Objective To assess the feasibility and effectiveness of home-based occupational therapy and physiotherapy programmes in children with cerebral palsy (CP), focusing on the upper extremity and reporting on child-related and/or parent-related outcomes.
Design Systematic review.
Data sources Electronic searches were performed in MEDLINE, EMBASE, CINAHL, PsycINFO, OTseeker and PEDro, and in ICTRP and CENTRAL trial registers, from inception to 6 June 2019.
Eligible criteria The review included all types of original studies concerning feasibility or effectiveness of home-based therapy in children aged <18 years with any type of CP. No language, publication status or publication date restrictions were applied.
Data extraction and synthesis Study and intervention characteristics and the demographics of participating children and their parents were extracted. Feasibility was assessed by outcomes related to acceptability, demand, implementation, practicality, adaptation, expansion or integration. Regarding effectiveness, child-related outcome measures related to any level of the International Classification of Functioning, Disability and Health, or parent-related outcomes were investigated. Two authors independently extracted the data. Risk of bias was assessed using the Downs and Black checklist and the Joanna Briggs Institute Critical Appraisal Checklist.
Results The search resulted in a total of 92 records: 61 studies and 31 conference abstracts. Feasibility studies reported mainly on acceptability and implementation. Overall compliance to home-based training programmes (implementation) was moderate to high, ranging from 56% to 99%. In the effectiveness studies, >40 different child-related outcome measures were found. Overall, an improvement in arm-hand performance within group across time was shown. Only two studies reported on a parent-related outcome measure. No increase in parental stress was found during the intervention.
Conclusions Based on the results of the included studies, home-based training programmes seem to be feasible. However, conclusions about the effectiveness of home programmes cannot be made due to the large variability in the study, patient and intervention characteristics, comparators, and outcome measures used in the included studies.
PROSPERO registration number CRD42016043743.
- rehabilitation medicine
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
Strengths and limitations of this study
This is the first review to be systematic as well as specifically focused on the feasibility and effectiveness of home-based occupational therapy and physiotherapy programmes in children with cerebral palsy.
Besides child-related outcomes, this review also included parent-related outcomes.
We were unable to perform a meta-analysis due to the large variability in study characteristics.
Over the last years, despite an increased survival rate of low birthweight infants, the overall prevalence of cerebral palsy (CP) has remained constant at 1.96 per 1000 live births.1 CP is the largest diagnostic group treated in paediatric rehabilitation. Social participation, independence and self-efficacy are restricted in children with CP as they experience limitations in the execution of daily activities.2 About 60% of children between 4 and 16 years have problems with effective use of the arm and hand during reach, grasp, release and manipulation of objects, resulting in limitations in performance of daily activities.3 4 Most currently applied upper extremity interventions aim at improving functionality and abilities towards independence. Studies examining these interventions have shown that the key ingredients for effective treatment constitute a high training intensity combined with meaningful goal-directed and task-specific training.5 Relevant context for children to learn new daily activities is usually the home environment, and interventions provided in this context are called home-based programmes.6 7 Home-based programmes are defined as ‘therapeutic activities that the child performs with parental assistance in the home environment with the goal to achieve desired health outcomes’.7
Home-based programmes are thought to be a useful addition or even replacement of centre-based therapy in the rehabilitation of children with CP.5 Home-based programmes provide a unique opportunity to train continuously, and specific tasks are trained in a relevant context. Furthermore, these programmes enable parents to incorporate training into their daily routine with the child, so no separate training moments are necessary, generalisation is fostered, and intensity and repetition of trained tasks can be high, which all enhance effective motor learning.8 In addition, increased amount of training may facilitate retention of established intervention effects. Furthermore, it may also increase parental involvement and empowerment, in turn contributing to reciprocal partnerships between parents and health professionals.9
Despite consensus on the importance of home-based programmes for children with CP, there is scarce information regarding programme characteristics that may influence family participation.10 For example, parents can be either a therapy provider in collaboration with a health professional (partnership home programme) or supervised by a health professional (therapist-directed home programme).11 When parents become therapy providers, the relationship between parents and the health professional changes: the health professional becomes the coach of the parents. Depending on the role of parents and their specific needs, the way and amount of coaching can vary from limited instruction only at the beginning of the programme, to extensive demonstration, feedback and coaching throughout the entire programme. Mode of coaching can vary from home visits by the therapist to remote coaching by email or telephone consultation.
