Single-case Design Studies in Children with Cerebral Palsy: A Scoping Review

ABSTRACT

Aim: To critically evaluate single-case design (SCD) studies performed within the population of children/adolescents with cerebral palsy (CP).

Methods: A scoping review of SCD studies of children/adolescents with CP. Demographic, methodological, and statistical data were extracted. Articles were evaluated using the Risk of Bias in N-of-1 Trials (RoBiNT) Scale and the Consolidated Standards of Reporting Trials (CONSORT) extension for N-of-1 trials (CENT 2015). Comments regarding strengths and limitations were analyzed.

Results: Studies investigated the effects of a wide range of interventions on various outcomes. Most SCD types were adopted in multiple studies. All studies used visual inspection rather than visual analysis, often complemented with basic statistical descriptives. Risk of bias was high, particularly concerning internal validity. Many CENT items were insufficiently reported. Several benefits and limitations of SCD were identified.

Conclusions: The quality of evidence from results of SCD studies needs to be increased through risk of bias reduction.

KEYWORDS:

• Single-case design
• risk of bias
• quality of reporting
• cerebral palsy
• review

Introduction

Cerebral palsy (CP) is a life-long disorder that results from a lesion occurring in the developing brain and encompasses a large group of childhood movement and posture disorders. The rather stabile prevalence is 2–3.5 per 1000 live births over countries.¹ The disorder varies largely between individuals: in timing of the lesion, the clinical presentation, the site and severity of impairments, and activity and participation limitations. Consequently, rehabilitation of children and adolescents with CP requires an individualized approach.

Knowledge translation from research to clinical practice and vice versa is essential. This is a challenge in children with CP, since research has to deal with small and heterogeneous populations. In contrast, research within this population is dominated by group-based designs, and the randomized controlled trial (RCT) has for a long time been considered superior in intervention research.^2,3 Because children may respond differently or even oppositely to the same intervention, application of group RCT designs leads to several problems in conducting and interpreting, such as very large confidence intervals for mean changes. In addition, RCTs generally include a narrow selection of the wide diversity in characteristics of children with CP and often end up with rather small study samples, limiting the generalizability of results to an even larger extent. Because the variability in treatment effects is masked, the applicability of RCT findings to clinical decision-making in individual children is very limited.³ The single-case design (SCD) provides an alternative to the traditional group-based research to establish intervention effectiveness in this diverse and relatively small population.^4,5

The SCD is a within-subject design, performed in one or a small number of cases. The participants serve as their own controls by comparing distinct experimental phases for the outcome of interest. This ordinarily includes a baseline phase to determine a representative stable state or trend, with which subsequent phases can be compared.^5,6 The independent variable is deliberately manipulated by the researcher.^7–9 The main outcomes are to be measured systematically and repeatedly within and across all phases, in order to assess individual change across different conditions.³ There is a variety of designs adopting single-case methodology: the A-B design, the withdrawal/reversal (e.g., A-B-A) design, the multiple baseline design, the alternating treatments design, and the changing-criterion design. The design characteristics have been described in detail elsewhere.^3,10 There is controversy over the types of single-case designs that are considered experimental. In the current study, the most prevalent view is adopted: the basic A-B design does not meet the prerequisites to establish a causal relationship between the independent and the dependent variable, and is therefore not a single-case experimental design (SCED). The other previously mentioned designs provide sufficient opportunity to demonstrate an experimental effect and are, therefore, regarded SCEDs.¹⁰ Other perspectives on this matter should also be acknowledged. For instance, the What Works Clearinghouse SCD standards specifies that a design should allow for at least three demonstrations of an effect at different points in time in order to assess experimental control and, thus, considers the A-B-A design non-experimental.¹¹

The specific design used and the study’s methodology determine the confidence in any causal relationship found.⁵ Visual inspection/analysis is a common methodology used for evaluating effectiveness of intervention effects within single-case designs,^11,12 and can be complemented with statistical approaches.⁵ Various means can be used to increase the study’s validity, such as multiple baselines, randomization, and replication within or between participants.^3,5 A response-guided approach is also a commonly accepted option.¹³ Some specific threats to internal validity make the SCD methodologically challenging. These include the issue of autocorrelation due to repeated measures within cases and the possibility of carry-over effects from interventions.^6,10 In developing children return to baseline is not necessarily expected and favorable carry-over effects are desirable: retention of improvements after the treatment has ended is what is aimed for. Equally important, it would be unethical to withdraw an effective intervention in a subsequent phase, especially from a developing child. Likewise, lengthy baseline periods can be undesirable, and provoke resistance from families, who advocate the best care at all times.³ These matters are to be carefully considered regarding SCD studies in children with CP.

SCD studies can provide a rigorous alternative to RCTs, respecting the heterogeneity regarding clinical performance and prognosis that characterize the population of children with CP.³ Replication of results among cases increases generalization of findings, whereas dissimilar results enable identification of factors that interfere with the children’s response to the intervention. Hence, SCD studies inform treatment providers whether an intervention is expected to benefit their individual patient, facilitating translation of research findings into clinical practice.³ Similarly, clinicians can employ a pragmatic SCD as a clinical decision tool to identify the best option for care for their patient.

The level of evidence from SCD studies is a subject of debate. Recently, there has been a shift towards acceptance of the SCD as a methodology equivalent to the RCT.¹⁴ Some even consider the level of evidence from SCD studies to be the highest for clinical decision-making.^3,14

Although SCD seems to be underused³, there has been no reviews of the use of SCDs within CP-related research, the methodological and statistical approaches applied, or the validity and the quality of reporting of these studies. A better understanding of the (dis)advantages of specific methodological aspects would contribute to epidemiological advancements in terms of design and analysis of SCD studies. This scoping review, therefore, aims to critically evaluate SCD studies performed within the population of children and adolescents with CP. Scoping reviews differ from systematic reviews because authors do not typically summarize and analyze the level of evidence of an intervention of interest within the included studies. Scoping studies also differ from narrative reviews in that the scoping process requires analytical reinterpretation of literature. Their aim could be to examine the extend, range, and nature of a certain research activity, like single-case methodology. That creates for us the possibility to incorporate the whole range of SCDs and to address questions beyond those related to intervention effectiveness.¹⁵ The research questions that guided this scoping review were:

1. How many SCD studies have been performed in the population of children and adolescents with CP and what were their aims?

2. What methodological and statistical approaches were used in the available SCD studies?

3. What are the validity of these SCD studies and the quality of their reporting?

4. What are strengths and limitations of these SCD studies according to their authors?

Materials and Methods

The current study is explorative in nature, intending to systematically map the available literature relevant to the topic of interest. Because of the broadness and complexity of this aim, a scoping review was deemed most appropriate.

The methodological framework of Arksey and O’Malley,¹⁶ and enhancements to this approach suggested by Levac et al.,¹⁵ as well as the Joanna Briggs Institute Reviewers’ Manual 2015 (Methodology for JBI Scoping Reviews),¹⁷ were adopted for designing this scoping review. Its objectives, inclusion criteria, and methods were specified in advance in a protocol.

Identification of Relevant Studies

Inclusion Criteria

Types of Participants

Studies with children and/or adolescents aged up to 21 years with CP were included. Studies of which >50% of the participants did not meet these criteria were excluded. At the outset of the study, the population of interest also consisted of children and adolescents with acquired brain injury. However, based on the more numerous results of the initial search than anticipated and to keep the study feasible, a protocol amendment was made to restrict the scoping review to CP only.

Concept

Given the scoping review design, it is appropriate to be as inclusive as possible in order to get a true overview of the application of SCDs in the research field. Although the A-B design does not comply with the criteria for SCED, it was expected that it is frequently used. Moreover, adopting methodological quality as an in-/exclusion criterion, thereby eliminating studies with low validity, is likely to positively bias the results in terms of the risk of bias assessment. Hence, studies reporting on any SCD were included.

SCD studies were defined as having the following characteristics, in accordance with the single-case designs technical documentation from What Works Clearinghouse¹⁸:

• “An individual ‘case’ is the unit of intervention and unit of data analysis. A case may be a single participant or a cluster of participants.”

• “Within the design, the case provides its own control for purposes of comparison.”

• “The outcome variable is measured repeatedly within and across different conditions or levels of the independent variable.”

An element of experimental manipulation by the researcher(s) was required to meet the definition of SCD.

Context

The focus was on pragmatic studies, carried out in clinical practice to determine the best option for care for the individual patient. Consequently, cross-sectional laboratory studies, in which different conditions were compared during the same measurement session, were not considered interventional and, therefore, excluded. No other limits were used for the setting, type of interventions, or outcomes.

Types of Sources

Completed or ongoing empirical research studies or protocols were included. Studies and dissertations published in English, Dutch, or German were considered for inclusion.

Search Strategy

The iterative process started with an initial limited search in the electronic databases EMBASE (Ovid interface) and MEDLINE (Ovid interface). The query included terms related to the types of participants (children with CP) and the concept (SCD) of interest. The text words of titles and abstracts along with the index words of retrieved reports were analyzed. Thereafter, the identified keywords and index words were used to perform a second amended search in the electronic databases CINAHL (EBSCO interface), EMBASE (Ovid interface), MEDLINE (Ovid interface) and PsycINFO (EBSCO interface), without restriction on the year of publication. The complete search query for all databases is presented in Appendix 1. The first author developed the search strategies, in consultation with a research librarian, and performed the searches.

In order to identify grey literature, the members of the European Academy of Childhood Disability (EACD) and the Australasian Academy for Cerebral Palsy and Developmental Medicine (AusACPDM) were requested by a newsletter to inform the project group of any published or unpublished SCD studies, executed either by their research group or associates.

Study Selection

The inclusion criteria were applied during two selection phases: (1) title and abstract screening, where in case of doubt the study remained included, and (2) assessing full texts for eligibility. The first and second author independently performed both phases of the selection process. One German article was evaluated by the first and last author instead. All cases of discrepancy were discussed until consensus was reached. An inter-rater agreement and reliability analysis using percentage of agreement and the Cohen’s kappa statistic, respectively, were performed to determine consistency between raters in assessing eligibility of the full-text publications. The online Covidence systematic review software (Veritas Health Innovation, Melbourne, Australia) was used for management of the study.

Data Extraction

Study Characteristics, and Methodological and Statistical Approaches

Study characteristics as well as methodological and statistical approaches were extracted. Two independent reviewers (first and second author) carried out the data collection by using a customized digital charting form. Any differences were resolved via consensus. A third reviewer (last author) mediated if required.

Risk of Bias and Quality of Reporting

The selected articles were graded for risk of bias using the Risk of Bias in N-of-1 Trials (RoBiNT) Scale. The RoBiNT Scale has been developed to critically appraise SCD studies, primarily in the field of neurorehabilitation. It contains a 7-item Internal Validity subscale and an 8-item External Validity and Interpretation subscale. The scales cover items such as ‘blinding of people involved in the intervention’ and ‘generalization’. For each item, 0, 1, or 2 points can be awarded, which enables a total score from 0 to 30 to be calculated. The higher the score, the lower the risk of bias. Inter-rater reliability and construct validity of the RoBiNT Scale are satisfactory and an extensive manual is available, providing scoring procedures and rating guidelines.^10,19

The Consolidated Standards of Reporting Trials (CONSORT) extension for N-of-1 trials (CENT 2015) was used to evaluate the quality of reporting of the full-text publications. This standard provides guidance to facilitate preparation and appraisal of, specifically, multiple crossover N-of-1 trials.²⁰ This is in contrast to the RoBiNT Scale, which has a wider application. The CENT comprises the topics ‘title and abstract’, ‘introduction’, ‘methods’, ‘results’, ‘discussion’, and ‘other information’, and is reported in a checklist as well as an ‘explanation and elaboration’ document.²⁰ In this review, all 25 items and 44 sub-items of the CENT were rated as sufficiently reported, insufficiently reported, or not applicable.

To increase consistency between the reviewers, the first, second, and last author independently assessed five studies. They reflected on the results and aligned their conceptualizations of the appraisal tools. Thereafter, the first and second author appraised the risk of bias and reporting quality for all studies independently. Disagreements between the reviewers’ assessments were debated and a final decision agreed upon. To determine inter-rater agreement of the risk of bias appraisal, percentage of agreement between the assessors was calculated per item of the RoBiNT Scale, as well as for the Internal Validity Subscale, External Validity and Interpretation Subscale, and total score.

