evaluated the conduct of the trial and the participants’ perceptions of the interventions and their satisfaction with the treatment received

Data analyses

All variables were analysed using the Statistical Package for the Social Sciences (SPSS) version 11.5 (SPSS Inc., Chicago, IL, USA). Intention-to-treat analysis was carried out by an investigator blinded to treatment allocation. Missing data were dealt with by the simple mean imputation method, whereby the missing value is replaced with the mean of the group.24 As the data were not normally distributed non-parametric tests were conducted. The Mann–Whitney and Friedman’s test are the non-parametric alternatives to the t-test for independent samples and the one-way repeated measures analysis of variance, respectively. The impairment and activity scores were compared to the minimal clinically important differ-ence (MCID) for the outcome measures used (i.e. the

at UZH Hauptbibliothek / Zentralbibliothek Zürich on June 30, 2014cre.sagepub.comDownloaded from

Crosbie et al. 801

minimum level of change of an outcome measure that is considered to be clinically relevant).25

ResultsSeventy-seven potential participants were contacted (see Figure 1 for flow diagram of study according to CONSORT statement26). Three people were deceased at time of contact. Fifty-six people were excluded for the following main reasons: unwilling-ness to participate (n = 23), non-response (n = 4) and ineligibility according to the inclusion criteria (n = 29). Of the 23 unwilling to participate 15 gave no specific reason; the remaining 8 identified

concurrently attending formal rehabilitation (3); personal time constraints (2) and problems with travelling to the research setting (3), as reasons for not participating. Thirty-eight per cent of potential participants did not meet the inclusion criteria: comorbidity (n = 15); more than 85 years (n = 2); Motricity Index score of less than 25 (n = 3); longer than 2 years post stroke (n = 6) and having a history of more than one stroke (n = 3). One participant dropped out after baseline measures and one treat-ment session, despite efforts to reschedule appoint-ments and to acquire further outcome data.

The demographic details of the participants are shown in Table 1. In summary, 18 participants were

Potential participantsn=77

Deceased n=3

Assess for eligibility n=74Excluded n=56(Ineligible n=29Unwilling n=23

Non-response n=4)Randomized n=18

Allocated to virtual reality(VR) group n=9Received VR n=9

Did not receive VR n=0

Followed up postintervention n=8At 6 weeks n=8

Loss to follow-up n=0Withdrawn n=1

Allocated to conventionaltherapy (CT) group n=9

Received CT n=9Did not receive CT n=0

Followed up postintervention n=9At 6 weeks n=9

Loss to follow-up n=0Withdrawn n=0

Analysed at follow-up n=9 Analysed at follow-up n=9

Figure 1. The flowchart of the trial according to the CONSORT statement.26

at UZH Hauptbibliothek / Zentralbibliothek Zürich on June 30, 2014cre.sagepub.comDownloaded from

802 Clinical Rehabilitation 26(9)

randomized into two groups (9 in each); 10 were male and 8 female, 7 with right and 11 with left side more affected by first stroke. The mean age of all participants was 60.3 years (range 38–77 years). The mean time since stroke was 10.8 months (range 2–24 months) and a mean time from being involved in formal rehabilita-tion was eight months (range 1–16 months).

The baseline range for all participants on the upper limb Motricity Index was 29–100 points (vir-tual reality group 71–100, conventional group

29–100), and on the Action Research Arm Test was 0–57 points (virtual reality group 32–57, conven-tional group 0–57). There were no significant or clinically important differences between the groups at baseline (see Table 2). Outcome data were obtained from 95% of participants at the end of treatment and at follow-up. One participant with-drew consent after one virtual reality session. Compliance in both groups was high, with all 17 participants completing all nine therapy sessions. Only two people reported side-effects of transient dizziness and headache from virtual reality expo-sure. Group median scores are presented in Table 2. The Mann–Whitney U-test indicated no significant differences between the groups for either the Motricity Index (P = 0.485) or the Action Research Arm Test (P = 0.139).

Group median scores indicated an improvement in upper limb Motricity Index scores for both of the groups, the virtual reality group by 7/100 points and the conventional groups by 8/100 points, at dis-charge. The improvement in Motricity Index scores was maintained in the virtual reality group at fol-low-up, whereas in the conventional group it decreased to the baseline level. Research has indi-cated that a change of 10% in an outcome score is clinically relevant.27 This would be a change of 10 points on the Motricity Index. Thus the trend in these results would not be clinically significant.