Parents are of great importance in home-based programmes. Although a survey among parents has shown that they do not have an unfavourable opinion concerning home programmes, these programmes may induce or enhance stress in parents.11 Parents may experience pressure to comply, especially when the programme is demanding. Furthermore, the altered parent–child interaction during training may cause additional tension.12 As the role of parents changes to that of a therapy provider, this may cause a conflict between their parenting style and their approach as a therapy provider. Consequently, loss of motivation by parents and/or child to complete training activities may affect compliance and probably effectiveness of the intervention. Because of the aforementioned factors, home-based interventions need to be carefully developed and implemented.
Feasibility is an important aspect that needs to be considered when implementing home-based programmes. Feasibility studies are used to determine whether an intervention is relevant, sustainable and appropriate for further testing.13 Several studies have investigated the feasibility of home-based programmes for children with CP and indicated that the programmes were feasible in terms of compliance and adherence.14 15 However, up until now no systematic overview is available of relevant feasibility components, such as satisfaction, acceptability or practicality, and even when these treatments appear feasible they are not necessarily effective. So far, effectiveness of home-based programmes in children with CP has been reviewed by Novak and Berry.7 They concluded that home-based programmes using goal-directed training are effective in improving motor and functional outcomes.7 Another review by Sakzewski et al5 on non-surgical upper extremity therapies in children with unilateral CP concluded that home-based programmes are an effective supplement next to centre-based interventions.
Supplementary to these two reviews, this systematic review aims to provide a clear summary on both feasibility and effectiveness of currently available home-based programmes in children with CP (aged <18 years), specifically focusing on the upper extremity. Effectiveness will be investigated on both child-related and parent-related outcomes, as parent involvement has received little research attention.
The following two objectives will be addressed:
To assess the feasibility of home-based occupational therapy and physiotherapy programmes in children with CP.
To assess the effectiveness of home-based occupational therapy and physiotherapy programmes that focus on the upper extremity in children with CP in child-related and parent-related outcomes.
The objectives and methods of this review were prespecified and registered in the International Prospective Register of Systematic Reviews (PROSPERO), as well as published in a protocol.16
Types of studies: all types of original studies concerning feasibility or effectiveness of home-based therapy in children with CP. An intervention was considered to be home-based if treatment was performed in the home setting without a healthcare provider being physically present. Studies that only included therapy provided at a healthcare facility, (pre)school or day care were excluded. In case the intervention took place in different settings, studies were only included if treatment in the home setting was a fundamental, prespecified element of the intervention. The studies included in this systematic review were categorised using the scale published by the American Academy for Cerebral Palsy and Developmental Medicine to hierarchise studies based on research design types of either intervention (group) studies or single-subject design studies.17
Types of participants: children aged <18 years with any type of CP. In case of a more heterogeneous study population, results of the target population must have been reported separately.
Types of intervention: home-based occupational therapy or physiotherapy intervention performed in the home setting without (continuous) physical presence of a healthcare provider. To investigate effectiveness, only upper extremity interventions were included.
Types of comparators: concerning feasibility, studies comprising all types of comparators or no control intervention were considered. In order to determine effectiveness, no therapy, care as usual, centre-based occupational therapy or physiotherapy, pharmacological intervention, and surgical procedure were considered. If a study comprised multiple distinct home-based programmes, the one of main interest was included as the experimental intervention and the other home-based programme(s) as comparator(s).
Types of outcome measures: to review feasibility, studies reporting on key areas as proposed by Bowen et al13 were considered: acceptability, demand, implementation, practicality, adaptation, expansion or integration. Regarding effectiveness, child-related outcome measures related to any level of the International Classification of Functioning, Disability and Health (ICF), or parent-related outcomes within the psychological and social domain including parenting, were investigated.18
Report criteria: no restrictions regarding language, publication status or publication date were applied. Conference abstracts that provided insufficient information to decide on selection were excluded, as well as records of which the full text could not be retrieved.
Records were identified using electronic databases MEDLINE (Ovid interface; 1946–present), EMBASE (Ovid interface; 1974–present), Cumulative Index to Nursing and Allied Health Literature (CINAHL) (EBSCO interface; 1981–present), PsycINFO (EBSCO interface), OTseeker and PEDro. Trial protocols were also identified through International Clinical Trial Registry Platform (ICTRP) and Cochrane Controlled Trials Register (CENTRAL). Moreover, reference lists of included papers, excluded reviews and meta-analyses were scanned. Finally, a bibliography of included records was sent to all corresponding and last authors of included studies. They were asked to provide any related study by either their own research group or associates.