Content Analysis

The first author searched the introduction, methods, and discussion sections of all included studies for comments regarding the strengths and limitations of studies, including rationales for using SCD approaches. Relevant sentences or paragraphs were tracked and coded per study. These coded extracts were analyzed for recurrent issues across studies and overarching themes were sought.

Consultation

A consultation element was embedded in the scoping review to enhance the validity of the study results. The stakeholder consulted (PO) was a professor of educational and behavioral statistics and methodology, with expertise in SCD research. A report describing the preliminary findings was shared with PO to inform the consultation, followed by a discussion of his perspectives. The provided insights were incorporated into this paper.

Results

Figure 1 displays the scoping review flow diagram. After removal of duplicates, the database search (last updated 13 September 2018) retrieved 417 citations. The newsletter item, which was sent to 1261 EACD and AusACPDM members, obtained no responses. Eighty published records were included in the review: 73 full-text publications (including one dissertation) and seven conference and dissertation abstracts. The records reported on 74 unique studies. One of the full-text publications reported on two separate studies and another described methods only. The inter-rater agreement of full-text assessment was found to be 86.9%. Inter-rater reliability was substantial: Cohen’s kappa = 0.72 (p < .0.001).

Figure 1. Flowchart of the study selection process.

Study Characteristics

Appendix 2 contains the aims and characteristics of the included studies. The studies focused on a wide range of interventions and outcomes. Interventions of interest in multiple studies were interactive video gaming or virtual reality (eight studies), orthotics (seven studies), seating and positioning (five studies), alternative and augmentative communication (four studies), neurodevelopmental therapy (three studies), Constrained Induced Movement Therapy (CIMT) (two studies), early intervention (two studies), electrical stimulation (two studies), and oral motor treatment (two studies). Besides establishing effects, a few studies aimed to answer feasibility questions, for instance concerning satisfaction of users with the treatment(s). The sample sizes varied from one to 14 children and adolescents. The demographics of participants were diverse with respect to types of CP, age, and gender.

Methodological and Statistical Approaches

As can be seen from Table 1, the withdrawal/reversal design was most frequently used (39.2%), followed by the A-B design (29.7%), the multiple baseline design (18.9%), and the alternating treatments design (9.5%). The changing-criterion design was seldom used (2.7%). The data analysis of all studies comprised visual inspection to a greater or lesser degree. Nevertheless, none of the studies systematically evaluated all six components (i.e., level, trend, variability, immediacy of the effect, overlap, and consistency of data patterns across similar phases) that are considered fundamental to demonstrate whether a causal relation exists and which would have been regarded as a visual analysis.^11,12,18 The descriptive two-standard deviation band method and the split-middle technique (celeration line) were the most frequently mentioned statistical approaches, used in 25.3% and 17.6% of the studies, respectively.

Table 1. Methodological and statistical approaches of studies.

Study	Design type	Statistical approach(es)
Akerstedt et al.^Citation21 (2010)	A-B design
Angelo^Citation22 (1992)	Alternating-treatments design	Kruskal-Wallis test
Angelo^Citation23 (1993)	Alternating-treatments design
Barbosa et al.^Citation24 (2008)	Withdrawal/reversal (A-B-A) design: A-B-C-A	Celeration Line (CL) Two-standard deviation band method
Beauregard et al.^Citation25 (1998)	Withdrawal/reversal (A-B-A) design	Celeration line
Brien et al.^Citation26 (2010)	Multiple-baseline design	Visual analysis of graphed data Statistical analysis using the two-standard deviation band method Analysis of clinical significance
Brien et al.^Citation27 (2011)	A-B design (including follow-up)	Autocorrelation coefficient to determine serial dependency Visual inspection: summarizing patterns of data on the basis of trend, level, and variability Two-standard deviation band method Computation of the differences between the mean scores at baseline and the mean scores at intervention and follow-up
Chen et al.^Citation28 (2007)	A-B design (including follow-up)	Autocorrelation coefficient to determine serial dependency Two-standard deviation band method Analysis of clinical significance
Coker et al.^Citation29 (2009)	Withdrawal/reversal (A-B-A) design: A-B-A-B
Corn et al.^Citation30 (2003)	A-B design	Linear regression models to determine the line of best fit T-values (difference between the slope of the baseline data and the slope of the intervention and intercept between the last baseline value and the first intervention value (effect))
Costigan et al.^Citation31 (2010)	Withdrawal/reversal (A-B-A) design: A-B-A-B	Changes in mean accurate selections between phases Changes in the level of the dependent variable between phases (i.e., difference in performance between the end of one phase and the beginning of the next phase) The trend of the data (i.e., the direction of the data for the dependent variable in each phase) The strength of the trend (i.e., the slope of the line connecting data points) Percentage of nonoverlapping data (i.e., the percentage of data points in the intervention phase that were greater than the highest data point in the baseline phase)
Crocker et al.^Citation32 (1997)	Withdrawal/reversal (A-B-A) design	Two-standard deviation band method Inter-rater reliability: Pearson correlation coefficients
Dada et al.^Citation33 (2009)	Multiple-baseline design
DeGangi et al.^Citation34 (1983)	Alternating-treatments design	Sign test
Dinomais et al.^Citation35 (2013)	A-B design	‘‘Nonoverlap of all pairs’’ (NAP) method
Do et al.^Citation36 (2016)	Withdrawal/reversal (A-B-A) design	Change in means between all phases.
Durfee et al.^Citation37 (1999)	Withdrawal/reversal (A-B-A) design: A-B-A-B	Least squares regression procedures were used to establish straight-line trends for each treatment phase
Fienup et al.^Citation38 (2013)	Two-phase study: multiple baseline design + changing criterion design
Fragala et al.^Citation39 (2002)	A-B design	Trend line
Ghorbani et al.^Citation40 (2017)	Multiple-baseline design	Percentage of Non-overlapping Data (PND) Split-Middle Method of trend estimation (celeration line)
Goodman et al.^Citation41 (1991)	A-B design	Autocorrelation coefficient to determine serial dependency
Hamill et al.^Citation42 (2007)	Multiple-baseline design
Harris et al.^Citation43 (1986)	Alternating-treatments design
Hartveld et al.^Citation44 (1996)	A-B design
Havstam et al.^Citation45 (2003)	Withdrawal/reversal (A-B-A) design
Hinderer et al.^Citation46 (1988)	A-B design: A-B-A-C design	Linear regression line
Iammatteo et al.^Citation47 (1990)	Withdrawal/reversal (A-B-A) design	Celeration line approach Statistical significance was determined through the use of Bloom’s probability table Autocorrelation coefficient to determine serial dependency
Jelsma et al.^Citation48 (2013)	Multiple-baseline design	Non-overlap of All Pairs (NAP) Repeated measures ANOVA Tukey post-hoc analysis
Johnson et al.^Citation49 (2014)	Multiple-baseline design	Split-middle lines of trend estimation
Kelly et al.^Citation50 (2009)	A-B design (including follow-up)	Two-standard deviation band method
Kenyon et al.^Citation51 (2018)	Withdrawal/reversal (A-B-A) design: A-B-A-B	Two-standard deviation band method
Kinghorn et al.^Citation52 (1996)	Withdrawal/reversal (A-B-A) design	Two-standard deviation band method C statistic
Ko et al.^Citation53 (2015)	A-B design	Two-standard deviation band method Paired t-test
Laessker-Alkema et al.^Citation54 (2016)	A-B design	Two-standard deviation band method Group level: Wilcoxon singed rank test
Lammi et al.^Citation55 (2003)	Multiple-baseline design	Serial dependency between the data points in the baseline using Bartlett’s test Celeration lines Linear regression using the c-statistic Comparison of the mean performance ratings for each phase of the study to the overall mean
Laskas et al.^Citation56 (1985)	Withdrawal/reversal (A-B-A) design	Split-middle technique to calculate the celeration line
Lephart et al.^Citation57 (2015)	Withdrawal/reversal (A-B-A) design	Dependent t tests (to determine whether physiological measures changed within treatment sessions) Split middle analyses (celeration line)
Lewis et al.^Citation58 (2016)	Multiple-baseline design
Lilly et al.^Citation59 (1990)	Alternating-treatments design	Friedman analysis of variance by ranks (modified Friedman test) Analysis of variance (ANOVA) Wilcoxon rank sum test
Lin et al.^Citation60 (2014)	Withdrawal/reversal (A-B-A) design: A-B-A-B	Kolmogorov–Smirnov statistical test
Lin et al.^Citation61 (2015)	Withdrawal/reversal (A-B-A) design: A-B-A-B	Kolmogorov–Smirnov statistical test
Mackey^Citation62 (1989)	Withdrawal/reversal (A-B-A) design	One-way analysis of variance test T-test for group effects
Man et al.^Citation63 (2007)	A-B design: A-B-C-D	Levene’s test
Matthews et al.^Citation64 (2009)	Withdrawal/reversal (A-B-A) design	Inferential statistical analysis: celeration line method
McCarthy et al.^Citation65 (2015)	A-B design
McConnell et al.^Citation66 (2014)	Withdrawal/reversal (A-B-A) design	Two-standard deviation band method
Owen et al.^Citation67 (1992)	Four-phased randomized design	Mann-Whitney U Test
Pinder et al.^Citation68 (1995)	Withdrawal/reversal (A-B-A) design
Pool et al.^Citation69,Citation70 (2014)	Withdrawal/reversal (A-B-A) design	Two-standard deviation band method C-statistic Autocorrelation coefficient to determine serial dependency Group-level: Wilcoxon signed rank tests
Pratt et al.^Citation71 (2016)	Alternating-treatments design	Autocorrelation coefficient to determine serial dependency Percentage of nonoverlapping data (PND) between conditions Celeration line Probability table Effect size
Prosser et al.^Citation72,Citation73 (2012)	A-B design (including follow-up)	Split middle technique: celeration lines Net change scores (subtracting baseline change from treatment change)
Ramani et al.^Citation74 (2014)	Multiple-baseline design
Ray et al.^Citation75 (1983)	Withdrawal/reversal (A-B-A) design	Split-middle technique to calculate celeration lines Autocorrelation coefficient to determine serial dependency
Reid et al.^Citation76 (1999)	Withdrawal/reversal (A-B-A) design	Group level: Repeated measures ANOVA
Retarekar et al.^Citation77 (2009)	Withdrawal/reversal (A-B-A) design	Two-standard deviation band method
Rigby et al.^Citation78 (2001)	Withdrawal/reversal (A-B-A) design
Rintala et al.^Citation79 (1994)	Withdrawal/reversal (A-B-A) design
Rivi et al.^Citation80 (2014)	A-B design
Sakemiller et al.^Citation81 (1998)	Withdrawal/reversal (A-B-A) design
Sheppard et al.^Citation82 (2007)	A-B design: A-B-A-C design	Interrupted time-series analysis (ITSA)
Shumway-Cook et al.^Citation83 (2003)	Multiple-baseline design	Hierarchical linear modelling Repeated measures ANOVAs
Siconolfi-Morris^Citation84 (2012)	A-B design	Two-standard deviation band method C-statistic (converted to a Z score) Improvement rate difference (IRD)
Smelt^Citation85 (1989)	Withdrawal/reversal (A-B-A) design	Auto-correlation coefficient (serial dependency) Standard deviation band (SD-band) C-statistic
Smithers^Citation86 (1991)	Withdrawal/reversal (A-B-A) design	Method of least squares (trend line)
Soto et al.^Citation87 (2008)	A-B design
Stansfield et al.^Citation88 (2015)	A-B design	Two-standard deviation band method
Stewart et al.^Citation89 (1992)	Alternating-treatments design	Two-tailed paired sample t-test for each subject and across all subjects
Stewart et al.^Citation90 (2000)	Withdrawal/reversal (A-B-A) design: A-B-A-C-A
Svedberg et al.^Citation91 (2001)	A-B design: A-B-B	Student’s t-test Linear regression
Symons et al.^Citation92 (2007)	A-B design: A-B-C-D
Thorpe et al.^Citation93 (2002)	Multiple-baseline design	“Sign” test on group level Split-middle method/trend line
Ustad et al.^Citation94 (2009)	Multiple-baseline design: A-B-A-B-A	Two-standard deviation band method Celeration lines
Voulgarakis et al.^Citation95 (2015)	Changing-criterion design
Ward et al.^Citation96,Citation97 (2013 & 2014)	Multiple-baseline design: A-B-C-A	Two-standard deviation band method Dual-criterion method of trend estimation with binomial test Modified Cohen’s d, based on the pooled standard deviation

The letters ‘A’, ‘B’, etc., designate the treatment phases.