For the Action Research Arm Test there was a positive trend in favour of the conventional group at discharge from intervention, with a change of 3/57 points. At follow-up the virtual reality group had improved their performance on the Action Research

Table 1. Baseline characteristics

Virtual reality group (9 participants)

Conventional group (9 participants)

Age (years) Mean 56.1 64.6 SD 14.5 7.4Sex Male 5 5 Female 4 4Time since stroke (months) Mean 10 11.7 SD 6.4 7.8Side most affected Right hemiplegia 4 3 Left hemiplegia 5 6Time from formal rehabilitation to trial entry (months) Mean 7.3 9.2 SD 3.6 7.3

Table 2. Outcome measure scores at each time point

Conventional group Virtual reality group

Baseline Post intervention Follow-up Baseline Post intervention Follow-up

Upper Limb Motricity IndexMedian and (IQR) 84 (7.5) 92 (16) 84 (8.5) 77 (13) 84 (15.5) 84 (25)Mean ± (SD) Range 77.4 (19.5) 85 (22.1) 81.6 (20.9) 81.7 (9.4) 84.9 (9.2) 85.2 (12)Action Research Arm TestMedian and (IQR) 53 (8.5) 57 (2) 57 (0) 54 (8.5) 54 (4) 57 (0)Mean ± (SD) Range 47.3 (18.1) 50.2 (18.9) 50.7 (19) 51.3 (8.2) 52.8 (6.9) 52.1 (7.9)

IQR, interquartile range; SD, standard deviation.

at UZH Hauptbibliothek / Zentralbibliothek Zürich on June 30, 2014cre.sagepub.comDownloaded from

Crosbie et al. 803

Arm Test by 3/57 points. The minimum clinically important difference for the Action Research Arm Test is 6 points,23 and likewise the results in the cur-rent study were not clinically significant.

Exit questionnaires from 4 of the 9 virtual reality group participants reported that movement of the more affected arm had improved and they could undertake some tasks that they had not been able to do previously (e.g. driving a car; opening kitchen cupboards). The virtual reality-based therapy appeared to have been acceptable to participants. Seven of the 9 in the standard therapy group said that the movement in their more affected arm had improved and 8 reported they could undertake some tasks that they had not been able to do previously.

Participants indicated their satisfaction with the conduct of the trial, with only one person in the vir-tual reality group stating they did not enjoy the experience. The majority of participants in both groups would have liked the therapy they received to have been part of their original rehabilitation pro-gramme. The virtual reality-based therapy appeared to have been acceptable to participants.

The current data were used to calculate numbers required to detect significant differences for impair-ment and activity scores between groups. For impairment using an estimated within-group SD of 25 points, an minimal clinically important differ-ence between groups of 10 Motricity Index points, alpha of 0.05 and power 85%, a total of 21 partici-pants would need to be recruited in a future trial. For activity using the same power and alpha level, an estimated within-group SD of 18 points, and a minimal clinically important difference between groups of 6 Action Research Arm Test points, a total of 29 participants would be needed. Therefore allowing for a 30% attrition rate it would be neces-sary to recruit a total of 78 participants (39 per group).

Discussion

This study reports that conducting a randomized controlled trial of virtual reality-based therapy for the upper limb compared to conventional therapy is feasible; and is the only known randomized

controlled trial to have been conducted in Ireland and the UK comparing these two treatments. The trial design and methodology were acceptable to participants, with minimal drop-out at the six-week follow-up (control 0%; experimental 10%).

Recruitment patterns were disappointing, although not unexpected given the nature of the sample. The pattern of recruitment to our study reflected patterns in other trials involving people with stroke. For example, our referral rate was 31%, similar to that of Stewart et al.27 (38%) and slightly lower than that for Hill et al.28 (48%). Another main reason for non-participation was the presence of comorbidity as an exclusion criteria (with 20% excluded on this basis). On reflection, it may have been possible to relax these criteria to include peo-ple with respiratory or cardiac problems. Revision of selection criteria for a future study would not exclude participants on the basis of comorbidity, apart from those with a cardiac pacemaker.