Search terms for population and intervention were combined for Medical Subject Headings (MESH) terms and text words in titles and abstracts (online supplementary appendix 1). Search strategies were created by LWMEB and revised after peer review by JK. A data search expert from Kleijnen Systematic Reviews conducted the search on 10 October 2016, and an update of this search was done on 6 June 2019.
The software platform Covidence was used to complete eligibility assessment. LWMEB and MLAPS independently executed the screening of titles and abstracts as well as the unblinded evaluation of full-text publications in duplicate. Any disagreements between reviewers were resolved through consensus and arbitrated by YJMJ-P, when necessary. Inter-rater agreement and reliability were calculated using percentage of agreement and Cohen’s kappa statistic to determine consistency between reviewers in assessing the eligibility of full-text publications.
Data collection process
LWMEB and MMEG collected data independently for each study. A data extraction form was developed a priori, pilot-tested on two records that were not eligible for this review, and refined accordingly. During data collection reviewers discussed any discrepancies and consulted YJMJ-P to mediate when necessary. Authors were contacted if essential information was missing from a study or if reports were inconsistent. Author names, intervention locations, intervention characteristics, sample sizes and outcomes were compared to identify duplicate publications. Multiple records reporting on different outcomes or time points of one study were combined. For records investigating the same outcomes and time points, only the record reporting the largest sample size was included.
General information was extracted from each included study: (1) study characteristics (author(s), publication year, study design, country, comparator, number of participants (in total and per study arm), outcomes, follow-up duration and measurement time points); (2) intervention characteristics (objective, therapy provider(s), coaching approach of parents, duration of programme, frequency and duration of sessions, treatment approach, and motor learning approach); (3) demographics of participating children (age, gender, diagnosis (type and topographical distribution of CP), Manual Ability Classification System (MACS) level, Gross Motor Function Classification System (GMFCS) level, Communication Function Classification System level); and (4) demographics of parents of participating children (age, gender and educational level).
Feasibility was assessed primarily by outcomes related to the feasibility area, whereas demand, implementation, practicality, adaptation, integration and expansion were of secondary interest. Definitions of these constructs are provided in the protocol.16 Concerning the effectiveness objective, child-related upper extremity outcomes within the ICF level activity were primary. Outcomes assessing body functions and structures, participation, and parent-related outcomes were of secondary interest.
Home-based programmes are often complex interventions, formed by multiple interacting components. For that reason, if results were reported separately for particular components of the intervention, this was also recorded.
Risk of bias in individual studies
The Joanna Briggs Institute (JBI) Critical Appraisal Checklist for Qualitative Research was used to determine risk of bias of qualitative studies.19 Studies with primary focus on intervention effectiveness were assessed by the Checklist for Measuring Quality by Downs and Black.20 Construct power was not included, since this item estimates precision rather than bias. Single items were summarised into overall scores, and each study was classified into excellent (24–28 points), good (19–23 points), fair (14–18 points) or poor (<14 points).21 All assessments were done at study level. LWMEB and MMEG performed the unblinded assessment independently. In case reviewers could not come to an agreement, YJMJ-P interceded.
For effectiveness studies included in the review, the risk of selective reporting was determined by comparing records on study results with previously published study protocols or registrations. Any discrepancies were listed.
Patient and public involvement
Patients and the public were not involved in our research.
The search resulted in 3077 records. After deduplication, a total of 2054 titles and abstracts were screened, resulting in 1779 irrelevant records. The remaining 275 records were full texts assessed for eligibility, of which 183 records did not meet the eligibility criteria. The search resulted in 92 records, some reporting on the same study. The flow chart is depicted in figure 1.
There were 83 corresponding and last authors contacted to provide any related studies. Of these authors, 49 (59%) responded with either a suggestion or no additions at all, resulting in 22 additional records, which are already included in the 92 records.
Inter-rater agreement of full-text assessment was found to be 83.3%. Inter-rater reliability was substantial (Cohen’s kappa 0.66).
Of the 92 records, 31 records22–52 were conference abstracts. Eight initial studies described in these abstracts22–24 31–34 44 developed into a full-text article (25.8%). The remaining 61 studies11 14 15 53–110 were included in this review, 30 feasibility studies11 14 15 53–71 98 99 101 102 105–108 (49.2%), 10 effectiveness studies87–96 (16.4%), and 21 studies72–86 97 100 103 104 109 110 that reported on both feasibility and effectiveness (34.4%).