Risk of Bias

Table 2 presents the results obtained from the risk of bias appraisal. The included studies scored on average 11.2 points on the RoBiNT Scale. Total scores ranged from 4 to 22 points. The Interval Validity Subscale mean score was 3.6 points. The mean score on the External Validity and Interpretation Subscale was 7.6 points, more than twice as high.

Table 2. Risk of bias per item of the RoBiNT Scale, and inter-rater agreement of RoBiNT Scale scores (n = 73).

	Mean (range)						% of agreement
Interval Validity Subscale	3.6 (1–10)						79
External Validity and Interpretation Subscale	7.6 (2–13)						72
Total score	11.2 (4–22)						75
	Score
Item	0 points		1 point		2 points
	Absolute	%	Absolute	%	Absolute	%
Interval Validity Subscale
1. Design with control	27	37.0	29	39.7	17	23.3	74
2. Randomization	63	86.3	0	0.0	10	13.7	93
3. Sampling of behavior	8	11.0	30	41.1	35	47.9	78
4. Blinding of people involved in the intervention	70	95.9	3	4.1	0	0.0	93
5. Blinding of assessor(s)	56	76.7	3	4.1	14	19.2	77
6. Inter-rater agreement	47	64.4	15	20.5	11	15.1	53
7. Treatment adherence	67	91.8	3	4.1	3	4.1	84
External Validity and Interpretation Subscale
8. Baseline characteristics	37	50.7	36	49.3	0	0.0	58
9. Setting	27	37.0	29	39.7	17	23.3	59
10. Dependent variable (target behavior)	0	0.0	13	17.8	60	82.2	78
11. Independent behavior (therapy/intervention)	5	6.8	23	31.5	45	61.6	67
12. Raw data record	12	16.4	43	58.9	18	24.7	68
13. Data analysis	34	46.6	24	32.9	15	20.5	59
14. Replication	25	34.2	20	27.4	28	38.4	89
15. Generalization	71	97.3	0	0.0	2	2.7	99

Internal Validity Subscale

For the item ‘design with control’, 37.0% of the articles scored 0 points, 39.7% scored 1 point and 23.3% scored 2 points. Many multiple baseline designs were rewarded 0 points since concurrency was not demonstrated. In as many as 86.3% of the studies, no randomization was applied. With regard to item 3, it was found that 41.1% and 47.9% of the studies used at least 3 or at least 5 data points or sets of alternating sequences, respectively. Blinding of either the participant or the practitioner involved in the intervention was not applied in 95.9% of the studies. The majority of studies (80.8%) also did not use a blinded assessor. As to inter-rater agreement, 0 points, 1 point, and 2 points were given to 64.4%, 20.5%, and 15.1% of the studies, respectively. Finally, it appeared that 91.8% of the studies did not (adequately) confirm treatment adherence.

External Validity and Interpretation Subscale

Although half of the studies did sufficiently report on relevant biomedical variables, none provided an analysis of the baseline conditions or behaviors using functional status, case formulation, or other systematic investigation. Only 23.3% of the studies described both the general location and the specific environment in which the intervention was conducted. A large proportion of the studies (82.2%) sufficiently defined the dependent variable (outcome), including the measuring method. Similarly, 61.6% of the studies described the content of the intervention and the procedure of delivery in detail. For the item ‘raw data record’ a score of 1 point was awarded to the largest group of studies. Many studies were penalized because the figures were of such poor quality that the data points could not be read. The data analysis of studies was rated 0 points in 46.6%, 1 point in 32.9%, and 2 points in 20.5% of the studies. Many studies did conduct statistical analyses, but justification was missing, which is a requirement for a maximum score of 2 points. A total of 34.2% of studies did not contain any replication. The study was replicated (across participants, settings, behaviors, or practitioners) once or twice in 27.4% of the studies, and three or more times in 38.4% of the studies. Just two studies indicated a generalization measure to determine transfer of intervention effects to other variables or settings.

Inter-rater Agreement

The inter-rater agreement between the two assessors per item of the RoBiNT Scale ranged from 53% to 99% (Table 2). The agreement for the Internal Validity Subscale, External Validity and Interpretation Subscale, and total score were 79%, 72%, and 75%, respectively.

Quality of Reporting

Table 3 compares the proportion of studies that sufficiently and insufficiently reported the items of the CENT.

Table 3. Quality of reporting appraisal, per item of the CENT (n = 72).

			Sufficient reporting		Insufficient reporting		Not applicable
	No.	Item	Absolute	%	Absolute	%	Absolute	%
Title and abstract
	1a	Identify as an “N-of-1 trial” in the title	17	23.6	55	76.4	0	0.0
	1b	Structured summary of trial design, methods, results, and conclusions (for specific guidance, see CENT guidance for abstracts)	0	0.0	71	98.6	1	1.4
Introduction
Background and objectives	2a.1	Scientific background and explanation of rationale	68	94.4	4	5.6	0	0.0
	2a.2	Rationale for using N-of-1 approach	21	29.2	51	70.8	0	0.0
	2b	Specific objectives or hypotheses	54	75.0	18	25.0	0	0.0
Methods
Trial design	3a	Describe trial design, planned number of periods, and duration of each period (including run-in and wash out, if applicable) In addition for series: Whether and how the design was individualized to each participant, and explain the series design	43	59.7	29	40.3	0	0.0
	3b	Important changes to methods after trial start (such as eligibility criteria), with reasons	19	26.4	53	73.6	0	0.0
Participant(s)	4a	Diagnosis or disorder, diagnostic criteria, comorbid conditions, and concurrent therapies. For series: Eligibility criteria for participants	64	88.9	8	11.1	0	0.0
	4b	Settings and locations where the data were collected	19	26.4	53	73.6	0	0.0
	4c	Whether the trial(s) represents a research study and if so, whether institutional ethics approval was obtained	27	37.5	45	62.5	0	0.0
Interventions	5	The interventions for each period with sufficient details to allow replication, including how and when they were actually administered	38	52.8	34	47.2	0	0.0
Outcomes	6a.1	Completely defined pre-specified primary and secondary outcome measures, including how and when they were assessed	55	76.4	17	23.6	0	0.0
	6a.2	Description and measurement properties (validity and reliability) of outcome assessment tools	10	13.9	62	86.1	0	0.0
	6b	Any changes to trial outcomes after the trial commenced, with reasons	11	15.3	61	84.7	0	0.0
Sample size	7a	How sample size was determined	0	0.0	72	100	0	0.0
	7b	When applicable, explanation of any interim analyses and stopping guidelines	0	0.0	72	100	0	0.0
Randomization
Sequence generation	8a	Whether the order of treatment periods was randomized, with rationale, and method used to generate allocation sequence	4	5.6	68	94.4	0	0.0
	8b	When applicable, type of randomization; details of any restrictions (such as pairs, blocking)	0	0.0	10	13.9	62	86.1
	8c	Full, intended sequence of periods	60	83.3	12	16.7	0	0.0
Allocation concealment mechanism	9	Mechanism used to implement the random allocation sequence (such as sequentially numbered containers), describing any steps taken to conceal the sequence until interventions were assigned	0	0.0	9	12.5	63	87.5
Implementation	10	Who generated the random allocation sequence, who enrolled participants, and who assigned participants to interventions	2	2.8	9	12.5	61	84.7
Blinding	11a	If done, who was blinded after assignment to interventions (for example, participants, care providers, those assessing outcomes) and how	13	18.1	59	81.9	0	0.0
	11b	If relevant, description of the similarity of interventions	2	2.8	3	4.2	67	93.1
Statistical methods	12a	Methods used to summarize data and compare interventions for primary and secondary outcomes	52	72.2	20	27.8	0	0.0
	12b	For series: If done, methods of quantitative synthesis of individual trial data, including subgroup analyses, adjusted analyses, and how heterogeneity between participants was assessed, (for specific guidance on reporting syntheses of multiple trials, please consult the PRISMA Statement)	8	11.1	46	63.9	18	25.0
	12c	Statistical methods used to account for carryover effect, period effects, and intra-subject correlation	9	12.5	63	87.5	0	0.0
Results
Participant flow	13a.1	Number and sequence of periods completed, and any changes from original plan with reasons	53	73.6	19	26.4	0	0.0
	13a.2	For series: The number of participants who were enrolled, assigned to interventions, and analysed for the primary outcome	50	69.4	4	5.6	18	25.0
	13b	For series: losses or exclusions of participants after treatment assignment, with reasons, and period in which this occurred, if applicable	43	59.7	10	13.9	19	26.4
Recruitment	14a	Dates defining the periods of recruitment and follow-up	4	5.6	68	94.4	0	0.0
	14b	Whether any periods were stopped early and/or whether trial was stopped early, with reason(s).	10	13.9	62	86.1	0	0.0
Baseline data	15	A table showing baseline demographic and clinical characteristics	28	38.9	44	61.1	0	0.0
Numbers analysed	16	For each intervention, number of periods analysed. In addition for series: if quantitative synthesis was performed, number of trials for which data were synthesized	64	88.9	8	11.1	0	0.0
Outcomes and estimation	17a.1	For each primary and secondary outcome, results for each period; an accompanying figure displaying the trial data is recommended.	59	81.9	13	18.1	0	0.0
	17a.2	For each primary and secondary outcome, the estimated effect size and its precision (such as 95% confidence interval) In addition for series: if quantitative synthesis was performed, group estimates of effect and precision for each primary and secondary outcome	3	4.2	67	93.1	2	2.8
	17b	For binary outcomes, presentation of both absolute and relative effect sizes is recommended	1	1.4	0	0.0	71	98.6
Ancillary analyses	18	Results of any other analyses performed, including assessment of carryover effects, period effects, intra-subject correlation In addition for series: If done, results of subgroup or sensitivity analyses	14	19.4	3	4.2	55	76.4
Harms	19	All harms or unintended effects for each intervention	11	15.3	61	84.7	0	0.0
Discussion
Limitations	20	Trial limitations, addressing sources of potential bias, imprecision, and, if relevant, multiplicity of analyses	40	55.6	32	44.4	0	0.0
Generalizability	21	Generalizability (external validity, applicability) of the trial findings	42	58.3	30	41.7	0	0.0
Interpretation	22	Interpretation consistent with results, balancing benefits and harms, and considering other relevant evidence	45	62.5	27	37.5	0	0.0
Other information
Registration	23	Registration number and name of trial registry	0	0.0	72	100	0	0.0
Protocol	24	Where the full trial protocol can be accessed, if available	0	0.0	72	100	0	0.0
Funding	25	Sources of funding and other support (such as supply of drugs), role of funders	35	48.6	37	51.4	0	0.0

Title and Abstract

Merely a quarter of the publications identified their study as a SCD in the title. According to the specific CENT guidance, the abstract should describe such items as the trial design, participant(s), interventions, objective, outcome, randomization, blinding, results, and conclusion. None of the included studies fully met these criteria. Particularly, randomization and blinding procedures, and harms were often missing.

Introduction

The majority of the studies reported the scientific background, explained the rationale, and specified the objectives or hypotheses. However, only 29.2% of the authors reported their rationale for using a SCD approach.

Methods

Although 59.7% of the studies sufficiently described the trial design, only 26.4% described whether and how methods changed after the start of the study. Patient characteristics or eligibility criteria were well reported. Still, in most studies information on settings and locations was lacking, and it was not addressed whether both a health research ethics board reviewed and approved the research study and parent and/or patient consent or assent was obtained. About half of the studies reported the interventions for each period insufficiently. Baseline intervention procedures were frequently not reported. While 76.4% of the studies completely defined pre-specified primary and secondary outcomes, the minority adequately described the measurement properties of outcome assessment tools and changes to outcomes after the trial commenced. None of the studies defined how sample size was determined nor explained interim analysis and stopping guidelines. Only four studies reported both whether and how randomization was applied. Consequently, item numbers 8b, 9, and 10 were often not applicable. Studies that did use randomization generally did not describe the type of randomization, allocation concealment mechanism, and implementation. Blinding was not mentioned in 81.9% of the studies. A total of 72.2% of the studies satisfactorily reported the statistical methods that were used to summarize data and compare interventions. On the other hand, methods of quantitative synthesis of individual trial data and assessing heterogeneity between participants (for series), and methods to account for carryover effect, period effects, and intra-subject correlation were reported in only 11.1% and 12.5% of the studies, respectively.

