Description of the condition
Children with developmental co-ordination disorder (DCD) (APA 2013) have significant difficulty in performing the essential motor tasks required for self care (for example, dressing), social and recreational activities (for example, riding a bicycle), and academic achievement (for example, handwriting) as compared with typically developing children of the same age. Additionally, the disturbance in movement skills is not explained by any known medical conditions (APA 2013).
A diagnosis of DCD is made if the child satisfies the diagnostic criteria from the Diagnostic and Statistical Manual for Psychiatric Disorders 5 (DSM-5) (APA 2013). The assessment involves taking a developmental history, performing a clinical examination to rule out possible medical conditions, assessing the child's functional motor skills (usually through parent or teacher report), and objectively assessing the child's motor competence using a performance-based motor assessment (Blank 2012). DCD is usually diagnosed between the ages of five and 16 years (Blank 2012). By definition, children with suspected DCD should be free from definite neurological conditions (Gibbs 2007); however, minor neurological dysfunctions are frequently reported in children with DCD, suggesting that early brain lesions might be causative (Hadders-Algra 2003). Moreover, studies on imaging report differences in neural networks and brain activation patterns between children with DCD and control children (Kashiwagi 2009; Zwicker 2010).
The prevalence of DCD has been cited as 6% of school-aged children (APA 2013) and the male to female ratio has been reported as 1.9 to 1 in a recent UK study of seven-year-olds (Lingam 2009). DCD may also be referred to as clumsy child syndrome, dyspraxia (Miyahara 2000), or specific developmental disorder of motor function (World Health Organization 2010). Currently, the DSM-5 criteria accept comorbidities of DCD with attention-deficit and/or hyperactivity disorder, communication disorders, intellectual disability, and specific learning disorders (APA 2013).
DCD is included in the manual of mental disorders because of its consequential avoidance behaviours and psychosocial impacts (Spitzer 1994). The self esteem of children with DCD, in terms of physical competence, is diminished to a greater extent than that of children with severe physical disabilities (Miyahara 2006). They are likely to be onlookers in playgrounds, isolated and solitary in the school yard (Smyth 2000). Rejection by their peers can lead to children with DCD missing out on important socialisation experiences, resulting in suboptimal social skills (Cummins 2005). They may be easy targets for bullies (Piek 2005). Their levels of depressive symptomatology and anxiety are higher than typically developing children (Schoemaker 1994) and adolescents (Cantell 1994). DCD influences children's physical functions and health status, as well as their emotional life and social participation, not only during childhood but also throughout adolescence (Losse 1991) and adulthood (Cousins 2003; Missiuna 2008). Their reduced levels of participation in physical activity (Cairney 2005) have secondary consequences, such as reduced cardio-respiratory fitness (Cairney 2006), and increased risk for obesity and coronary vascular disease (Cairney 2007). While the motor difficulties of children with DCD may appear to be less debilitating than those experienced by children with severe physical disabilities (for example, cerebral palsy), it is the high prevalence of DCD, and its impact on children's socio-emotional well-being and future health status, that makes DCD a significant condition in need of appropriate intervention.
DCD is often measured using performance-based and impairment-based motor outcomes. Performance-based outcomes assess general motor ability, which underpins activities of daily living and academic performance. These measures employ neutral tasks which vary slightly from real-life functional tasks to avoid item bias. They are also standardised, objective, and sensitive to change. Some measures of task performance, such as the Canadian Occupational Performance Measure (COPM) or the Goal Attainment Scale (GAS), offer a self report perspective of task-related outcomes and are used to complement objective measures of task performance. Impairment-based measures, an historic way of approaching intervention and assessment, cover the spectrum of WHO categories (Impairment, activity limitations, participation restrictions).