Of the effectiveness studies, 2 studies76 95 (6.5%) were large randomised controlled trials (RCTs), 24 studies72–75 77–79 81–88 90 92 93 97 100 103 104 109 110 (77.4%) smaller RCTs, 4 studies89 91 94 96 (12.9%) were single-subject designs, and 1 study80 (3.2%) used a pretest–post-test cohort design, with the participants serving as their own controls (see table 1).
Methodological quality of studies with a primary focus on intervention effectiveness, assessed by the Downs and Black checklist, is depicted in online supplementary appendix 2. According to this scale, 5 studies75–77 85 86 (16.1%) were rated as good, 15 studies73 74 78 79 81–83 87 88 92 95 97 100 103 110 (48.4%) were fair and 11 studies72 80 84 89–91 93 94 96 104 109 (35.5%) were poor. The 13 qualitative studies11 55 57–59 61–64 67 70 72 101 found were scored with the JBI Critical Appraisal Checklist to determine risk of bias. A positive answer to the first five questions of this checklist is crucial for the assessment of risk of bias. Scores are given in online supplementary appendix 3. In only five qualitative studies59 62–64 101 (38.5%), the first five questions of the JBI checklist could be answered. In other words, risk of bias in these five studies was clear, whereas in eight studies11 55 57 58 61 67 70 72 (61.5%) this risk could not be estimated from the data provided. Records on study results were compared with previously published study protocols or registrations. Chiu et al98 stated that therapy sessions lasted 20 min, while they stated in the trial registration that therapy sessions lasted 25 min. Several other studies showed a discrepancy in the amount of outcome measures reported. They reported either less or more outcome measures in the trial registration than in actual study results.
Most studies targeted children with unilateral spastic CP, but there was a large variation in other child characteristics such as age, MACS and GMFCS classification. The vast majority of studies did not report any parent characteristics. Only two studies54 101 reported on age, gender and educational level of parents. Only 16% of the studies reported on gender characteristics, and only 7% reported on educational level. The number of study participants ranged from 1 to 147, with a maximum of 105 in an effectiveness study. All participant characteristics are shown in table 1.
In table 2 intervention characteristics of the included studies are shown. One should note that all characteristics described in the tables and the results apply to the parent-delivered part of the intervention only. A more detailed description of the intervention is provided in online supplementary appendix 4.
The treatment approach used in the studies was predominantly (modified) Constraint-Induced Movement Therapy (CIMT) (32.8%),55 65 72–75 78–80 84–87 89 91–95 97 and several studies60 67–69 71 76 82 88 98 99 102 104 107 also used computer-based rehabilitation (eg, virtual reality, 22.9%). Very few studies used goal-directed (n=2)53 83 or bimanual (n=3)56 100 110 training. Comparators used were none (feasibility studies), other home-based programmes, care as usual, centre-based occupational therapy or physiotherapy interventions. The objectives of the intervention were mostly unspecified, but when specified the focus was mainly on ICF activity level. The use of motor learning principles was often not mentioned; only 20 studies55 56 72–80 83–87 93 97 100 110 (32.8%) reported that their intervention was based on motor learning principles. Training duration of home-based programmes varied from 2 weeks to 6 months (all parent-delivered), and intensity ranged from 70 min to 56 hours a week (all parent-delivered). Therapy was mostly provided by parents (55.7%), but there were also programmes combining parent-delivered and therapist-delivered sessions (41%). In the latter, the main part of sessions were delivered by parents. Coaching of parents was often unspecified (49.2%). Some studies mentioned different modes that were used by therapists to coach parents, such as course/training, manual or other form of written instructions, DVD, reviewing of logbooks, email, telephone or Skype calls, home visits, computer feedback, and mutual discussion of goals and therapeutic activities.