Results

Overall, the participant flow was well reported. Dates of the recruitment and follow-up period were not defined, with the exception of four studies. Further, just 13.9% of the studies reported whether and why any periods or the study itself were stopped early. Baseline demographic and clinical characteristics were too limited or not provided at all in 61.1% of the studies. Most studies sufficiently reported the numbers analyzed. The assessment of outcomes and estimation revealed opposing results: 81.9% of studies presented results for each outcome and each period, while no more than three studies estimated an effect size and its precision. The items concerning binary outcomes and ancillary analysis were mainly irrelevant. Few studies (15.3%) reported any harms or unintended effects of interventions.

Discussion

Limitations, generalizability, and interpretation of the results were discussed satisfactorily in comparable proportions of studies: 55.6%, 58.3%, and 62.5%, respectively.

Other Information

A registration or protocol of the study was not referred to in any of the publications. Sources of funding were named in about half of the studies.

Development over Time

From 1983 to 2018, the number of published SCD studies fluctuated between 0 and 7 per year, with an increase over time (Figure 2). The mean RoBiNT Scale total score of studies published each year remained stable over the last decades (Figure 2).

Figure 2. Development of SCD studies over time (including linear least squares trend lines).

Content Analysis

The analysis of the authors’ commentaries on the utilization of SCDs resulted in a synthesis of the findings in two themes: applicability of the SCD and methodological considerations.

Applicability of the SCD

The most mentioned argument for applying a SCD was the appropriateness of this design for the heterogeneous population of children with CP and related variability in responses to interventions. Authors specified that the SCD approach met the demands of individualized (evaluation of) treatments, thereby being highly relevant for clinical practice. A secondary reason for using a SCD was limited availability of potential participants. Finally, some expressed the SCD to be a rigorous design with a high level of internal validity.

Authors frequently discussed limited generalizability of their findings, either to patients comparable to the children under research or to a wider population, as a major limitation of their study. Likewise, a small sample size was mentioned as a limitation. Replication or a follow-up study in a larger and/or more heterogeneous sample and using a longer follow-up period were often recommended. Several studies explicitly planned or recommended an RCT following the SCD study. In agreement with this, many authors referred to their work as preliminary, exploratory, or a pilot study.

Methodological Considerations

Practical concerns and constrains limited the possibilities of the research design. These were often inherent to the clinical nature of the study. Additionally, in many cases illness, holidays, or other events interrupted or otherwise disturbed the study procedures. Another recurrent example was the lack of a blinded assessor, due to insufficient resources.

Concerning the study population, authors noticed that motivation, compliance, rapport with the assessor, and other mood and behavioral aspects of the child affected the assessment process. Second, confounding effects of growth, maturity and the ongoing process of motor development were recognized in the interpretation of findings. Researchers explained that they carefully considered the time and effort required of participants in order to ease the burden on children and their families, sometimes at the expense of study rigor.

The relevance as well as the challenge of standardized measurement conditions (e.g., same time each day) were acknowledged. Moreover, authors balanced the importance of outcome measures being sensitive to change with clinical relevance and ease of administration. A few specifically stated that the Gross Motor Function Measure (GMFM) was not responsive to changes in the case of short intervals between measurements.

In several publications, it was argued that withdrawal of treatment was inappropriate, either ethically or because permanent change of the child’s functioning was intended. In some other studies, where researchers did withdraw an intervention, a carry-over effect was demonstrated. Furthermore, some authors identified other limitations of their studies related to the SCD: the absence of a control group, too few data points collected, the variability of data points within phases and thus no stable baseline, and practice effects.

Discussion

With respect to the first research question of this study, 74 SCD studies have been identified. These studies aimed to investigate the effects of a range of interventions on a variety of outcomes. The second question focused on the methodological and statistical approaches of the included studies. Except for the changing-criterion design, all SCD types were adopted in multiple studies. All studies used visual inspection rather than visual analysis, often complemented with inconsistent, basic statistical approaches. The third question sought to determine the risk of bias and quality of reporting. The overall risk of bias was relatively high, particularly in terms of internal validity. Additionally, many items of the CENT were insufficiently reported in the included full-text publications. Although the annual number of studies increased, their risk of bias did not decrease over time. The fourth and last question was designed to identify the strengths and limitations of SCD studies. Benefits and limitations of the applicability of the SCD in the population of interest were determined and several methodological considerations discussed.

The A-B design and the non-concurrent multiple baseline design, which can be seen as a series of A-B designs,¹⁰ were included in this scoping review because they encompass single-case methodology. It should be emphasized, though, that generally causal inferences cannot be derived from these designs.¹⁰ A recent article argued that A-B designs in which the start point of the treatment phase is randomized can be valid experimental designs under certain conditions,⁹⁸ but the numerous studies included in this review that used this design did not comply with these criteria. The analysis of SCD is controversial in terms of the applicability of visual and (advanced) statistical analyses available. Beyond this debate, there is abundant room for improvement in the reporting of the analyses used, their rationales, and the techniques used. Due to the nature of the interventions in pediatric rehabilitation, in most studies, neither families nor professionals involved in delivery of the intervention could have been blinded. Except for corresponding Item 4 of the RoBiNT Scale, studies missed opportunities to comply with quality criteria for SCDs. Contrary to expectations, this review did not find an increase of the total score of the RoBiNT Scale over time. This may be explained by the fact that guidelines for SCD design and reporting were only published some 10 years ago.^18,99 One of the findings of the content analysis was that SCD studies in the clinical setting are facing practical restrains and limited resources, which negatively affects the capability to perform the experiment in conformity with recommended procedures. This is a possible explanation for the high risk of bias of studies. Similarly, acceptable measurement burdens for participating families may have been prioritized over measurement intensity or phase lengths.

Next to the systematic and thorough search and data charting process, the consultation phase in which an expert in SCD statistics and methodology reviewed the results strengthened the validity of this scoping review. Unfortunately, the lack of response to the newsletter items did not allow for the inclusion of grey literature. It is possible, therefore, that unpublished studies were missed. The CENT reporting standard was chosen along with the RoBiNT Scale to cover both the policies of medical and behavioral sciences. Unlike the alternative Single-Case Reporting Guideline In BEhavioural Interventions (SCRIBE), the CENT has been specifically designed for the medical N-of-1 trial, i.e., a multiple crossover design. As a consequence, not all items were equally relevant for each type of SCD. Hence, results of the quality or reporting assessment need to be interpreted with caution to avoid overestimating the inadequacy of reporting. Last, the inter-rater agreement of the risk of bias appraisal was slightly lower than the agreement of 81% between trained novice raters in an inter-rater reliability study of the RoBiNT Scale.¹⁰

We are convinced that adopting SCD methodology is the way forward for clinically meaningful intervention research in the heterogeneous population of children and adolescents with CP. Many authors of included studies confirmed this potential of the SCD. Gaps and recommendations for using this type of design are also discussed in a recent publication.³ The authors state that single-case methodology could be very useful in pediatric rehabilitation, because it describes the variability of responses within and between individuals. The importance of the research question as leading for the choice of a design type within the field of single-case methodology is underlined, as multiple designs adopting single-case methodology are applicable to CP-research. A typical research question in this field should include a clear definition of the independent and dependent variable: are changes initiated by the independent variable associated with changes in the outcome to be measured? An outcome measure should be able to pick up change over time on an individual basis. However, that still is a point of debate, as responsiveness is not yet investigated for a lot of outcome measure in the field of CP, and the approach for responsiveness on an individual basis is not yet clear. More research on this point is needed. The paper of Romeiser-Logan et al. discusses some SCD examples in an extensive way.³

Epidemiological principles and design standards are available to guide the planning of a SCD study and increase the credibility of results. General tools include the RoBiNT Scale, the SCRIBE, the CENT, and the SCD standards by What Works Clearinghouse.^{10,18–20,100} So far only one CP-specific guideline is available, which is rather concise, with limited testing for validity.¹⁰¹ Hence, standards should be advanced for the complex rehabilitation setting to fit the specific needs of pediatric rehabilitation. To initiate a trend towards improvement of validity, it is imperative that researchers follow available recommendations for SCDs. In addition, the importance of generic principles of intervention research should not be underestimated, such as clinical relevance of treatments and reporting of harms.

A number of issues is still under debate among experts and requires consensus, for instance, consistent SCD terminology and appropriate statistical approaches per design. Statistics developed for this type of studies, next to the more traditional focus on the visual techniques for analysis, will open the possibility to combine studies in a meta-analysis approach. Best practices need to be agreed upon for CP-related research, supplementing earlier work.³ Some specific concerns need particular attention. On the one hand, the non-reversal multiple baseline design seems most convenient, since many CP-treatments are intended to produce a permanent change. On the other hand, if no immediate effect is expected, it will reduce the suitability of the multiple baseline design. Hence, directions for the applicability of SCD designs to certain research situations are desired. Furthermore, the frequently repeated measures require instruments with specific psychometric properties that are at the same time clinically relevant and convenient to administer often. A core set of suitable measurement instruments could aid in the selection from the many options available in pediatric rehabilitation.

The SCD has potential in the context of personalized evidence-based medicine in children and adolescents with CP, provided that the level of evidence of results is increased by minimizing risk of bias. Therefore, it is imperative that researchers comply with recommendations and guidelines. This requires researchers and health-care givers in the field to thoroughly invest in gaining knowledge about the typical features of single-case methodology. We invite the field to fill the gaps identified by us by building the evidence base for a SCD methodology that is highly valid and relevant to clinical practice.

Declaration of interest statement

The authors report no conflicts of interest.

Appendix 1.

Search query

EMBASE database (Ovid interface)

1. Cerebral Palsy/

2. ((cerebral or brain or central) adj2 (pals$ or paraly$)).ti,ab,ot.

3. 1 or 2

4. Infant/or exp Child/or Adolescent/or Young Adult/or Pediatrics/

5. (infant$ or child$ or adolescent$ or p?ediatric$ or youth).ti,ab,ot.

6. 4 or 5

7. (“n of 1” or “single subject$” or “single case$” or “single patient$” or “within subject$”).ti,ab,ot.

8. 3 and 6 and 7

9. limit 8 to humans

10. remove duplicates from 9

MEDLINE database (Ovid interface)

#1 Cerebral Palsy[mesh:noexp]

#2 ((cerebral[tiab] OR brain[tiab] OR central[tiab]) AND (pals*[tiab] OR paraly*[tiab]))

#3 (#1 OR #2)

#4 (Infant[mesh:noexp] OR Child[mesh] OR Adolescent[mesh:noexp] OR Young Adult[mesh:noexp] OR Pediatrics[mesh:noexp])

#5 (infant*[tiab] OR child*[tiab] OR adolescent*[tiab] OR pediatric*[tiab] OR paediatric*[tiab] OR youth[tiab])

#6 (#4 OR #5)

#7 (“n-of-1*”[tiab] OR “single subject*”[tiab] OR “single case*”[tiab] OR “single patient*”[tiab] OR “within subject*”[tiab])

#8 (#3 AND #6 AND #7)

CINAHL and PsycINFO databases (EBSCO interface)

S1 MH “Cerebral Palsy”

S2 TI (cerebral OR brain OR central) N2 (pals* OR paraly*) OR AB (cerebral OR brain OR central) N2 (pals* OR paraly*)

S3 S1 OR S2

S4 MH “Infant” OR MH “Child, Preschool” OR MH “Child” OR MH “Adolescence” OR MH “Young Adult” OR MH “Pediatrics”

S5 TI (infant* OR child* OR adolescent* OR p#ediatric* OR youth) OR AB (infant* OR child* OR adolescent* OR p#ediatric* OR youth)

S6 S4 OR S5

S7 TI (“n-of-1*” OR “single subject*” OR “single case*” OR “single patient*” OR “within subject*”) OR AB (“n-of-1*” OR “single subject*” OR “single case*” OR “single patient*” OR “within subject*”)

S8 S3 AND S6 AND S7

Appendix 2.