Another factor limiting recruitment was refusal to participate. Schulz et al.29 recognized that recruit-ing participants with chronic disabling conditions into research studies is a complex process. Unfortunately, 30% of people contacted did not specify a reason for their unwillingness to partici-pate. It might be that this diagnostic group of par-ticipants would require longer than 24 hours to consider participation in this intervention involving computer technology, which may be unfamiliar to a predominately older age group. It might have been worth using a short video of a participant using the equipment as a reassurance for those who may have felt technologically naive.

Our study adds to the limited evidence base for virtual reality-based therapy for the upper limb, as indicated by two systematic reviews in 2007.8,9 This study has shown that it is feasible to conduct a main trial comparing virtual reality therapy to a well-designed conventional therapy programme for the upper limb following stroke. Although a previous trial has been reported which compared virtual real-ity to conventional care,14 it was unclear whether the conventional care followed best evidence and carefully controlled the dose between the two groups, as in this study. Two other studies11,12 have compared two forms of virtual reality and are

at UZH Hauptbibliothek / Zentralbibliothek Zürich on June 30, 2014cre.sagepub.comDownloaded from

804 Clinical Rehabilitation 26(9)

therefore less relevant to the question of how virtual reality compares to current best evidence-based practice in stroke rehabilitation. It would be impor-tant for future studies to consider what the most appropriate form of control intervention might be, either an active control, such as conventional ther-apy, or a no intervention group.

In terms of our feasibility testing we identified an important aspect that would need to be addressed in a main trial. With the outcome measures we chose it was not feasible to show clinically important dif-ferences between each group, which suggests that the measures we used were not sensitive enough. It was disappointing not to have recruited participants with a wider range of disabilities, with most scoring towards the ceiling of both the Motricity Index and Action Research Arm Test at baseline. Although it was not the intent to exclude the more disabled par-ticipants (upper limb Motricity Index <25/100) unfortunately the sample recruited tended to be the more mildly affected participant, with the majority scoring highly on both measures from baseline.

These particular measures were not sensitive enough to detect small to moderate changes with these higher functioning participants. At inception of this trial the literature strongly recommended the use of the Action Research Arm Test as an outcome measure for this population. It would have been useful to incorporate a timed measure (e.g. the Wolf Motor Function Test,30 the Jebsen Hand Function Test31 or the Nine Hole Peg Test32). The timed ele-ments in these tests may be more likely to detect the level of change amongst the type of participants recruited to this trial.

Further limitations include the use of opaque envelopes following group allocation by a computer generated randomization table. This was conducted by an independent colleague outside of the research team. However, the use of opaque envelopes has been suggested to still be open to bias. In this study group allocation was verified post trial and was found to match the original randomized list. The outcome assessor was blind to group allocation and the research therapist delivering the intervention was blind to outcome scores. The outcome assessor conducted all participant screening and baseline scores and is another limitation of this study. In a

main trial it would be advantageous to have separate individuals conduct these elements.

There may also have been some element of bias as the same therapist delivered both inter-ventions, perhaps then favouring one intervention over the other. However, this potential weakness was counteracted by the video recording of a sample of six participants from the conventional group. This allowed an independent and experi-enced physiotherapist to monitor the content of these sessions and to verify that practice recorded did not vary from that which would be deemed equivalent to usual practice in stroke rehabilita-tion units in the UK.

A larger trial would require at least 39 partici-pants in each arm in order to have the power to demonstrate statistically significant changes in both impairment and activity of the more affected upper limb. Using the same method of sample size calculation a similar number would be required when employing the Wolf Motor Function Test. In such a trial we would anticipate recruiting from multiple stroke units across Belfast. Although drop-out rate in this particular study was low, it would be more prudent to esti-mate for a drop-out of 30%.

Based on these preliminary results, with a small number of participants, it is not possible to defini-tively advocate virtual reality-mediated therapy over conventional therapy for the rehabilitation of the upper limb. The literature to date suggests that virtual reality-based therapy may be more appropri-ate for people following the acute phase post stroke. The length of virtual reality intervention period may have been insufficient to detect clinically signifi-cant changes. The three-week dose was determined from a review of the existing literature,8 at the time of the study inception, which was inconclusive. The team in this case made a decision to use the mean length of exposure from the published studies reviewed. In addition, the study was based in a research clinic and there may be value in proposing to conduct a larger trial in a stroke unit or clinical outpatient setting.

