Abstract
Aim
A novel interactive exergaming device enables children with cerebral palsy (CP) to exercise whilst standing. A small study showed gains in ankle motion after training with the device. This study explores the feasibility of a community-based randomised controlled trial (RCT) using this device to improve mobility in children with CP (Registered ISRCTN80878394).
Method
Fifteen children with CP Gross Motor Function Classification System (GMFCS) levels I-III were randomised 1:1 to either 10-weeks training with the device or usual care. Two potential primary outcome measures were assessed at 10-weeks with 20-week follow up: motion of the centre-of-mass estimate while stepping (Next Step) and Pediatric Balance Scale (PBS).
Results
Twenty-one children were assessed for eligibility, three declined to participate, two withdrew and one did not receive the intervention but remained in follow-up. Two serious adverse events occurred (n=2). Participant recruitment was 1.2 per month, limited by device availability. Completion of PBS and Next Step was 100% and 87% at baseline, dropping to 75% and 65% respectively at 10-weeks. The intervention group showed clinically meaningful changes in median (interquartile range) PBS at 10-weeks 2.7 (1) compared to the control 1.7 (0.67).
Conclusion
This study found multiple feasibility issues and further work is needed prior to implementing a RCT. There were signs of efficacy of the intervention, but findings are limited by small sample size.
What this paper adds
• Explores the feasibility of a community-based exergaming RCT • Investigates the acceptability of the interactive exergaming device • Establishes a primary outcome measure to assess balance in a definitive trial • Measures the safety of the intervention through reporting adverse events • Reports indicators of efficacy of the exergaming device as a physiotherapy intervention
Main text
Cerebral palsy (CP) is a childhood-onset disability, affecting 2-3 per 1000 live births worldwide.1,2 It results from a non-progressive neurological disturbance occurring in the developing infant’s brain, affecting the development of posture and movement. The Gross Motor Function Classification System (GMFCS) provides a common language to describe and predict motor ability. 3 Children with GMFCS levels I-II may walk independently while children with GMFCS level III walk with aids. Motor impairments associated with CP are spasticity, weakness and reduced selective movement, which interfere with various aspects of balance. Children with CP commonly undertake physiotherapy exercises to enhance their motor skills. Adherence to home exercise can present multiple challenges to family life. Rates of adherence to home exercise differ depending on the child’s motivation, parental support and the child’s environment. 4
Serious games and exergaming have gained increasing attention as innovative tools for rehabilitation, with a growing evidence base demonstrating their potential to enhance motivation, engagement, and motor outcomes across diverse populations..5–8 Exergaming describes digital games that require body movements to play the games. 6 Systematic review of the literature indicates that exergames can improve muscular strength and physical function. However, despite this general support, evidence specific to children with CP remains limited. Only a small number of studies have explored applications tailored to this group. 5 There is a need for larger, robust studies to establish the evidence base for exergaming in children with CP. 8
A novel interactive device (Figure 1) integrates a supportive standing frame with exergaming. It has a range of motivating games controlled by leg movements. The device offers a way for children with CP to train in a functional standing position, less reliant on adult support. The child operates games by moving the footpads (dorsiflexion and plantar flexion) and kneepads (forwards and backwards). Additionally, sensors under the feet allow certain games to be operated by shifting their weight side-to-side. The device has motors that assist or resist the movement at the knee and ankle. Physiotherapists may tailor the device settings to target training for strength, range of motion, selective motor control or balance during exergaming sessions. However, this intervention currently lacks evidence of efficacy in improving motor function.
9
Initial engagement work with families showed a preference for trialling this novel intervention at school or home, rather than through clinic attendance. Therefore, this study was designed to explore the feasibility of the intervention for a definitive trial in a community setting. Happy Rehab ™ device. Photographs reproduced with permission from Innovaid. The middle image shows one of the games on the laptop which operates the exergaming device.
The ACCEPT feasibility randomised controlled trial (fRCT) explored a 10-week community-based physiotherapy intervention using the Happy Rehab™. 10 The objectives of the feasibility study were to determine the acceptability of the intervention, to test the processes in the protocol and examine the appropriateness of the proposed outcome measures. This feasibility work was designed with reference to the Medical Research Council framework for developing complex interventions. 11 The study followed CONSORT guidelines, had approval from North of Scotland Research Ethics Committee (20/NS/0018), was registered ISRCTN80878394 and was funded by National Institute of Health and Care Research (ICA-CDRF-2017-03-041). Parents or carers have given informed consent to the research and to publication of the results.