Feasibility studies mainly reported on the key areas of acceptability and implementation, and some on demand and practicality. None of the studies reported on the areas of adaptation, integration or expansion. Overall compliance to home-based programmes (implementation) was moderate to high, ranging from 56% to 99%.14 54 56 60 61 70 71 98 99 106 108 Majority of studies reported that parents found it easy to carry out the programme and enjoyed seeing their children improve (acceptability). Some studies reported on the demand and mainly on the recruitment rate, which ranged between 45% and 83%.98 106 One study reported on the safety (practicality) of the programme. During the programme no serious injuries occurred; children only experienced muscle soreness and were more fatigued.98
In the effectiveness studies, more than 40 different child-related outcome measures were found. Child-related outcome measures on ICF activity level were considered to be primary outcome measures in this review. There were 15 different primary outcome measures found, that is, Quality of Upper Extremity Skills Test (17×), Assisting Hand Assessment (15×), Canadian Occupational Performance Measure (10×), Melbourne Assessment of Unilateral Upper Limb Function (7×), Goal Attainment Scaling (4×), Pediatric Motor Activity Log (4×), ABILHAND-Kids (4×), video observation (3×), Shriners Hospital for Children Upper Extremity Evaluation (1×), Assessment of Motor and Process Skills (1×), Functional Inventory (1×), Box and Blocks Test (1×), Jebsen-Taylor Hand Function Test (1×), test of sensation (1×) and Children’s Hand-use Experience Questionnaire (1×). The vast majority of these outcome measures showed an improvement in arm-hand performance within group, across time, that is, before and after intervention. However, in case of effectiveness, this improvement (within group) was not always sufficient to identify a difference between the interventions investigated (between groups).
Except for Hsin et al74 and Novak et al,81 who reported on the results of Cerebral Palsy-Specific Quality of Life (parent-proxy version) and Children’s Assessment of Participation and Enjoyment, respectively, none of the studies included outcome measures on ICF participation level. Both studies reported gains in health-related quality of life. All other outcome measures were on ICF function level. Again, majority of studies showed a positive change in hand function, within group, before and after intervention, but a difference in effectiveness between interventions could not always be confirmed.
In contrast to the large amount of child-related outcome measures, only two studies56 79 reported on a parent-related outcome measure, that is, Parenting Stress Index-Short Form. Lin et al79 and Ferre et al56 found no increase in parental stress during the intervention.
A detailed description of the results of feasibility studies, effectiveness studies and studies that reported on both feasibility and effectiveness is given in tables 3–5. Furthermore, the completed data extraction form can be obtained from the authors.
This systematic review aimed to assess both the feasibility and effectiveness of home-based occupational therapy and physiotherapy programmes in children with CP, specially focusing on upper extremity. The objective was to investigate all relevant feasibility components according to Bowen et al,13 not only whether home programmes were feasible in terms of compliance and adherence, as is most commonly reported. However, only a few studies mentioned the feasibility outcomes demand and practicality. None of the included studies reported on the other aspects. Based on the implementation and acceptability results of the included studies, home-based programmes seem to be feasible. Overall compliance to home-based programmes was moderate to high, ranging from 56% to 99%. Farr et al99 and Lorentzen et al,60 who found the lowest compliance (56% and 62%, respectively), reported that technical problems and the fact that children were sometimes too tired or upset to complete the virtual reality training were the main reasons for the difference between the actual amount and intended amount of training. The high compliance (96.1%) reported by Ferre et al56 may be due to the fact that they employed a strict selection of participants. Eleven parents and their children met the inclusion criteria and were willing to commit to the programme requirements. One family dropped out after 4 weeks because the programme was too demanding. Adjoining, they provided intensive coaching sessions to parents. Chiu et al98 reported a compliance of 99%. This may be due to the fact that the therapy demand was low: only 20 min a session, three times per week, over 8 weeks. In addition, both parents and children were highly satisfied with the therapy. Overall, studies reported that parents were positive about their experiences with the programmes. They found it easy to carry out the programme and enjoyed seeing their children improve. However, there were also parents who found it difficult to incorporate the programme in their daily life routine. Parents indicated that it was difficult to find enough hours in a day to perform the programme next to their daily activities.55 When the parent who delivered the programme got support and help from other family members, it was easier for them to implement the training in their daily routine.66 Despite these difficulties reported, general parental stress did not increase during the intervention.56 58
Conclusions about the effectiveness of home programmes cannot be made due to the large variability in the study, patient and intervention characteristics, comparators, and outcome measures used in the included studies. Even within the same treatment approach, frequency and duration of the interventions varied. As training intensity is an important predictor for treatment success, improvement in arm-hand function and performance can therefore not be solely attributed to the intervention approach.