In conclusion, this pilot study has demonstrated the feasibility of a randomized controlled trial in this area, if careful consideration is given to the

at UZH Hauptbibliothek / Zentralbibliothek Zürich on June 30, 2014cre.sagepub.comDownloaded from

Crosbie et al. 805

recruitment methods and outcome measures. Larger trials are needed to provide high-quality evidence for the specific effect of virtual reality-mediated therapy in upper limb stroke rehabilitation. It may be that there are advantages over conventional ther-apy for people with stroke-related disabilities, in terms of improving the level of motor impairment or activity. Its value may lie in providing an alterna-tive method of therapy delivery and potentially a useful source of additional practice, particularly after the subacute phase, when people tend to have limited access to further rehabilitation.

Clinical messages

• Virtual reality seems to be a safe interven-tion for people in the chronic phases of stroke.

• It was an acceptable intervention to people with stroke.

• A larger trial with at least 78 participants (39 per group) would be needed to reach a conclusion on effectiveness.

Funding

This research was funded by project funding from the Northern Ireland Chest, Heart and Stroke Association (Application number: 2002004) and the Strategic Priority Grant, Department for Employment and Learning, Northern Ireland.

Acknowledgements

We would like to acknowledge the following individuals: Dr MI Wiggam for assistance in the acquisition of funding; Dr L Pokluda and Dr M Ma for supporting the technologi-cal aspects of the project; Ms E Gardner for statistical advice; Mrs G McMillan for assistance with the recruit-ment of participants.

References 1. Sunderland A, Tinson D and Bradley L. Arm function after

stroke: an evaluation of grip strength as a measure of recov-ery and a prognostic indicator. J Neurol Neurosurg Psychi-atry 1989; 52: 1267–1272.

2. Taub E, Miller N and Novack T. Technique to improve chronic motor deficit after stroke. Arch Phys Med Rehabil 1983; 74: 347–354.

3. DeLuca SC, Echols K, Ramey SL and Taub E. Pediatric constraint-induced movement therapy for a young child with cerebral palsy: two episodes of care. Phys Ther 2003; 83: 1003–1013.

4. Sisto S, Forrest GF and Glendinning D. Virtual reality applications for motor rehabilitation after stroke. Top Stroke Rehabil 2002; 8: 11–23.

5. Holden MK, Dyar TA and Dayan-Cimadoro L. Telerehabil-itation using a virtual environment improves upper extrem-ity function in patients with stroke. IEEE Trans Neurol Syst Rehabil Eng 2007; 15: 36–42.

6. Merians AS, Poizner H, Boian R, Burdea GC and Adamovich SV. Sensorimotor training in a virtual reality environment: does it improve functional recovery post-stroke? Neurorehabil Neural Repair 2006; 20: 252–267.

7. Reid DT. Benefits of a virtual play rehabilitation environ-ment for children with cerebral palsy on perceptions of self-efficacy: a pilot study. Paediatr Rehabil 2002; 5: 141–148.

8. Crosbie JH, Lennon S, McDonough SM and Basford JR. Virtual reality in stroke rehabilitation: Still more virtual than real. Disabil Rehabil 2007; 29: 1139–1146.

9. Henderson A, Kerner-Bitensky N and Levin M. Virtual reality in stroke rehabilitation: a systematic review of its effectiveness for upper limb motor recovery. Top Stroke Rehabil 2007; 14: 52–61.

10. Kuttuva M, Boian R, Merians A, et al. The Rutgers Arm, a rehabilitation system in virtual reality: a pilot study. Cyber-psych Behav 2006; 9: 148–151.

11. Houtsma JA and Van Houten FJ. Virtual reality and a haptic master-slave set-up in post-stroke upper limb rehabilitation. J Engl Med 2006; 220: 715–718.

12. Jang SH, You SH, Hallett M, et al. Cortical reorganization and associated functional motor recovery after virtual reality in patients with chronic stroke: an experimenter-blind prelim-inary study. Arch Phys Med Rehabil 2005; 86: 2218–2223.

