Efficacy of Interventions to Improve Motor Development in Young Children: A Systematic Review

Identification of Studies

A literature search of interventions to improve motor development in preschool children (mean age: <5 years) was conducted of 14 electronic databases (A+ Education, ERIC [Ovid], Medline [Ovid], Health Reference Centre Academic, Sports Discuss, Current Contents Connect, PubMed, ISI Web of Science, CINAHL [Cumulative Index to Nursing and Allied Health Literature], PsychInfo, Australian Family and Society Plus, Expanded Academic ASAP, Dissertation Abstracts, and Ebsco Megafile Premier) from January 1987 to June 2007. The search strategy focused on 3 elements: motor development (key words: play, coordination, motor development, motor skills), age (key words: toddler, preschool), and intervention (key words: program, trial, intervention). There was also a citation search of included articles and relevant published reviews.

Criteria for Inclusion/Exclusion

Articles were included when they were reported on an intervention that targeted children with a mean age of <5 years and measured motor development objectively (eg, fundamental movement skill proficiency). Articles were excluded when they were published in a language other than English or published before. 1987; when the intervention targeted children with a specific disease or health problem; when the full article was not available; or when the intervention was <4 weeks, had <10 participants, had no control group, or reported only qualitative data.

Articles were initially excluded on screening of title and abstract. When appropriateness of the article could not be determined, the full article was obtained. Three reviewers independently evaluated, using a standardized checklist, full text copies of all obtained articles to determine whether they met the inclusion criteria. When opinions differed, a consensus was reached through discussion.

Assessment of Methodologic Quality

Included articles were assessed for methodologic quality by using a 10-item quality assessment scale¹⁹ (Table 1). For each article, 2 reviewers independently assessed whether the article scored positively or negatively for each item. When an item was insufficiently described, it was allocated a negative score. Agreement between reviewers for each article was set a priori at 80%^19,20; that is, for each article, reviewers were required to agree that the items were either present or absent for 8 of 10 items. In the case of <80% agreement, consensus was reached by additional discussion. The scores were then summed for each article to determine the overall quality of the article. An article was deemed to have high methodologic quality when it scored ≥5 for a controlled trial and ≥6 for a randomized, controlled trial.¹⁹

Comparing Results

To facilitate comparison between studies, we extracted the following data from each article: (1) design, including randomization procedures and settings; (2) methodologic quality (described already); (3) intervention components, including sample size, length of the intervention, hours of instruction, program deliverer, and parental involvement; (4) effectiveness of the intervention immediately after intervention and at subsequent follow-up (eg, 12 months); and (5) alignment with the Consolidated Standard of Reporting Trials (CONSORT) and Transparent Reporting of Evaluation with Nonrandomized Designs (TREND) statements.^21,22

This review includes interventions that report motor skill development by using different instruments, thereby allowing more interventions to be included; however, this does not make it possible to pool the results. Rather than using a set change standard for assessing whether an intervention was or was not efficacious, each intervention's efficacy was assessed on the basis of its own reported results. An intervention was deemed efficacious when the intervention group showed a statistically significant improvement in comparison with the control group.

Published Articles

Table 2 summarizes the information in this section.

Unpublished Articles

Table 3 summarizes the information in this section.

Design

In both published and unpublished studies, there were more controlled trials than randomized, controlled trials; however, controlled trials were substantially more prominent in the unpublished studies. The fact that all unpublished studies were retrieved as postgraduate theses, which are typically restricted by budget and time, may explain the high percentage of controlled trials. Although this is understandable, randomization is no doubt important, because it minimizes bias during and after assignment of treatments and also prevents confounding biases.²⁰

Overall, 65% of studies were implemented in a child care setting (nurseries, elementary schools, and preschools), 60% of which were effective. In older children, physical activity interventions delivered in a variety of settings have generally been effective^18,40; however, in young children, we hypothesize that the setting in which the intervention is delivered may be highly influential in having an impact on outcomes.