Method
Eligibility criteria.
Key- Medical Research Council (MRC), Gross Motor Function Classification Scale (GMFCS).
Written informed consent was taken by the research physiotherapists before baseline assessment. Participants were randomly allocated to intervention or control by the unblinded assessor, using a computer-generated randomisation model (REDCAP™), at a ratio 1:1. Randomisation minimised for GMFCS level (I and II versus III) and age (above or below 9 years). Children in the intervention arm were asked to train using the Happy Rehab™ for 20 minutes, three times per week for 10 weeks and to record their training in an e-diary, using a pseudonym to maintain confidentiality. The unblinded assessor enrolled participants, arranged for the transport of the device and booked the initial appointment with their usual physiotherapist to set up the Happy Rehab™ device. The control group continued with usual care, 12 typically involving the prescription of a home/school programme of active stretching and strengthening exercises.
Two potential primary outcome measures were assessed at 10-weeks, with follow up at 20-weeks by a physiotherapist blinded to allocation. They were medio-lateral motion of the centre of mass (COM) estimate and centre of pressure (COP) while stepping onto one of two targets (Next Step test) 13 and the Pediatric Balance Scale (PBS). 14 Secondary outcomes included lower limb range of motion, spasticity and strength using goniometry, Modified Tardieu Score (MTS), and dynamometry. Additionally gait kinematics, goal-setting with the Canadian Occupational Performance Measure (COPM) 15 and the CHU-9D quality of life measure. 16 Data were collected at community physiotherapy departments. Reportable adverse events included pain or injury to the legs lasting more than one hour and fatigue lasting more than a day following training. Data were analysed using descriptive statistics as per the statistical analysis plan. 10 Qualitative interviewing took place as part of this feasibility study to explore the experiences of participants.
Planned recruitment varied from the published protocol as the commencement of the study was delayed, due to the COVID-19 pandemic. The recruitment targets reduced from forty children over five sites to sixteen children over two sites.
Results
Recruitment was carried out at each site until saturation was reached. There were three sizes of device with different weight limits. During the baseline assessment, the child’s height and weight were measured to determine which frame would be needed. Therefore, potential participants who expressed an interest were held on a waiting list until a suitable device became available.
Participant characteristics at baseline.
Key- Standard deviation (SD), number (n).
The retention and reason for drop out of the participants is shown in the CONSORT diagram (Figure 2). Recruitment rate was limited by the number and size of available devices. Three families declined: two were too busy and one did not want to stop usual care. In the intervention group one child trained at school while the others chose to use the device at home. One participant did not receive the intervention, as the device would not fit through an inner doorway. They declined to come into the Physiotherapy department to train due to travel and time constraints, but the participant continued with follow-up. CONSORT Diagram showing the recruitment and retention to the ACCEPT fRCT.
The Shaprio-Wilk test showed that the PBS (and PBS change) data were not normally distributed. The Next Step Medio-lateral (ML) COP (and ML-COP change) was normally distributed except for two timepoints: intervention group at 10 weeks and control group change score at 10 weeks.
At baseline, all participants reached maximum scores (ceiling) on the first four items of the PBS, and change was only measurable over the last eight items of PBS. Two participants from the control group reached the ceiling PBS score at baseline and one from the intervention group at 10 weeks.
Completeness and change in the two potential primary outcome measures at baseline, 10 and 20 weeks.
Key-number (n), interquartile range (IQR), percentage (%), score greater than minimal detectable difference (*) standard deviation (SD), millimetres (mm), Confidence Interval (CI)
The MDD for peak ML-COP ranges from 23.7-29.6mm depending on the position of the target. 13 All increases in ML-COP were below MDD. Both groups showed an increase in ML-COP grand average scores at 10 weeks with the intervention group increasing further at 20 weeks (Table 3).
Changes in clinical impairment measures can be found in the supplemental table (S1). Overall, there was a slightly higher increase in passive range of motion and muscle strength in the intervention group. The CHU-9D score was similar between groups at baseline with no notable change at 10 weeks. Children in the intervention group improved their COPM performance scores by a mean 0.6 (SD 0.9) points and satisfaction by 1.1 (1.2) points at 10 weeks. Gait kinematics showed no notable change at 10 weeks.