Description of the intervention
Existing interventions range from movement-based therapies and education (usually provided by physiotherapists, occupational therapists, and physical educators) to pharmacology, dietary supplements, and counselling. Traditionally, the movement-based approaches have been classified in accordance with the emphasis of the intervention; that is, task-oriented or process-oriented. Interventions that focus on the performance of specific movement tasks or 'occupations', such as tying shoelaces, ball catching, and handwriting, are collectively called task-oriented approaches. Within the task-oriented approach are task-specific training (Revie 1993), cognitive motor approach (Henderson 1992), cognitive orientation to daily occupational performance (CO-OP) (Missiuna 2001), neuromotor task training (NTT) (Schoemaker 2003), and ecological intervention (Sugden 2007). The common theme of the task-oriented approaches resides in the employment of specific tasks in an attempt to improve corresponding skills. The differences between the task-oriented approaches depend on where the relative emphasis is placed, such as task-specificity in motor skill learning (Revie 1993), the interaction between cognitive, affective, and motor competence (Henderson 2007), child-centred cognitive strategies (Missiuna 2001), analysis of neuromotor processes underlying motor control (Schoemaker 2003), and making the task relevant and ecologically valid (Sugden 2007). In contrast, process-oriented approaches work on the principle that there is an underlying deficit, which must be remediated before functional change can take place. One of the most popular approaches in this category is sensory integration therapy, first devised by Ayres in the 1960s, which aims to improve the effectiveness and efficiency of processing and coordinating sensory information input in order to improve motor performance (Ayres 1979). However, there is more evidence against the effectiveness of this approach than in favour of it (Zimmer 2012). In this review, we will evaluate existing research on the more recently proposed task-oriented approaches in comparison to these other process approaches so that consumers and professionals have the opportunity to make informed decisions.
How the intervention might work
Based on principles of motor control and learning, task-oriented approaches involve concentration on the tasks, or group of tasks, to be mastered. In essence, they capitalise on the assumption that learning and skill acquisition is strongest when the learner understands the meaning of the training, and finds the task to be useful or relevant to his or her life. Thus, aspects of motivation and engagement are catered for, as well as the current understanding around brain plasticity, which supports the idea that learning effectiveness is enhanced when the individual perceives the goal, or likely reward, as functional and beneficial (Hoerzer 2012). At the behavioural level, the intervention effects are explained in terms of the variables involved in motor learning, such as repetition, duration, intensity, frequency of practice, and the types of feedback given (Keogh 1985; Henderson 1992; Revie 1993; Schoemaker 2003). At the cognitive level, the improvement of motor skills is explained in terms of intellectual understanding of motor tasks and verbal mediation, or talking through movements in the process of perceiving stimuli, and preparing and executing movements (Cratty 1989; Henderson 1992; Missiuna 2001). The impact of incorporating ecological aspects involves adapting or manipulating the environment and context to reproduce, as closely as possible, the actual learning task environment. This ensures contextual relevance and meaning, and thus is ecologically valid to the child with the support of significant others such as parents and teachers (Sugden 2007).
Why it is important to do this review
Parents of children with DCD need a readily understandable review to help them make informed decisions about the best available interventions, as do service providers. Since the publication of the recent systematic (Hillier 2007) and meta-analytic reviews (Pless 2000) of the intervention effects for children with DCD, new evidence has accumulated. The latest systematic and meta-analytic reviews (Smits-Engelsman 2013; Wilson 2013) include the recent evidence only, and do not evaluate these data together with older evidence. The meta-analytic studies also considered the intervention effects of the foregoing studies altogether, rather than examining the differential intervention effects of the children's age, the environment of intervention, and interventionist (Hillier 2007). The identification of differential intervention effects would allow service providers and consumers to make more informed decisions. It is of clinical and theoretical interest whether the intervention effects are transferred from the specific intervened tasks to general motor ability.
To assess the effectiveness of task-oriented interventions on movement performance, psychosocial functions, activity, and participation for children with DCD. To examine differential intervention effects as a factor of age, sex, severity of DCD, intervention intensity, and type of intervention.
Criteria for considering studies for this review
Types of studies
Randomised clinical trials and quasi-randomised trials.
Types of participants
Children aged four to 18 years, diagnosed with DCD as defined by the DSM-IV (APA 1994), DSM-5 (APA 2013), and/or children referred to as clumsy, physically awkward, or with dyspraxia who otherwise meet the criteria.
Types of interventions
We will include studies where the intervention is described as task-oriented and formally requires practice of a specific task or occupation as the principal form of intervention. This may include task-specific training, cognitive motor approach, ecological Intervention, neuromotor task training (NTT), and cognitive-orientation to occupational performance (CO-OP). If a trial intervention appears to be task-oriented but is not formally labelled as such, we will still include it if all authors agree.