Many different treatment approaches were found in the included studies. Majority of studies reported on the effectiveness of (modified) CIMT, whereas only three studies56 100 110 investigated the effect of bimanual training. Both treatment approaches have shown to be effective in clinical rehabilitation. However, most daily activities require bimanual use of hands. Therefore, an intervention focusing on the coordinated use of both hands in bimanual activities may have more impact on the child’s daily life than a modified CIMT programme focusing on improving the capacity of the affected hand.
According to Sakzewski et al,5 upper limb interventions in children with unilateral CP should be goal-directed, adequately dosed and based on motor learning approaches that use activity-based therapy. Most studies found in this review did not specify whether their intervention was based on motor learning principles. Some studies indicated that they used shaping and repetitive task practice, implying that the intervention was based on motor learning principles. The question which motor learning approach in the specific context of parent-delivered programmes is best suitable, remains, therefore, unanswered. Protocols from existing intramural programmes may not always be feasible in a home setting, where parents are supervising the training of the child. They need to instruct their children and prompt the use of the affected hand over and over again. Continuous prompting may pose an important stress factor on parents.111 Studies on basic motor learning in children with movement disorders have shown that implicit motor learning has positive effects on motivation and compliance and may therefore be better suited for a home setting.112–114 There is also evidence indicating that children with CP often have problems with working memory, making it difficult for them to learn in an instruction-driven way.115 Moreover, implicit learning may lead to increased self-efficacy, which is important for motivation and compliance. Parents and clinicians rate motivation as the most influential personal characteristic, determining outcome and treatment adherence.116 An implicit motor learning approach seems very promising and should be explored in future studies.
Coaching of parents is a key element of home-based programmes. When parents are effectively coached by therapists and guided throughout the training period, parents become more confident in carrying out the home-based programme and find it easier to implement the programme in their daily routine.11 66 Surprisingly, information on how parents were coached to be therapy providers was lacking in a lot of the reported studies. Perhaps coaching received little attention during the interventions. Information on parent characteristics was also hardly given. Inferences about why some parents find it easy to carry out a home programme while others struggle with finding ways to do so cannot be made. The fact that only two studies56 79 reported on a parent-related outcome measure is also surprising given the major role of parents in the execution of a home-based programme.
In conclusion, one can state that a detailed description of home-based training protocols in most intervention studies is lacking. An extensive description of interventions tested may take up many words, but provides crucial information that increases our understanding on the working mechanism of an intervention. We therefore plea in favour of writing protocol papers before publishing results.
Due to the large variability in study, participants and intervention characteristics, as well as child-related outcome measures found in the included studies, a meta-analysis on outcome measures was not possible. Although home-based training seems to be promising as most studies showed positive changes in child-related outcome measures, hard evidence on the effectiveness of these programmes cannot be given. This also means that guidelines to improve existing home-based programmes or to develop new home programmes are still awaited. As no synthesis of evidence was possible, the Grading of Recommendations Assessment, Development and Evaluation guidelines to judge the quality of evidence was not relevant and could not be used.117 With this, the review deviates from the protocol published by Beckers et al.16
Recommendations for future research would be to develop a core set of outcome measures incorporating all ICF levels to investigate the effects of interventions. In addition, the outcome measures should be validated for the total population of children with CP, including all types of CP, and should have good usability. Furthermore, parent-related characteristics, intervention elements and outcome measures should be part of and described in detail in studies investigating home-based programmes. Finally, future studies should focus on the comparison of two different home-based programmes using a different motor learning approach while keeping aforementioned characteristics the same.
We would like to thank Lisa Stirk for performing the literature searches, and also Tuba Aliskan Turkbey for screening two Turkish abstracts, Helena Gunther for screening one Russian abstract and Cinara Sacomori for screening one Portuguese abstract.
LWMEB and MMEG contributed equally.
Contributors LWMEB and JK developed the search strategy. JK, EAAR, MLAPS, RJEMS and YJMJ-P provided critical insights and reviewed the protocol and manuscript, making important intellectual contributions. All authors read and approved the final version.
Funding This work was supported by ZonMw (the Netherlands Organisation for Health Research and Development; grant number 630000001), funded by HandicapNL (former Revalidatiefonds; R2016006), Johanna Kinderfonds and Stichting Rotterdams Kinderrevalidatie Fonds Adriaanstichting. HandicapNL and Stichting Vooruit (18–05/YvH/NS) provided additional funding.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information. Additional data (the completed data extraction form) are available upon reasonable request.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.