At baseline, the PBS, MTS and hand-held dynamometry had complete data sets. The Next Step test data was complete in the control group at baseline; however, one child from the intervention group declined this test. A child with autistic spectrum condition (ASC) was not able to complete all PBS tasks due to their autism, so their results were excluded. Gait kinematic data was incomplete at baseline due to technical difficulties and two children declining to participate. The PBS and MTS had the most complete data sets across all time points. There was one instance of unblinding where the child told the assessor about their allocation.
The intervention group recorded a mean of 20.5 (SD 17.3) minutes per session, undertaking a mean of 1.32 (SD 1.30) sessions per week. The control group recorded a mean of 35.1 (SD 24.5) minutes per session, over a mean 3.5 (SD 2.5) sessions per week. The control group recorded that they exercised 51% more than the amount of exercise the intervention group were asked to do. The intervention group only recorded 34.2% adherence to the recommended exercise intensity. One participant was unable to undertake the intervention, and one child recorded inability to train on one occasion due to faulty device. Physiotherapists recorded that the intervention group required a mean 2.3 (SD 2.1) hour support over 10 weeks compared to 1.5 (SD 1.9) in the control group. One child sustained a bruised leg while getting out of the exergaming device unsupervised. The bruises resolved within a few days and did not require treatment. One child in the control group sustained a fractured metatarsal, unrelated to trial activities.
Discussion
To our knowledge this is the first study of a community-based resistance exergaming intervention aimed at improving walking and balance in children with CP. Previous work with this exergaming device explored the effect of a clinic-based intervention on ankle range of movement and gait. 9 Other studies of community-based exergames used virtual reality or commercially available games that involve exercise without resisted movement8,18 and cycling. 19 Robotic gait training19,20 is unfeasible for community use; therefore, this novel interactive exergaming device offers a unique opportunity as a community-based intervention.
This study found multiple challenges to the implementation of the trial. The trial was initially planned to run between 1/5/2020 to 1/8/2022. However, the COVID-19 pandemic affected the opening and duration of the trial. This resulted in a shorter recruitment period, with the recruitment target reduced from 40 to 15, and the number of sites was reduced from five to two. Britain’s exit from the European Union (Brexit) resulted in the UK distributer withdrawing the device from their product sales. This meant that the trial was run without the expected technical support, however, repairs and support were accessed through the product designer. Physiotherapists spent more time delivering the intervention than usual care, requiring support to problem-solve, such as the game appearing not to respond to the controls. The solution was to allow the device to recalibrate without the child standing in it. Other studies using rehabilitation devices at home have required high levels of technical support.21,22 Implementation of the exergaming device into clinical practice or a definitive trial would be unfeasible without effective technical support in place.
A randomisation model 1:1 was used. Consideration was given to contamination such as where several participants could be in in one place, for example, in a special school or where there were eligible siblings. However, these problems were not encountered in practice. For a larger study, consideration should be given about whether to use cluster randomisation in these situations. Minimisation of age over or under 9 years of age and GMFCS levels I and II versus GMFCS level III was used. Only one participant of GMFCS level III was recruited so this limited the effect of the minimisation process. In a larger trial, this would still be an appropriate model to use.
Withdrawals and loss to follow up related to services reopening after the COVID-19 pandemic. Families found that there was a sudden increase in appointments and Physiotherapists delivering the trial experienced increased pressure from re-opening services.
This feasibility study evaluated a complex intervention. 23 A key component of this intervention was the clinician’s ability to set up the intervention and coach the child and family to use it. Previous studies have shown the importance of pre-trial training and education for the clinicians and manualised interventions. 24 A future trial would need to include a more detailed and flexible training programme and the development of a ‘champion’ with more in-depth skills who could support their colleagues.
A strength of this study was the inclusion of a wide range of children, including those with additional communication needs. Four children with co-morbidities were recruited; hearing impairment, epilepsy, learning disability and ASC and these are common in children with CP.