We will include studies that compare the task-oriented intervention with either (1) an inactive control intervention, such as usual care or a waiting list control, or (2) an active control intervention, for example, a process-oriented approach such as sensory integration therapy (Ayres 1979), pharmacology, counselling, or dietary advice.
Types of outcome measures
We will use both movement performance- and impairment-based measures to examine changes in fine and gross motor function following intervention.
Changes in fine and gross motor function following intervention as measured by standardised performance outcome tests such as the following:
- McCarron Assessment of Neuromuscular Development (MAND) (McCarron 1997)*
Changes in fine and gross motor function following intervention as assessed by the following:
- Measures of impairment (for example, sensation as measured by tests such as stereognosis or pressure detection, muscle strength as measured by tests such as one repetition maximum, or co-ordination as measured by tests such as the Purdue pegboard).
- Adverse effects or events: no studies to date have identified adverse events resulting from task-oriented interventions. By its very nature, everyday tasks are performed under closer than usual scrutiny, supervision, or both, and therefore are assumed to be safer than those encountered in everyday life. However, we will record any reports of adverse events, which conceivably could include musculoskeletal injury, falls, or pain.
- Measures of participation (academic level, sporting participation, recreation), such as LIFE-H (Fougeyrollas 1998), or teacher and family reports of level of participation.
*Outcomes to be included in a 'Summary of findings' table as recommended by the Cochrane Handbook for Systematic Reviews of Interventions (Schünemann 2008).
Search methods for identification of studies
We will search the following electronic databases and trials registers.
- Cochrane Central Register of Controlled Trials (CENTRAL), part of T he Cochrane Library
- Ovid MEDLINE
- ProQuest Dissertations & Theses: A & I
- Science Citation Index - Expanded
- Social Sciences Citation Index
- Conference Proceedings Citation Index - Science
- Conference Proceedings Citation Index - Social Science and Humanities
- ClinicalTrials.gov (clinicaltrials.gov/)
- Current Controlled Trials (controlled-trials.com)
- WHO International Clinical Trials Registry Platform (who.int/ictrp/en/)
- Australian New Zealand Clinical Trials Registry (anzctr.org.au/)
We will use the following sensitive search strategy to search Ovid MEDLINE and adapt it for other databases. We will apply no language or date limits.
1 Motor Skills Disorders/
2 Psychomotor Disorders/
3 ((coordination or co-ordination) adj3 disorder$).tw.
4 (motor skill$ adj3 disorder$).tw.
5 (motor function$ adj3 disorder$).tw.
7 (clumsy or clumsiness).tw.
8 (physical$ adj1 awkward$).tw.
9 (incordination or in-cordination).tw.
10 (motor adj1 (competence or impair$ or difficulty or difficulties or proficiency)).tw.
11 (movement$ adj1 (difficulty or difficulties)).tw.
12 exp Apraxias/
15 Developmental Disabilities/ (14882)
16 (co-ordination or coordination or motor$).tw.
17 15 and 16
18 14 or 17
19 exp child/
21 (child$ or preschool$ or pre-school$ or boy$ or Girl$ or teen$ or adolescen$ or young people$ or youth$).tw.
23 18 and 22
24 randomized controlled trial.pt.
25 controlled clinical trial.pt.
28 drug therapy.fs.
33 exp animals/ not humans.sh.
34 32 not 33
35 23 and 34
Searching other resources
We will distribute an email to members of the International Society for Research in DCD and ask them to provide any unpublished studies (including studies in languages other than English) that meet our inclusion criteria. We will also search the reference lists of relevant papers found by the literature search. We will also search relevant websites identified by international experts, such as advocacy groups or education resource listings, which may have identified unpublished trials.
Data collection and analysis
Selection of studies
Two authors (SH and MM) will independently assess all studies identified by the search strategy for inclusion. Disagreements will be resolved by discussion with the third and the fourth authors (LP, SN). We will perform a first screening by reading the titles and abstracts of the identified studies. We will determine final inclusion by reading the full paper. We will report reasons for exclusion for all studies considered to have closely missed inclusion during the selection process.