13. Fischer HC, Stubblefield K, Kline T, et al. Hand rehabilita-tion following stroke: a pilot study of assisted finger exten-sion training in a virtual environment. Top Stroke Rehabil 2007; 14: 1–12.

14. Piron L, Turolla A, Agostini et al. Exercises for paretic upper limb after stroke: A combined virtual-reality and tele-medicine approach. J Rehabil Med 2009; 41: 10–16.

15. Hodkinson HM. Evaluation of a mental test score for assessment of mental impairment in the elderly. Age Ageing 1972; 1: 233–238.

16. Bhavani GCJ and Lincoln N. The reliability of the Riv-ermead Assessment and implications for some commonly used assessments of perception. Br J Occup Ther 1983; 46: 17–19.

17. Demeurisse G, Demol O and Robaya E. Motor evaluation in vascular hemiplegia. Eur Neurol 1980; 19: 382–389.

18. Sanchez-Blanco I, Ochoa-Sangrador C and Loez-Munain L. Predictive model of functional independence in stroke

at UZH Hauptbibliothek / Zentralbibliothek Zürich on June 30, 2014cre.sagepub.comDownloaded from

806 Clinical Rehabilitation 26(9)

patients admitted to a rehabilitation unit. Clin Rehabil 1999; 13: 464–475.

19. Crosbie JH, McNeill MDJ, Burke J and McDonough SM. Utilising technology for rehabilitation of the upper limb following stroke: the Ulster Experience. Phys Ther Rev 2009; 14(5): 336–347.

20. Winstein CJ, Rose DK, Tan SM, Lewthwaite R, Chui HC and Azen SP. A randomised controlled comparison of upper-extremity rehabilitation strategies in acute stroke: a pilot study of immediate and long-term outcomes. Arch Phys Med Rehabil 2004; 85: 620–628.

21. Lyle RC. A performance test for assessment of upper limb function in physical rehabilitation and treatment research. Int J Rehabil Res 1981; 4: 483–492.

22. Collin C and Wade D. Assessing motor impairment after stroke: a pilot reliability study. J Neurol Neurosurg Psy-chiatry 1990; 53: 576–579.

23. Lang CE, Wagner JM, Dromerick AW and Edwards DF. Measurement of upper-extremity function early after stroke: properties of the action research arm test. Arch Phys Med Rehabil 2006; 87: 1605–1610.

24. Nielsen SF. Nonparametric conditional mean imputation. J Stat Planning Inference 2001; 99: 129–150.

25. Pavy S, Brophy S and Calin A. Establishment of the mini-mum clinically important difference for the Bath Ankylosing

Sponylitis Indices: A prospective study. J Rheum 2005; 32: 80–85.

26. Moher D, Schulz KF and Altman DG. The CONSORT statement; revised recommendations for improving the quality of reports of parallel-group randomised trails. Lan-cet 2001; 357: 1191–1194.

27. Stewart R, Prince M, Brayne C and Mann A. Stroke, vascu-lar risk factors and depression. Br J Psychiatry 2001; 178: 23–28.

28. Hill KM, House AO and Hewison J. Recruiting a truly rep-resentative stroke cohort. www.leedsstrokedatabase.net/adrticale1.html (accessed 4 October 2007).

29. Schulz CH, Wasserman J and Ostwald SK. Recruitment and retention of stroke survivors; The CAReS experience. Phys Occup Ther Geriatr 2006; 25: 17–29.

30. Wolf SL, Lecraw DE, Barton LA and Jann BB. Forced use of hemiplegic upper extremities to reverse the effect of learned non-use among chronic stroke and head-injured patients. Exp Neurol 1989; 104: 125–132.

31. Jebsen RH, Trieschmann RB and Heward LA. An objec-tive and standardized test of hand function. Arch Phys Med Rehabil 1969; 50: 311–319.

32. Mathiowetz V, Weber K and Kashman N. Adult norms for the nine-hole peg test of finger dexterity. Occup Ther J Res 1985: 5: 24–37.

at UZH Hauptbibliothek / Zentralbibliothek Zürich on June 30, 2014cre.sagepub.comDownloaded from