Aims and characteristics of included studies

Table

Study	Country	Aim of the study	Study sample	Outcomes	Interventions^a
					ICF level	Treatment approach	Specified
Akerstedt et al.^Citation21 (2010)	Sweden	To evaluate single event multilevel surgery (SEMLS) and rehabilitation using clinical- and self-reported functional measures and HRQOL in youth with CP.	n = 11 Children with CP: 2 Adolescents with CP: 9 Gender: 10 male, 1 female Age: mdn 15y (range 9–18y) Type of CP: 11 spastic TD of CP: 1 hemiplegia, 10 bilateral	Energy cost: Physical Cost Index (PCI) Gross motor capacity: Gross Motor Function Measure (GMFM-66) Walking ability: Classification of Function, Disability and Health (ICF) Maximum outdoor gait distance: estimated in meters Health-related quality of life: Child Health Questionnaire (CHQ) Satisfaction of care: questions regarding fulfilment of expectations and worth of effort; questionnaire regarding estimated time for training and activities for daily life (ADL) during a week; estimated perceived physical and mental effort of the overall rehabilitation-process for the moment	BFS + A	Single event multilevel surgery with therapy	Active and passive movements, muscle strength, stretching, balance and gait towards more complex functional tasks
Angelo^Citation22 (1992)	USA	To compare the three basic modes of scanning -automatic, inverse, and step- regarding accuracy in adolescents with CP.	n = 6 Adolescents with CP: 6 Gender: NS Age: range 18–20y Type of CP: 3 spastic, 3 dyskinetic TD of CP: NS	Accuracy	A	Alternative and augmentative communication	Regular scanning
					A	Alternative and augmentative communication	Inverse scanning
					A	Alternative and augmentative communication	Step scanning
Angelo^Citation23 (1993)	USA	To determine the effect of tilt-in-space on head control for a child with cerebral palsy.	n = 1 Children with CP: 1 Gender: 1 female Age: 9y Type of CP: NS TD of CP: NS	Head control: Time sample recording (percentage of time the subject directly faced the activity)	A + P	Seating and positioning	Zero degrees
					A + P	Seating and positioning	Fifteen degrees
					A + P	Seating and positioning	Thirty degrees
Barbosa et al.^Citation24 (2008)	Brazil	To document the effects of electrical stimulation applied to the wrist muscles of hemiplegic children on hand function.	n = 2 Children with CP: 1 Adolescents with CP: 1 Gender: 1 male, 1 female Age: mdn 9.3y (range 6.9–11.7y) Type of CP: NS TD of CP: 2 hemiplegia	Active wrist extension range of movement (ROM): Goniometer Hand function: three tasks based in the Jebsen Taylor Hand Function Test	BFS	Electrical stimulation (ES, NMES, FES)	Neuromuscular electrical stimulation (FESMED dual-channel NMES unit): wrist extensors
					BFS	Electrical stimulation (ES, NMES, FES)	Neuromuscular electrical stimulation (FESMED dual-channel NMES unit): alternately wrist extensors and flexors
Beauregard et al.^Citation25 (1998)	USA	To examine the reaching movement in two children with CP and upper extremity movement dysfunction while they were playing a game and while performing a rote movement.	n = 2 Children with CP: 2 Gender: 2 female Age: mdn 4y 8mo (range 4y 4mo – 4y 11mo) Type of CP: 2 spastic TD of CP: 2 diplegia	Quality of reach: movement time, movement units, displacement, peak velocity, and the percentage of reach at which peak velocity occurred	BFS	Other	Reaching: Rote movement
					BFS	Other	Reaching: Playing a game
Brien et al.^Citation26 (2010)	Canada	To evaluate an intensive short-duration virtual reality (VR) intervention for balance and functional mobility in adolescents with cerebral palsy (CP) in Gross Motor Functional Classification System (GMFCS) Level I.	n = 4 Adolescents with CP: 4 Gender: 4 male Age: range 13y – 18y Type of CP: 3 spastic, 1 dyskinetic TD of CP: 3 diplegia, 2 NS	Community Balance and Mobility Scale (CB&M) 6-minute walk test (6MWT) Gross Motor Function Measure (GMFM)-Dimension E	BFS + A	Interactive video gaming or virtual reality
Brien et al.^Citation27 (2011)	Canada	To examine the effect of an intensive virtual reality (VR) training program on functional balance and mobility in 13- to 18-year-old adolescents with CP classified at GMFCS level I.	n = 4 Adolescents with CP: 4 Gender: 4 male Age: mdn 15y 10mo (range 13y 9mo – 18y 9mo) Type of CP: 3 spastic, 1 dyskinetic TD of CP: 3 diplegia, 1 NS	Community Balance and Mobility Scale (CB&M) 6-Minute Walk Test (6MWT) Timed Up and Down Stairs (TUDS) GMFM Dimension E	BFS + A	Interactive video gaming or virtual reality
Chen et al.^Citation28 (2007)	Taiwan	To investigate the training effects of a VR intervention on reaching behaviors in children with spastic CP.	n = 4 Children with CP: 4 Gender: 3 male, 1 female Age: mdn 5y 10mo (range 4y 8mo – 8y 6mo) Type of CP: 4 spastic TD of CP: 1 hemiplegia, 3 quadriplegia	Reaching kinematics: movement time (MT), path length (PATH), peak velocity (PV), and movement units (MUs) Fine Motor Domain of the Peabody Developmental Motor Scales–Second Edition (PDMS-2)	BFS + A	Interactive video gaming or virtual reality
Coker et al.^Citation29 (2009)	USA	To evaluate the effectiveness of a mCIMT program for a child less than one year of age diagnosed with congenital hemiplegic cerebral palsy.	n = 1 Children with CP: 1 Gender: 1 male Age: 5mo Type of CP: NS TD of CP: 1 hemiplegia	Peabody Developmental Motor Scales-2 (PDMS-2) Gross Motor Function Measure-88 (GMFM-88)	A	CIMT (Constrained Induced Movement Therapy)	Modified CIMT
Corn et al.^Citation30 (2003)	Australia	To investigate the effectiveness of Second Skin lycra splints to improve the quality of upper limb movement for children with spasticity.	n = 4 Children with CP: 1 Adolescents with CP: 1 Other: 2 Gender: 1 male, 1 female Age: mdn 12y (range 8y – 16y) Type of CP: 2 spastic TD of CP: 2 quadriplegia	Quality of upper limb movement (Melbourne Assessment)	BFS + A	Orthotics (splints)	Lycra splints
Costigan et al.^Citation31 (2010)	USA	To investigate the effect of seated position on the accuracy and efficiency of upper-extremity access to AAC for a child with CP.	n = 1 Children with CP: 1 Gender: 1 male Age: 5y Type of CP: NS TD of CP: 1 quadriplegia	Frequency of Accurate Selection Response Time to Accurate Selection	A	Seating and positioning	Typical seating
					A	Seating and positioning	Conventions of functional seating
Crocker et al.^Citation32 (1997)	Canada	To determine the effect of restriction of the less-involved upper extremity on spontaneous use of the more-involved upper extremity in children with spastic cerebral palsy.	n = 2 Children with CP: 2 Gender: 1 male, 1 female Age: mdn 2.5y (range 2y – 3y) Type of CP: 2 spastic TD of CP: 2 hemiplegia	Specific upper-extremity behaviors (videotaping): Stabilization and weight bearing; grasp and release; sensory exploration; bimanual assist; index finger use in isolation; push and pull; communication and gesture Use of the more-involved extremity for fine motor tasks: The Peabody Developmental Fine Motor Scale (PDMS-F). Daily log (feeding task)	A	Forced-use therapy (FUT)	Custom resting splint, to be worn for most of the waking hours.
							Regular therapy sessions
Dada et al.^Citation33 (2009)	South Africa	To evaluate the effect of an aided language stimulation program on the acquisition of target vocabulary items in children with LNFS (little or no functional speech).	n = 4 Children with CP: 2 Adolescents with CP: 2 Other: 1 Gender: 1 male, 2 female Age: mdn 8.5y (range 8.1y – 10.1y) Type of CP: NS TD of CP: 3 hemiplegia	Acquisition of the target vocabulary items (probe test)	A	Alternative and augmentative communication	Aided language stimulation (simultaneously pointing to symbols on a communication board in conjunction with providing ongoing spoken language stimulation) Three activities: arts and crafts, food preparation, and story time
DeGangi et al.^Citation34 (1983)	USA	To measure the effects of NDT compared to nonspecific play on qualitative motor performance (immediately after a therapy session), in children with CP.	n = 4 Children with CP: 4 Gender: 4 female Age: mdn 18mo (range 10mo – 22mo) Type of CP: 2 spastic, 2 NS TD of CP: 1 hemiplegia, 1 diplegia, 1 quadriplegia, 1 NS	Postural tone Weight shift and weight bearing in developmental postures Transitional movements Functional gains		Neurodevelopmental therapy (NDT, Bobath)
						Play therapy
Dinomais et al.^Citation35 (2013)	France	To examine the ‘‘clinical utility’’ of a virtual rehabilitation tool on upper-limb function in patients with cerebral palsy (CP).	n = 2 Adolescents with CP: 2 Gender: 2 male Age: 16y Type of CP: 2 spastic TD of CP: 2 hemiplegia	Hand function: Box and Block Test (BBT) Hand function: Jebsen–Taylor test (JTT)	A	Interactive video gaming or virtual reality	Training of both upper extremities
Do et al.^Citation36 (2016)	South Korea	To evaluate the effect of virtual reality based bilateral upper limb exercise on upper limb functions and bilateral hand coordination ability in hemiplegic cerebral palsy children.	n = 3 Children with CP: 3 Gender: 2 male, 1 female Age: mdn 6y (range 5y – 7y) Type of CP: 3 spastic TD of CP: 3 hemiplegia	Hand movement ability with complicated movement tasks: Wolf Motor Function Test (WMFT) Independent use of affected upper limb: Pediatric Motor Activity Log (PMAL) Bilateral hand coordination ability: putting a basketball through the hoop Bilateral hand coordination ability: moving large lightweight boxes	A	Interactive video gaming or virtual reality	Training of both upper extremities
Durfee et al.^Citation37 (1999)	USA	To compare the effect of the Touch Window versus the mouse with an enlarged on-screen arrow and cursor when matching upper-case letters of the alphabet for a child with both motor control impairments and visual impairments.	n = 1 Children with CP: 1 Gender: 1 male Age: 9y Type of CP: 1 spastic TD of CP: 1 quadriplegia	The percentage of letters correctly matched per session The total number of correct matches per letter Rate, expressed as the number of letters per minute accurately matched	A	Computer training	Touch Window
					A	Computer training	Mouse
Fienup et al.^Citation38 (2013)	USA	To compare the effect of two interventions targeting money-counting in a student with brain injury.	n = 1 Adolescents with CP: 1 Gender: 1 male Age: 15y Type of CP: NS TD of CP: NS	Counting accuracy: correctly stating the total value of multiple coins or bills Sorting accuracy: organizing bills and coins based on their denomination in a decreasing order Sequential counting: counting bills and coins in descending values	A	Education (money counting skills)	Education as usual
					A + P	Education (money counting skills)	Pre-teaching organization and error-correction
Fragala et al.^Citation39 (2002)	USA	To examine the effect of BtA injections (botox A injections) on impairment, spasticity, disability and parent satisfaction in children with cerebral palsy and lower-extremity spasticity.	n = 7 Children with CP: 5 Adolescents with CP: 2 Gender: 6 male, 1 female Age: mdn 4y 7mo (range 2y 9mo – 11y 10mo) Type of CP: 7 spastic TD of CP: 3 hemiplegia, 4 diplegia	Passive range of motion (hip flexion, extension, and abduction; popliteal angle; knee flexion and extension; ankle dorsiflexion): goniometry Spasticity (hip flexors, extensors and adductors; knee flexors and extensors; and ankle plantar flexors): MAS Disability: COPM Performance score Parents satisfaction: COPM Satisfaction score	BFS + A	Botulinum toxin (BoNT-A)	BoNT-A in the lower extremity
Ghorbani et al.^Citation40 (2017)	Iran	To investigate whether the CO-OP approach leads to motor-based, occupational performance goal achievement and improvement in motor skills in children with CP.	n = 6 Children with CP: 6 Gender: NS Age: mdn 7y 5mo (range 7y – 9y) Type of CP: 3 spastic, 1 dyskinetic, 1 NS TD of CP: 3 hemiplegia, 1 triplegia, 1 NS	Motor skills including gross motor, fine motor, and a combination of both: Bruininks-Oseretsky Test of Motor Proficiency Measure (BOTMP) Change in performance and satisfaction of personal goals: Canadian Occupational Performance Measure (COPM); Goal Attainment Scale (GAS)	A and P	Other	Cognitive Orientation to daily Occupational Performance (CO-OP)