Data extraction and management
Two review authors (MM and SN) will independently extract data from the included trials using a standardised data extraction form specifically designed and piloted for this review. We will seek translations where necessary. Extracted data will include the following information from the included studies:
- methods - including aim, design, unit of allocation;
- participants - including inclusion and/or exclusion criteria, number randomised, withdrawals and exclusion, sample characteristics;
- intervention - type of intervention (for example, NTT or CO-OP), mode of delivery (individual or group), personnel (health, education, or non-trained staff), location (clinic, hospital, school, home), duration, frequency, and intensity;
- outcomes - including time points measured, unit of measurement, power;
- other - source of funding, possible conflicts of interest;
- 'Risk of bias' assessment - including details of sequence generation, allocation concealment, blinding, completeness of outcome data, selective outcome reporting;
- data and analysis - including length of follow-up, loss to follow-up, unit of analysis, statistical methods used.
In order to assess the effects of the intervention, we will extract data for outcomes of interest (means and standard deviations for continuous outcomes and number of events for dichotomous outcomes) where available in the published reports. Two review authors (MM and SN) will enter extracted data from each study into RevMan and verify this. Any inconsistencies will be resolved by discussion. A third author (SH) will assist in making a decision, if necessary.
Assessment of risk of bias in included studies
Two authors (MM and SH) will independently assess the risk of bias for each study and overall risk of bias, using The Cochrane Collaboration 'Risk of bias' tool (Higgins 2011a) for randomised controlled trials (RCTs). We will assess the risk of bias for each included study against key criteria: random sequence generation; allocation concealment; blinding of outcomes; incomplete data; and selective outcome reporting. Where possible, we will obtain trial protocols for comparison of planned outcome assessment to the outcome data available from each trial, to enable us to evaluate whether all outcomes assessed in a trial have been reported. We will also attempt to contact study authors if no outcome data relevant to the primary or secondary outcomes of the review have been published for a trial (Kirkham 2010). We will report the full 'Risk of bias' assessment and whether or not it reduces the confidence in the effects being analysed. If the risk of bias of the individual included studies is high, we will interpret the results of the studies with caution. We will explicitly judge each of the listed domains for risk of bias as: low risk of bias; high risk of bias; or unclear risk of bias (either lack of information or uncertainty over the potential for bias). Two independent assessors (LP and SN) will determine overall risk of bias and resolve disagreements by consensus using a third review author if necessary.
Measures of treatment effect
Continuous outcome data
We will obtain standardised mean differences (SMD), more precisely Hedges g (Borenstein 2009), from the primary trial data. SMD is dimensionless so we will be able to combine outcomes from studies which used different measurements. We will calculate SMDs (Hedges g) and associated sampling error variance from the available information in the included studies. Such information includes descriptive statistics (means, standard deviations (SDs), and sample sizes) and test statistics (t, F, P values etc.). In the case of no relevant information being available, we will request information from the authors.
Multiple outcome data
For studies with post-intervention data at multiple time points, we will extract the SMD from both post-intervention and follow-up phases, but we will analyse them separately in meta-analyses. The immediate post-intervention data will be considered primary.
We will calculate odds ratios and convert them to SMDs (Borenstein 2009) so we are able to combine and compare these outcomes with the aforementioned outcomes.
If results cannot be summarised as above, we will report them as 'other data' narratively and we will not include them in the meta-analysis. In all cases, we will perform primary statistical analysis with Review Manager 5.2 (Review Manager 2012).
Unit of analysis issues
For cross-over trials, we will calculate SMDs and associated variances by accounting for a carry-over effect (Higgins 2011b). This can be done by obtaining (inter-individual) correlation coefficients between pre- and post-intervention periods. If such information cannot be gathered from the publications, we will contact the authors.
For cluster-randomised trials, we will appropriately integrate the effect when calculating the SMDs and corresponding variances by obtaining intra-cluster correlation coefficients (from the authors of the studies as necessary).
Dealing with missing data
If we find any missing, inconsistent, or incomplete data (for example, missing outcomes, missing summary data, missing study characteristics), we will contact the correspondence author. We will record all relevant information on missing data and drop-outs for each study as part of our 'Risk of bias' assessment. We will also examine the reasons for missing data and drop-outs, and take those reasons into account when drawing conclusions.