There is an emerging body of evidence to support ‘serious games’ in therapy.7,25 Similar therapeutic games reported in the literature use low cost and portable virtual reality games that might be used at home, while others focus on expensive lab-based robotic movement training.18,20,26 The former may have little guided movement or support, where the latter has fully supported and guided movement. This device sits somewhere in between, as it involves supporting the child in standing position while they operate the games by using movements assisted or resisted by motors in the footplates and kneepads. It is important to note that this device offers potentially unique treatment options for children who may not be able to otherwise sustain training in an upright standing position without support. Most children were highly motivated by the games and virtual rewards in the devices games. These included virtual ‘badges’ to represent their scores and positive messages that celebrated their level of achievement. Studies have shown that the inclusion of gaming such as Virtual Reality games can increase the repetitions of therapeutic exercise. 27 Physiotherapists delivering the intervention needed to pitch the challenge of the training at the right level of difficulty, also called the ‘just right challenge’. Where the challenge is too high or perceived as unachievable, this can have a de-motivating effect. Setting the challenge too low may limit physiological training effects. It is essential to be able to tailor rehabilitation games to meet the needs individual patients, and to be able to progress their training. 28
Nearly all parents were willing to accommodate the device for a 10-week period, except when it was in direct competition with the family’s needs for space, such as at Christmas. Parents have previously expressed a willingness to sacrifice space or the aesthetic of their home to accommodate large equipment, where they consider it benefits their child 29 parents felt that their child should not use the device at school as it might single them out. Goodwin et al., 2019 found that teachers were concerned about the competing demands of using rehabilitation devices on the curriculum as well as space. 30
One weakness in the protocol was that the intervention group stopped accessing the intervention at 10 weeks, but the control group were able to continue training. This may have affected the results. One parent declined for his son to participate because he was concerned about stopping his usual care. Children with CP often have well-established exercise programmes at home and school. It would be feasible to deliver this exergaming intervention as a burst of treatment in addition to usual care or to reinstate the usual care programme after the cessation of the intervention.
Children found the gaming aspect motivating and enjoyable. Parents and physiotherapists agreed that the exciting potential of gaming to enhance exercise training. 31
The exergaming device has the potential to support children to undertake exercise that targets multiple impairments. The feasibility trial findings show some indications that balance functions, flexibility and muscle strength may improve with training using the exergaming device. The exergaming device may offer a unique opportunity for children with CP to train in a supported standing position. In this study we only recruited one child with GMFCS III. This is because we had a small sample size and because the incidence of GMFCS level III is only 11.5% of all GMFCS levels. 32 Therefore, a larger study is also likely to have more children GMFCS II than III and would need to stratify randomisation and analysis to address this. The PBS and measures of clinical impairment were the outcomes most acceptable to participants. The PBS captures changes in functional balance tasks in children GMFCS II and III, however the PBS ceiling was reached in some children with GMFCS I. A definitive trial should focus on these children using the PBS as the primary outcome with stratification by GMFCS level and age.
Future trial design would include children GMFCS II and III who are able to access the device at home or school. It should include a manualised training package for clinicians, sufficient availability of devices in the three sizes and responsive technical support.
Conclusions
This trial tested the feasibility and acceptability of a trial exploring a novel interactive training device. This novel exergaming device presents an opportunity for children to undertake a personalised training programme targeting their underlying impairments, while training in a functional position.
There were multiple feasibility issues during the study, some related to the COVID-19 pandemic and Brexit. This resulted in the small sample size, baseline imbalance, high dropout and missing data rates, low adherence in the intervention group. Additionally, some families raised concern about accommodating the device at home or within an educational setting and physiotherapists required more training and support to confidently set up and adjust the device settings. This study found that a full RCT would not be feasible without adequate and timely technical support and requires further feasibility work prior to a definitive trial.
Supplemental material
Supplemental material - A feasibility randomised controlled trial of an exergaming device aimed at improving mobility in children with cerebral palsy
Supplemental material for A feasibility randomised controlled trial of an exergaming device aimed at improving mobility in children with cerebral palsy by Rachel Rapson, Bernie Carter, Harriet Hughes, Jos M. Latour, Wendy Ingram and Jonathan Marsden in Digital Health.
Footnotes
Acknowledgements
Thank you to the families involved in the steering group and to the participants who took part in this study. Thanks to the National Institute for Health and Care Research for funding this work. The authors acknowledge the NIHR for funding this work (ICA-CDRF-2017-03-041). Innovaid (Denmark) have given written permission for the publication of their photographs.
Ethical considerations
This study was approved by North of Scotland Research Ethics Committee (20/NS/0018).
Author contributions
Rachel Rapson, Jonathan Marsden, Harriet Hughes, Jos Latour, Wendy Ingram and Bernie Carter have contributed to the design, writing, drafting of the protocol and all gave final approval to the publication of this protocol.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Institute for Health and Care Research Clinical Doctoral Research Fellowship (ICA-CDRF-2017-03-041).
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
Full data is available on request from the lead author.
Supplemental material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