Goodman et al.^Citation41 (1991)	USA	To measure the effects of a short thumb opponens splint on hand function and on the underlying aspects of hand function in a child with moderate spastic quadriplegic CP.	n = 1 Children with CP: 1 Gender: 1 female Age: 4y Type of CP: 1 spastic TD of CP: 1 quadriplegia	Active range of motion of the carpometacarpal joint in radial abduction, palmar abduction, and opposition Grip (spherical grasp) strength: Martin Vigorimeter Tip, lateral, and palmar pinch: B&L Pinch Gauge Scores on the Box and Block Test of manual dexterity The average number of I-in. cubes stacked in three trials Observation of qualitative aspects of prehension: the Erhardt Developmental Prehension Assessment and photographs.	BFS + A	Orthotics (splints)	Short thumb opponens splint
Hamill et al.^Citation42 (2007)	USA	To examine the effects of hippotherapy on sitting and parental perception in young children with CP classified in Level V on the Gross Motor Function Classification System (GMFCS).	n = 3 Children with CP: 3 Gender: 2 male, 1 female Age: mdn 3y 6mo (range 2y 3mo – 4y 6mo) Type of CP: 1 dyskinetic, 1 ataxic, 1 mixed TD of CP: 3 quadriplegia	Sitting: Gross Motor Function Measure-88 item version (GMFM-88) Dimension B (Sitting) Gross motor function: Gross Motor Function Measure-88 item version (GMFM-88) Postural control and functional control of arms and hands for different sitting positions: Sitting Assessment Scale (SAS) Anticipated and perceived effects and perceived value of the hippotherapy program: parent questionnaires (Likert scale and open-ended questions)	A	Hippotherapy
Harris et al.^Citation43 (1986)	USA	To examine the effects of inhibitive orthoses on both quantitative and qualitative aspects of independent standing for a child with spastic quadriplegic CP.	n = 1 Children with CP: 1 Gender: 1 male Age: 4.5y Type of CP: 1 spastic TD of CP: 1 quadriplegia	Independent standing balance (timed)			Barefoot
					A	Orthotics (splints)	Inhibitive ankle-foot orthoses
Hartveld et al.^Citation44 (1996)	UK	To examine the effects of Standing balance and weightshifting practice on the Compex board on the standing balance of children with diplegic CP.	n = 4 Children with CP: 2; Adolescents with CP: 2 Gender: 2 male, 2 female Age: mdn 9y (range 5y – 16y) Type of CP: 3 spastic, 1 mixed TD of CP: 4 diplegia	Standing balance (a quantitative test for standing balance after Atwater et al.)			Regular physiotherapy maintenance programme of functional training and some stretching exercises
					BFS	Other	Shifting weight from left to right and forwards and backwards while standing on the current production version of the Compex board, using a variety of games.
Havstam et al.^Citation45 (2003)	Sweden	To examine the ffect of the speech recognition system Dragon Dictate on computer use and speech development in individuals with CP with (severe) dysarthria.	n = 2 Adolescents with CP: 1 Other: 1 Gender: 1 male Age: 12y Type of CP: 1 spastic TD of CP: 1 quadriplegia	Computer access system efficiency (five minute typing test) Perceived performance and comfort (participant rated) Level of success in speech recognition (five-minute test) Speech production (videotape recording) Speech production (word repetition and word naming)	A + P	Alternative and augmentative communication	Original system (a computer, a head switch and the scanning software SAW (Switch Access to Windows) for text entry and other forms of computer access; a Bliss chart and a light pointer attached to the eyeglasses)
					A + P	Alternative and augmentative communication	The speech recognition system Dragon Dictate
Hinderer et al.^Citation46 (1988)	USA	To compare the effects of tone-reducing vs. standard immobilization casts on the gait patterns of children with CP.	n = 2 Children with CP: 2 Gender: 1 male, 1 female Age: mdn 4y 8mo (range 3y 6mo – 5 y 9 mo) Type of CP: 2 spastic TD of CP: 2 diplegia	Gait (footprint data) Fine motor skills (timed performance)			No casting
					BFS + A	Casting	Tone-reducing casts, operationally defined as casts which maintain the ankle at zero degrees and stabilize the toes and foot in neutral alignment by incorporating a footplate which supports the toes and the metatarsal, peroneal and longitudinal arches of the foot
					BFS + A	Casting	Standard casts, defined as casts which hold the ankle at zero degrees but do not include a footplate
Iammatteo et al.^Citation47 (1990)	USA	To examine the effect of techniques used to facilitate mouth closure and swallowing on drooling and bilabial sounds produced in children with CP.	n = 2 Children with CP: 2 Gender: 2 male Age: mdn 2y 9mo (range 2y 7mo – 2y 11mo) Type of CP: 2 spastic TD of CP: 1 diplegia, 1 quadriplegia	Drooling (saliva weight) Bilabial vocalization (counted)			No intervention
					BFS	Oral motor treatment	Techniques used to facilitate mouth closure and swallowing recommended by Morris and Mueller
Jelsma et al.^Citation48 (2013)	South Africa	To compare the effect of a Nintendo Wii Fit intervention compared to conventional physiotherapy on balance and gross motor function in children with spastic hemiplegic CP performing at GMFCS level I or II.	n = 14 Children/adolescents with CP: 14 Gender: 8 male, 6 female Age: mean 11.4y (range 7y – 14y) Type of CP: 14 spastic TD of CP: 14 hemiplegia	Balance and gross motor function: a set of adapted, criterion-referenced activities based on the Bruininks–Oseretsky test of Motor Proficiency 2nd Edition (BOT-2) Balance: Subtests 5 Running speed and agility, RSA: 6 of the BOT-2 Functional mobility: the timed up and down stairs test (TUDS)			Conventional physiotherapy (children continued their standard physiotherapy regimen of twice weekly 30min sessions)
					BFS + A	Interactive video gaming or virtual reality	Nintendo Wii Fit balance board
Johnson et al.^Citation49 (2014)	USA	To evaluate the effects and optimal duration of plyometric training on gross motor abilities in children with CP.	n = 3 Children with CP: 2 Adolescents with CP: 1 Gender: 3 male Age: mdn 9y 11 mo (range 8y 9mo – 10y) Type of CP: 3 spastic TD of CP: 3 hemiplegia	Gross motor ability: GMFM D, E, 66, and percentile rank Running speed: 20-m running-start sprint Agility: 10 × 5-m Sprint Muscle power: basketball throw, broad jump, vertical jump Self-selected goal Safety	BFS + A	Strength training (resistance)	Plyometric training intervention
Kelly et al.^Citation50 (2009)	Canada	To examine the effect of a community-based group aquatic exercise program on energy expenditure and children’s self-reported levels of fatigue in children with CP.	n = 5 Children with CP: 1 Adolescents with CP: 4 Gender: NS Age: mdn 10y (range 9y – 11y) Type of CP: 5 spastic TD of CP: 2 hemiplegia, 2 diplegia, 1 quadriplegia	Energy expenditure: Energy Expenditure Index (EEI) Children’s perception of fatigue: Pediatric Quality of Life Multidimensional Fatigue Scale (PedsQL-Fatigue Scale) Children’s performance and satisfaction ratings of a movement goal: Canadian Occupational Performance Measure (COPM)	BFS + A	Other	Aquatic aerobic exercise program
Kenyon et al.^Citation51 (2018)	USA	To evaluate the impact of power mobility training using an alternative power mobility device with a child who has multiple, severe impairments on the child’s spectrum of EEG activity and mastery motivation.	n = 1 Children with CP: 1 Gender: 1 female Age: 3y 2mo Type of CP: 1 spastic TD of CP: 1 quadriplegia	Mastery motivation: dimensions of mastery questionnaire (DMQ) EEG data collected under various conditions Paediatric evaluation of disability inventory–computer adaptive test (PEDI-CAT) Canadian occupational performance measure (COPM) Assessment of learning powered mobility use (ALP) Maternal interview to determine the mother’s perceptions of her child’s response to power mobility training	A and P	Other	Power mobility training
Kinghorn et al.^Citation52 (1996)	Canada	To investigate the effects of an upper-extremity weight-bearing splint on muscle tone and functional hand skills in a child with CP.	n = 1 Children with CP: 1 Gender: 1 male Age: 20mo Type of CP: 1 spastic TD of CP: 1 quadriplegia	Hand contact with a surface (hand tracing) Arm-hand posture: elbow and hand position (observation rating scale) Hand function: grasping of a 1-in. block cube; batting or trapping of a ball	BFS + A	Other	Weight bearing without splint
					BFS + A	Other	Weight bearing with upper-extremity weight-bearing splint
Ko et al.^Citation53 (2015)	South Korea	To evaluate the effect of AST (Adeli suit treatment) on spatial-temporal gait parameters, 10-meter walking speed, pediatric balance scale (PBS) and GMFM in a child with CP.	n = 1 Children with CP: 1 Gender: 1 female Age: 8y Type of CP: 1 spastic TD of CP: 1 diplegia	Spatial-temporal gait: parameters gait speed, stride time, stride length and stride width Walking speed: 10-meter walking test Motor function: Gross motor functional measure (GMFM) Balance: Pediatric balance scale (PBS)			Regular physical and occupational therapy treatments
					BFS + A	Other	Adeli suit treatment (and regular physical and occupational therapy treatments)
Laessker-Alkema et al.^Citation54 (2016)	Sweden	To examine the efficacy of knee orthoses on extensibility of the hamstrings in children with spastic CP.	n = 10 Children with CP: 4 Adolescents with CP: 6 Gender: 6 male, 4 female Age: mdn 12.0y (range 5.8y −14.7y) Type of CP: NS TD of CP: 4 diplegia, 6 NS	Passive ROM (hamstrings and knee extension): hand-held plastic goniometer Muscle tone (hamstrings): Modified Ashworth Rating Scale (MAS) Gross motor function: Gross Motor Function Measure (GMFM), GMFM-88 version	BFS	Orthotics (splints)	Knee orthoses for both legs (intervention position: sitting, well supported, with the hips up to 90◦ of flexion, the knees extended, and the legs slightly in abduction)
Lammi et al.^Citation55 (2003)	Canada	To determine if occupational therapy intervention, using modified FCFT (Family-Centred Functional Therapy), focused on changing only the task and the environment, could improve the functional performance of preschool aged children with CP.	n = 3 Children with CP: 3 Gender: 2 male, 1 female Age: mdn 3y 6mo (range 3y – 3y 6mo) Type of CP: NS TD of CP: NS	Performance and satisfaction of tasks: Canadian Occupational Performance Measure (COPM) The level of independence of the child: Paediatric Evaluation of Disability Inventory (PEDI); Caregiver Assistance Scale Level of achievement for task performance and the quality of performance: Performance Quality Rating Scale (PQRS)	A + P	Goal-directed training/functional training	Family-Centred Functional Therapy (FCFT)
Laskas et al.^Citation56 (1985)	USA	To examine the effect of particular NDT activities on heel to floor contact and dorsiflexor muscle activity in a child with quadriplegic CP.	n = 1 Children with CP: 1 Gender: 1 male Age: 2.5y Type of CP: 1 spastic TD of CP: 1 quadriplegia	Dorsiflexor muscle activity during three posterior equilibrium reactions: biofeedback Heel contact when coming to a standing position			Routine physical therapy services (once a week)
					BFS + A	Neurodevelopmental therapy (NDT, Bobath)	Routine physical therapy services (once a week) continued during intervention phase
Lephart et al.^Citation57 (2015)	USA	To compare the effects of 2 seating systems used in a school setting on cardiopulmonary functioning (body function and structure), time to activate communicationswitches (activity), and accuracy of answering questions participation) for a young adult with quadriplegic CP and neurological scoliosis.	n = 1 Adolescents with CP: 1 Gender: 1 male Age: 19y Type of CP: 1 spastic TD of CP: 1 quadriplegia	SaO2: digital fingertip pulse oximeter Heart rate (HR): digital fingertip pulse oximeter Respiration rate (RR): number of chest excursions in 1 minute Body temperatoure BT: factory-calibrated noncontact infrared thermometer Activity: the time it took to reach a switch from a standardized starting position after a question was asked Participation: the number of correct answers to 5 questions	BFS + A + P	Seating and positioning	Original chair position (OCP): Contoured Advance Seat