Further, as part of a sensitivity analysis, we will use a multiple imputation technique for missing subgroup information, assuming the data are missing at random (Pigott 2012), although such missing data are not expected.
Assessment of heterogeneity
- MM, SH, and LP (each of whom possesses clinical experience) will assess clinical heterogeneity, evaluating the variability across participants (age, gender, severity of DCD, interventions (frequency, duration, types), and outcomes (types).
- We will assess methodological heterogeneity by evaluating variability in research designs and risk of bias (SN and SH). We will examine any difference in effect size between studies that use adequate randomisation, allocation concealment, and blinding, and the studies that do not perform them adequately. We will also group the reviewed studies into high and low risk of bias groups, and evaluate the difference in effect sizes.
- We will identify statistical heterogeneity by visual inspection of the forest plots, and by using the Chi
2test and I 2statistic. We will use a P value of 0.10 to determine the statistical significance of the Chi 2test for a small sample size (Deeks 2008). We will evaluate the importance of the I 2by an observed I 2value > 40%, the magnitude and direction of effects, and the evidence for heterogeneity from the Chi 2test. To address heterogeneity, we will conduct a random-effects meta-analysis and explore the causes of heterogeneity by conducting subgroup analyses and meta-regression.
Assessment of reporting biases
We will explore whether there are any small study effects and conduct visual assessment of funnel plot asymmetry to identify possible publication bias (Sterne 2008). If funnel asymmetry is due to a lack of data points in the non-statistically significant region of a funnel plot, we will interpret the funnel plot asymmetry as an indicator of possible publication bias; that is, multiple or singular publication of research findings, depending on the nature and direction of the results. We will only create a funnel plot for a meta-analysis that contains at least 10 studies (SMDs). To avoid reporting biases, we will search other sources as described above.
We will use both fixed-effect and random-effects meta-analyses to combine SMDs (Hedges g) and report results from both models using Review Manager 5.2 (Review Manager 2012). This will only be performed if studies are sufficiently similar with regard to clinical heterogeneity and if there are more than two studies available. A main meta-analysis will combine all data points (SMDs) except for data points from follow-up periods. We will conduct subsequent meta-analyses on subgroups according to the criteria described in the next section. Also, we will provide narrative descriptions of those studies that are unsuitable for meta-analysis.
Subgroup analysis and investigation of heterogeneity
If we find sufficient studies, we will perform the following subgroup analyses.
- Age (preschool versus junior; primary versus senior; primary versus secondary or high school).
- Sex (male versus female).
- Severity of DCD in terms of cutoffs used for standard performance outcome tests and questionnaires (for example, second percentile; fifth percentile; 15th percentile).
- Intervention intensity calculated as a combination of frequency and duration (for example, < 3 times a week versus ≧ 3 times a week; or < 6 weeks versus ≧ 6 weeks).
- Type of intervention (for example, NTT versus CO-OP).
We also plan sensitivity analyses to explore the impact of study quality (risk of bias), for example, for studies with a high risk of:
- assessment bias (associated with issues of blinding);
- attrition bias (associated with completeness of data);
- selection bias (associated with sequence generation or allocation concealment).
The University of Otago and the University of South Australia: the development and publication of this protocol and forthcoming review is made possible thanks to a salary from each author's university.
Cochrane Developmental, Psychosocial and Learning Problems Group: provided advice and assistance in producing this protocol.
Richard German (Reference Librarian), Health Sciences Library (Medical and Dental), University of Otago: assisted with searching the databases.
Contributions of authors
MM drafted the protocol and all authors contributed advice. MM will perform the initial search, and LP and SH will perform second review author roles with MM from inclusion through to risk of bias and data extraction stages. SN will provide statistical advice. MM, SH, and LP will write the final review text.
Declarations of interest
Motohide Miyahara - none known.
Susan L Hillier - none known.
Liz Pridham - none known.
Shinichi Nakagawa - none known.
Sources of support
- University of Otago, New Zealand.In the form of a salary
- University of South Australia, Australia.In the form of a salary
- No sources of support supplied