					BFS + A + P	Seating and positioning	Custom-molded back (CMB)
Lewis et al.^Citation58 (2016)	Australia	To assess the effect of the dynamic elastomeric fabric orthoses (DEFO) in the management of shoulder instability in children with CP, GMFCS levels IV and V.	n = 3 Children with CP: 1 Adolescents with CP: 2 Gender: 2 male, 1 female Age: mdn 11y (range 6.5y – 16y) Type of CP: 3 dyskinetic TD of CP: 3 quadriplegia	Distance from the tip of acromion process to humeral head (mm) Shoulder passive range of motion Participant/Parent evaluation: rating scale Care and Comfort Hypertonicity Questionnaire A daily log of compliance	BFS	Orthotics (splints)	Dynamic elastomeric fabric orthoses (DEFO)
Lilly et al.^Citation59 (1990)	USA	To measure the effects of NDT and play intervention on dressing skills in children with CP.	n = 2 Children with CP: 2 Gender: 2 female Age: mdn 29.5mo (range 27mo – 32mo) Type of CP: 2 spastic TD of CP: 2 diplegia	Sets of measurement criteria for dressing skills	BFS + A	Neurodevelopmental therapy (NDT, Bobath)
					BFS + A	Play therapy
Lin et al.^Citation60 (2014)	Taiwan	To assess whether physical activities increase by employing MaKey–MaKey-based task systems in kindergarten children with severe CP.	n = 2 Children with CP: 1 Other: 1 Gender: 1 female Age: 5y 9mo Type of CP: NS TD of CP: NS	The number of times the participant applied his/her physical strength to touch any part of the conductive material as a physical activity	A	Interactive multimedia training	MaKey–MaKey-based task systems
Lin et al.^Citation61 (2015)	Taiwan	To assess whether interactive augmented reality using Scratch 2.0 increases physical activitiy in children with developmental disabilities.	n = 3 Children with CP: 2 Other: 1 Gender: 1 male, 1 female Age: mdn 5y (range 3y 11mo – 6y) Type of CP: NS TD of CP: NS	Number of correct movements	A		Evoking of movements without interactive augmented reality
					A	Interactive augmented reality	Scratch 2.0 software
Mackey^Citation62 (1989)	UK	To evaluate the difference in the ability of children with CP to carry out a motor control task whether it is performed with therapist feedback alone or with the addition of computer-assisted feedback.	n = 6 Children with CP: 5 Adolescents with CP: 1 Gender: 4 male, 2 female Age: mdn 7y (range 4y −12y) Type of CP: 6 spastic TD of CP: 2 diplegia, 4 quadriplegia	The amount of time spent pressing down a target	BFS + A	Other	Pressure on a switchbox with only oral feedback by the therapist
					BFS + A	Other	Pressure on a switchbox with visual and auditory cues from the computer feedback in addition to oral feedback by the therapist (computer-assisted feedback)
Man et al.^Citation63 (2007)	China	To compare the effects of various available non hand-held pointer interfaces on communication time, precision and comfort in students with severe disabilities.	n = 2 Adolescents with CP: 2 Gender: 1 male, 1 female Age: mdn 14y (range 13y – 15y) Type of CP: 2 dyskinetic TD of CP: 2 quadriplegia	Movement time and accuracy: WinFitts multidirectional point-and-click test Level of comfort: questionnaire	A	Other	Computer-Access Solution: nonhandheld pointing interfaces (CameraMouseTM)
					A	Other	Computer-Access Solution: ASL Head-Array Mouse Emulator
					A	Other	Computer-Access Solution: CrossScanner
					A	Other	Computer-Access Solution: Quick Glance Eye Gaze Tracker 2SH version 4.2
Matthews et al.^Citation64 (2009)	UK	To identify gait and mobility-related changes in a group of children with CP who had been prescribed DEFO leggings.	n = 8 Children/adolescents with CP: 8 Gender: 4 male, 4 female Age: mdn 5.5y (range 3y – 13y) Type of CP: 8 spastic TD of CP: 8 diplegia	Gait velocity: Ten Metre Walking Test (10MWT) Perceived changes in gait: Visual Analogue Scale (VAS) Energy expenditure during gait: Physiological Cost Index (PCI) Mobility-related functional goal: Patient Specific Functional Scale (PSFS) Perceived salient features of mobility and the total number of hours of orthotic wear each day: daily diary	BFS + A	Orthotics (splints)	Dynamic elastomeric fabric orthotic (DEFO) leggings
McCarthy et al.^Citation65 (2015)	USA	To investigate if providing computerized sounding out, or auditory providing the sound sequences of words through a digitized recording of a human voice, during spelling would increase spelling accuracy for children with CP related dysarthria, who did and did not use AAC.	n = 8 Children with CP: 6 Adolescents with CP: 1 Not specified: 1 Gender: 5 male, 3 female Age: mdn 8y (range 5y 8mo – 11y 3mo) Type of CP: NS TD of CP: NS	Correctly spelled pseudo-words at the whole word level Percentage of correctly used phonology (i.e. sounds in the pseudo-words)	A	Other	Spelling education without computerized sounding out
					A	Other	Spelling education with the computerized sounding out
McConnell et al.^Citation66 (2014)	UK	To evaluate the efficacy of reduced intensity CIMT on unilateral upper limb function in children 8–15 years old with hemiplegic CP and exploring the acceptability of the CIMT protocol from the perspectives of children, parents and education staff.	n = 7 Children with CP: 4 Adolescents with CP: 3 Gender: 4 male, 3 female Age: mdn 9y (range 9y – 11y) Type of CP: 7 spastic TD of CP: 7 hemiplegia	Quality of unilateral UL function: Melbourne Assessment of Unilateral Upper Limb Function (MUUL) Strength: Manual Muscle Testing (MMT) Tone: Modified Ashworth Scale (MAS) Active range of motion (AROM): goniometry Participation: shortened parent-proxy Assessment of Life-Habits (Life-H) questionnaire (version 1.0) adapted for children five to 13 years Interviews	BFS + A + P	CIMT (Constrained Induced Movement Therapy)	Reduced intensity CIMT
Owen et al.^Citation67 (1992)	Australia	To evaluate the usefulness of benztropine on reduction of drooling in school aged children with CP.	n = 3 Children with CP: 2 Adolescents with CP: 1 Gender: NS Age: mdn 9.8y (range 5.11y – 12.6y) Type of CP: NS TD of CP: NS	Saliva volume: observational rating scale 0–2	BFS	Other	Benztropine
							Placebo
Pinder et al.^Citation68 (1995)	USA	To investigate whether treatment for shaping intentional communicative signals could effectively influence the development of intentional communicative signals produced by young children who demonstrated different degrees of motor disability.	n = 4 Children with CP: 4 Gender: 2 male, 2 female Age: mdn 13mo (range 11.5mo – 13.5mo) Type of CP: 2 spastic, 2 mixed TD of CP: 2 diplegia, 2 NS	Communicative signals: in two different context (choice and request opportunities), with toys and with food and drink	A	Early intervention	Treatment for shaping intentional communicative signals
Pool et al.^Citation69,Citation70 (2014)	Australia	To determine the effects of functional electrical stimulation (FES) with the Walk Aide on the main impairments affecting gait and participation (trips and falls) in children with spastic hemiplegic CP.	n = 12 Children with CP: 7 Adolescents with CP: 5 Gender: 7 male, 5 female Age: mdn 9y 4mo (range 4y 10mo – 15y 5mo) Type of CP: 12 spastic TD of CP: 12 hemiplegia	Passive range of motion: goniometer Spasticity: dynamic ROM (Modified Tardieu Scale); Australian Spasticity Assessment Scale Strength: handheld dynamometry; single-limb heel raises Selective Motor Control: assessment as described by Boyd and Graham Balance: single-limb balance assessment Gait: 2D video using the Observational Gait Scale (OGS) Self-reported Toe Drag and Falls: questionnaire	BFS + A	Electrical stimulation (ES, NMES, FES)	A portable Functional electrical stimulation (FES) (the Walk Aide)
Pratt et al.^Citation71 (2016)	USA	To investigate the effect of ADA-compliant versus noncompliant playgrounds on the physical activity of children with ambulatory CPduring playground play.	n = 5 Children with CP: 5 Gender: 4 male, 1 female Age: mdn 8y 1mo (range 7y 4mo – 8y 5mo) Type of CP: 5 spastic TD of CP: 2 hemiplegia, 2 diplegia, 1 quadriplegia	Physical activity (Stride Count and time active): StepWatch activity monitor	A + P	Other	Noncompliant playground
					A + P	Other	ADA-compliant playground
Prosser et al.^Citation72,Citation73 (2012)	USA	To investigate the feasibility, tolerability and effect on motor development of a novel mobility training intervention incorporating infant motor learning and neurorehabilitation principles in toddlers with CP.	n = 5 Children with CP: 5 Gender: NS Age: mdn 17mo (range 13mo – 27 mo) Type of CP: NS TD of CP: 5 bilateral	Motor function: Gross Motor Function Measure (GMFM-66) Feasibility: completion rate (percentage of therapy sessions attended); participation rate (percentage of therapy time that participants were actively engaged in motor activity)			Maintenance of the therapy services they were already receiving
					A	Other	Mobility training by means of an advanced weight support system
Ramani et al.^Citation74 (2014)	India	To study the impact of low vision care intervention on the reading progress of children with multiple disabilities and visual impairment (MDVI) as measured by reading speed, accuracy percentage and fluency.	n = 3 Children with CP: 1 Adolescents with CP: 2 Gender: 3 male Age: mdn 13y (range 9y – 14y) Type of CP: NS TD of CP: NS	Reading speed: read words/letters per minute Reading accuracy: error rate Reading fluency: multidimensional fluency rating scale	BFS + A	Other	Low vision care intervention
Ray et al.^Citation75 (1983)	USA	To determine the effect of facilitating mouth closure on drooling in an adolescent with mental retardation and CP.	n = 1 Adolescents with CP: 1 Gender: 1 male Age: 11y 8mo Type of CP: 1 spastic TD of CP: 1 quadriplegia	Saliva volume: collected by an absorbent bib (in centigrams)			Play
					BFS	Oral motor treatment	Methods described by Mueller to facilitate jaw closure
Reid et al.^Citation76 (1999)	Canada	To evaluate the effect of a rigid pelvic stabilizer compared to a lap belt on task performance and satisfaction in wheelchair-dependent children with CP.	n = 6 Children with CP: 3 Adolescents with CP: 3 Gender: 2 male, 4 female Age: mdn 10.21y (range 8.33y – 12.75y) Type of CP: 6 spastic TD of CP: 1 diplegia, 5 quadriplegia	Self-perception of children and parents of performance and satisfaction relative to problem areas: Canadian Occupational Performance Measure (COPM) The caregivers’ perception of improvement regarding their child’ s performance: a single-item measure which was part of a structured interview developed for this study	A + P	Seating and positioning	Standard single-anchor-point wheelchair lap belt
					A + P	Seating and positioning	Rigid pelvic stabilizer (RPS)
Retarekar et al.^Citation77 (2009)	USA	To evaluate the effects of an aquatic aerobic exercise program on participation, gross motor function, walking endurance, and energy expenditure in an ambulatory child with CP	n = 1 Children with CP: 1 Gender: 1 female Age: 5y Type of CP: 1 spastic TD of CP: 1 diplegia	Mother’s perception of performance and satisfaction in the areas of self-care, productivity, and leisure: Canadian Occupational Performance Measure (COPM) Gross motor function: Gross Motor Function Measure-66 (GMFM-66) Submaximal functional exercise capacity: 6-Minute Walk Test Energy expenditure during walking: Modified Energy Expenditure Index Physical activity: Physical Activity Questionnaire			Usual care only (physical therapy and physical education)
					BFS + A	Other	Aquatic aerobic exercise
Rigby et al.^Citation78 (2001)	Canada	To evaluate task performance and caregiver assistance for children with CPusing a rigid pelvic stabilizer system compared with a lap belt on their adaptive wheelchair seating system.	n = 6 Children/adolescents with CP: 6 Gender: 2 male, 4 female Age: mean 10y 4mo Type of CP: 6 spastic TD of CP: 1 diplegia, 5 quadriplegia	The amount of caregiver assistance required by the child to engage in each of the occupational performance areas identified on the COPM: Caregiver Assistance Scale modified from the PEDI	A + P	Seating and positioning	A seating system lap belt, anchored at two points
					A + P	Seating and positioning	Rigid pelvic stabilizer (RPS)
Rintala et al.^Citation79 (1994)	NS	To document the effects of a 4-month physical activity program on posture control in children with CP.	n = 8 Children with CP: 3 Adolescents with CP: 5 Gender: 5 male, 3 female Age: mdn 10.5y (range 7y – 11y) Type of CP: NS TD of CP: 5 hemiplegia, 3 diplegia	Postural sway during standing: velocity of the center of pressure during 1) normal bipedal standing; 2) normal bipedal standing with eyes closed; and 3) toe-to-heel standing			No organized physical activities
					BFS + A	Other	Physical activity program, including ball skills, balance, and physical fitness
Rivi et al.^Citation80 (2014)	Italy	To assess the effectiveness of the standing frame on constipation in a chronically constipated child with quadriplegic CP.	n = 1 Children with CP: 1 Gender: 1 male Age: 5y Type of CP: NS TD of CP: 1 quadriplegia	Constipation (including evacuation, induction and pain): daily diary Type of stool: The Bristol Stool Scale Burden on caregivers because of the evacuation management: self-developed questionnaire			No use of the standing frame
					BFS	Other	Use of a standing frame
Sakemiller et al.^Citation81 (1998)	USA	To explore whether a meaningful game would elicit a higher degree of functional vertical neck and back extension than would rote exercises in children with hypotonic CP.	n = 2 Adolescents with CP: 2 Gender: 2 female Age: mdn 10.9y (range 10.6y – 11.2y) Type of CP: NS TD of CP: NS	Neck extension: videotaped two-dimensional biomechanical analysis Back extension: videotaped two-dimensional biomechanical analysis Total extension: videotaped two-dimensional biomechanical analysis	BFS + A	Other	A verbal request to hold the head up.
					BFS + A	Other	Meaningful games to motivate head and back extension
Sheppard et al.^Citation82 (2007)	Australia	To determine whether the treatment approaches NDT versus BIT increased spontaneous hand use and improved upper limb movement patterns in children with hemiplegic CP.	n = 3 Children with CP: 3 Gender: 1 male, 2 female Age: mdn 9y 10mo (range 8y 1mo – 9y 11mo) Type of CP: 3 spastic TD of CP: 3 hemiplegia	Hand use: 1) frequency of use/percentage of time the affected hand was used; 2) absence or presence of movement characteristics (during specified test tasks)			No therapy
					BFS + A		NDT in combination with Bilateral Isokinematic Training (BIT)
					BFS + A	Other	Bilateral Isokinematic Training (BIT)
Shumway-Cook et al.^Citation83 (2003)	USA	To examine the effect of massed practice using a moving platform on recovery of stability in school-age children with CP.	n = 6 Children with CP: 4 Adolescents with CP: 2 Gender: 4 male, 2 female Age: mdn 8y 9mo (range 7y 5mo – 12y 11mo) Type of CP: 6 spastic TD of CP: 2 hemiplegia, 4 diplegia	Reactive balance control: center of pressure area and time to stabilization following a perturbation on a force platform Motor function: GMFM dimension D			Current schedule of therapies
					BFS	Other	Moveable forceplate system to train reactive balance control
Siconolfi-Morris^Citation84 (2012)	USA	To explore the use of a therapist direct, video game based, multi-directional balance training intervention on the balance, strength, GV and activities of children with CP in the home environment.	n = 5 Children with CP: 4 Adolescents with CP: 1 Gender: 3 male, 2 female Age: mdn 8.1y (range 7.0y – 10.4y) Type of CP: 5 spastic TD of CP: 2 hemiplegia, 3 diplegia	Balance, weight bearing symmetry (WBS): assessed using the WiiFit Plus software and Wii balance board (WBB) from the Nintendo Wii gaming console Balance, center of pressure and time to boundary: ground reaction force data was assessed on an Accusway forceplace Balance: timed up and go (TUG) Gait velocity: time to ambulate a 3m distance Isometric strength of the hip abductors: hand-held dynamometer Activities: Activities Scale for Kids, performance version (ASKp) Enjoyment and difficulty of the intervention: Enjoyment and Difficulty Questionnaire (EDQ) (adapted version)			Current therapies received
					BFS + A	Interactive video gaming or virtual reality	Balance training intervention utilizing the WiiFit Plus software and the WBB
Smelt^Citation85 (1989)	USA	To determine the effect of an upper extremity inhibitive weight-bearing mitt in conjunction with NDT, as compared to NDT alone, on the muscle tone and functional performance of a child with spastic CP.	n = 1 Children with CP: 1 Gender: 1 male Age: 17mo Type of CP: 1 spastic TD of CP: 1 hemiplegia	Passive range of motion in the hand: grip circumference Tone: hand weight-bearing surface area Functional performance: the ability to perform several play activities	BFS + A	Neurodevelopmental therapy (NDT, Bobath)	NDT alone
					BFS + A	Orthotics (splints)	Inhibitive weiht-bearing mitt in conjunction with NDT
Smithers^Citation86 (1991)	UK	To evaluate the effect of facilitation techniques on the rolling ability of a boy with athetoid CP and associated learning difficulties.	n = 1 Children with CP: 1 Gender: 1 male Age: 7y Type of CP: 1 dyskinetic TD of CP: NS	Quantity of supine to side-lying rolls: number of successfully completed rolls (frequency ticklist) Quality of supine to side-lying rolls: a modified version of a video-assessment of rolling drawn up by Montgomery and Connolly			No facilitating techniques
					A	Other	Facilitating techniques for rolling
Soto et al.^Citation87 (2008)	USA	To investigate the effect of an intervention program designed to improve the narrative skills of children who use AAC on the narrative performance of a 12-year-old who used AAC.	n = 1 Adolescents with CP: 1 Gender: 1 female Age: 12y Type of CP: 1 dyskinetic TD of CP: 1 quadriplegia	Linguistic Complexity: 1) total number of different words (NDW); 2) number of clauses; and 3) clausal density Story Complexity, 1) cohesiveness of the narratives: the presence of cohesion markers between clauses; 2) story coherence: the use of story grammar elements and the organization of story structure Ecological validity of the intervention: post-treatment interview Social validity of the intervention: a postintervention questionnaire			No intervention
					A + P	Other	A comprehensive narrative intervention program
Stansfield et al.^Citation88 (2015)	USA	To investigate whether the use of motion- based VR gameplay results in improved measures of motor performance in an adolescent with CP.	n = 1 Adolescents with CP: 1 Gender: 1 male Age: 10y Type of CP: NS TD of CP: NS	Speed of reach with the affected upper extremity: motion tracker data Upper extremity skills: the dissociated movement subscale from the Quality of Upper Extremity Skills Test (QUEST) Bilateral integration of the upper extremities in a functional task: the Goal Attainment Scale (GAS) Changes noted in the participant’s overall functioning, upper extremity functioning, mood and demeanor: daily log Satisfaction Rating Scale	BFS + A	Interactive video gaming or virtual reality	A motion-based VR game
Stewart et al.^Citation89 (1992)	Australia	To compare the effect of an isometric joystick and a displacement joystick on response time in wheelchair driving in adolescents with CP.	n = 6 Adolescents with CP: 6 Gender: 3 male, 2 female Age: mdn 18y (range 16y – 20y) Type of CP: 2 spastic, 3 dyskinetic TD of CP: 5 quadriplegia	Speed: tracking task (Skill Evaluator and Trainer) Accuracy: tracking task (Skill Evaluator and Trainer)	A	Other	Wheelchaire controls: isometric joystick
					A	Other	Wheelchaire controls: displacement joystick
Stewart et al.^Citation90 (2000)	Australia	To compare the effect of three different types of symbol selection on speed and accuracy of symbol selection in scanning voice output device users with CP.	n = 3 Children with CP: 2 Adolescents with CP: 1 Gender: 1 male, 2 female Age: mdn 8y 4mo (range 6y 8mo – 10y 2mo) Type of CP: 1 dyskinetic, 2 mixed TD of CP: 3 quadriplegia	Speed: the time in seconds for the subject to complete each sentence Errors: the number of symbols that were incorrectly selected	A + P	Alternative and augmentative communication	No-prediction
					A + P	Alternative and augmentative communication	Regular prediction
					A + P	Alternative and augmentative communication	Internal prediction
Svedberg et al.^Citation91 (2001)	Sweden	To assess acupuncture effects on peripheral skin temperature in children with cold feet associated with CP, progressive encephalopathy or Smith-Lemli-Opitz syndrome.	n = 6 Children with CP: 2 Adolescents with CP: 2 Other: 2 Gender: 4 female Age: mdn 10y (range 4y – 16y) Type of CP: 3 spastic, 1 dyskinetic TD of CP: 2 diplegia, 1 quadriplegia, 1 NS	Skin temperature: at the forehead, the suprasternal fossa, the palm and the dorsal surface of the hand, and the sole and dorsal surface of the foot			Normal continuation of pharmacological and physical treatment only
					BFS	Acupuncture	Acupuncture
Symons et al.^Citation92 (2007)	USA	To compare the effect of low dose methylphenidate, high dose methylphenidate and placebo on disruptive classroom behaviour in school-age children with CPand comorbid ADHD symptoms.	n = 3 Children with CP: 2 Adolescents with CP: 1 Gender: 2 male, 1 female Age: mdn 9y 8mo (range 8y 5mo – 11y 2mo) Type of CP: 2 spastic, 1 NS TD of CP: 1 triplegia, 1 quadriplegia, 1 NS	Stereotyped behavior: video-taped direct classroom observation Disruptive behavior: video-taped direct classroom observation Task-related behavior: video-taped direct classroom observation	A + P	Other	Placebo
					A + P	Other	Low dose of methylphenidate
					A + P	Other	High dose of methylphenidate
Thorpe et al.^Citation93 (2002)	USA	To examine the effects of presenting augmented information to children with CPon the learning of a novel motor gross motor task.	n = 13 Children with CP: 10 Adolescents with CP: 3 Gender: 5 male, 8 female Age: mdn 8.67y (range 6.00y – 12.67y) Type of CP: NS TD of CP: 9 diplegia, 4 quadriplegia	Total backward displacement of the Pedalo, travelled in 10 seconds	A	Other	No augmented information
					A	Other	Knowledge of performance (KP)
					A	Other	Knowledge of performance enhanced by a cognitive strategy (KP+CS)
Ustad et al.^Citation94 (2009)	Norway	To examine effects of intensive physiotherapy on the acquisition of gross motor functional skills compared to standard care in infants with CP.	n = 5 Children with CP: 5 Gender: 4 male, 1 female Age: mdn 7mo 2wk corrected age (range 5mo 3wk – 9mo 2wk) Type of CP: 4 spastic, 1 mixed TD of CP: 4 hemiplegia, 1 bilateral	Gross motor function: Gross Motor Function Measure (GMFM), both versions			Early intervention (physiotherapy as usual)
					A	Early intervention	Intensive physiotherapy (daily)
Voulgarakis et al.^Citation95 (2015)	USA	To examine the effect of an escape extinction and negative reinforcement-based approach to treating food refusal in a child with CP.	n = 1 Children with CP: 1 Gender: 1 male Age: 8y Type of CP: NS TD of CP: NS	Number of bites accepted within 30 minutes	A + P	Other	An escape extinction and negative reinforcement-based approach during mealtime 5 bites required – 7 bites required – 10 bites required – 7 bites required – 12 bites required
Ward et al.^Citation96,Citation97 (2013 & 2014)	Australia	To evaluate the effectiveness of a motor-speech intervention program (Prompts for Restructuring Oral Muscular Phonetic Targets) on perceptual changes in speech production accuracy in children with moderate-to-severe speech impairment associated with CP.	n = 6 Children with CP: 5 Adolescents with CP: 1 Gender: 3 male, 3 female Age: mdn 5y 3mo (range 3y – 11y 9 mo) Type of CP: 5 spastic, 1 dyskinetic TD of CP: 4 hemiplegia, 1 quadriplegia, 1 NS	Accuracy of the targeted motor speech movement pattern (MSMP): speech probes Perceptual accuracy (PA): speech probes			Regular therapy services
					A	Other	Prompts for Restructuring Oral Muscular Phonetic Targets (PROMPT)

^a for studies in which the baseline condition is not specified, only the experimental intervention(s) is described; NS = not specified; mdn = median; TD = topographical distribution; BFS = body functions and structures level of the International Classification of Functioning, Disability and Health (ICF); A = activity level of the ICF; P = participation level of the ICF.