Methodologic Quality

Of the 17 studies reviewed, <20% had high methodologic quality. This is concerning considering that only 5 of the 10 items for controlled trials and 6 of the 10 items for the randomized trials were needed for a study to be categorized as having high methodologic quality. Furthermore, although the criteria used to assess methodologic quality have been used in several other large systematic reviews^18,41 and are considered a robust checklist, they are not exhaustive. For example, it is unknown whether these studies were adequately powered for statistical analysis. If additional items had been included, then the methodologic quality may have been even lower. Seven of the 10 items used to assess methodologic quality were consistently missing. We suggest that results that are reported from methodologically poor studies be interpreted with caution, because potentially fatal flaws within a study method may mask, overinflate, or bias findings; however, we acknowledge that the criteria used to measure quality are extremely specific. For example, 1 point was awarded when the study compared groups at baseline on 3 key characteristics (age, gender, and at least 1 relevant outcome). When only 2 of these were present, a point was not awarded. Despite the specificity of the methodologic quality criteria, we suggest that it is a helpful tool to ensure transparent reporting.

Intervention Components

Intervention components varied considerably for all studies. Sample sizes were more diverse for published studies (range: 24–545), compared with unpublished studies (range: 19–85). Recruiting whole groups or child care centers may be a beneficial mechanism of maximizing sample size to allow appropriate statistical analyses.

Studies were on average 12 weeks in duration and incorporated ∼1 hour of instruction per week. For older children, longer and more intense physical activity interventions have been recommended.^42,43 This may also be true for interventions that target young children; however, their reduced cognitive capacity and shorter concentration levels need to be considered. Although it is not known whether longer and more intense physical activity interventions that target young children are beneficial, it is clear that longer follow-up periods are advisable.^18,39

Setting staff (eg, teachers) delivered the intervention in only 5 of 17 studies reviewed. There are several advantages to having setting staff deliver interventions: (1) maximizing potential sustainability of the program; (2) minimizing cost associated with implementation; and (3) enhancing participation levels. As suggested already, young children are more likely to be physically active when they have positive interactions with their teacher and peers and are in a better resourced play environment.⁴⁴ Although these are advantages, it is important to note that successful implementation requires substantial additional responsibilities for staff, adding to their already crowded work schedules. Furthermore, we suggest, on the basis of our own research, that staff may lack confidence in delivering physical activity interventions, particularly those that focus on motor development and may require substantial training and guidance.⁴⁵ Using a combined or partnership approach, which has been shown to be beneficial in interventions that target older children,⁴⁶ may be helpful. This approach, which involves researchers and setting staff working together, distributes responsibility between key stakeholders, minimizes burden, and offers a forum of support for teachers who may not be confident in delivering motor development physical activity interventions.

Current recommendations for children who are aged 0 to 5 years emphasize the critical role that parents play in providing opportunities, encouragement, and support for physical activity.^14,15,47,48 Despite this, only 3 studies—<20%—involved parents. Although involving parents in interventions is logistically difficult and often complex, it is imperative that they be involved to some extent, even if it is indirectly as in the study by Reilly et al.²⁴ Using the home environment to support what happens in the intervention setting is essential to ensuring that instructional messages received are consistent and reinforced, maximizing the sustainability of changes in outcomes. This notion has been supported in other recent reviews that focused on physical activity among school children and adolescents.^18,39,49

Efficacy

More than half of the interventions were efficacious in significantly improving motor skills. Of the efficacious interventions, most were longer than 8 weeks,* and the intervention was implemented ≥ times per week.^{23,25,27,31,32,34,35,37} In addition, 5 of the successful studies were implemented by teachers and/or with the assistance of the researcher^{23–25,27,29}; however, it is difficult to identify specific intervention components that contributed to efficacy, because each study was uniquely distinct in design, length, instructional time, and participants (eg, some studies included disadvantaged children or those at risk for disadvantage²⁶). Efficacy may have been affected by a number of other factors. First, although most studies used a validated measure of motor development, there were variations in the number of skills assessed and reported. Some studies assessed <3 skills^32,35,38; others assessed and reported only object-control skills (eg, strike, kick, catch, throw, bounce, roll),^30,34 whereas others assessed both object-control and locomotor skills.^26,27 Second, the poor methodologic quality of most studies possibly biased the results away from the null. For example, it was difficult to ascertain who assessed the skills and whether assessors were blind to group allocation. Assessors who were not blind to group allocation may have biased results. Third, the demographics of the facilitator (eg, gender, experience, competence and confidence, training) may have influenced how the intervention was implemented and how participants received the intervention. These details were absent for almost all of the studies. Finally, the differing underpinning theoretical framework may have also affected the outcomes. The theoretical framework was reported only for 3 studies,^24–26 and each of these reported a different framework.

Alignment With the CONSORT and TREND Statements

The CONSORT²¹ and TREND²² statements provide guidelines for transparent reporting of randomized, controlled trials and controlled trials, respectively. They have been developed to ensure accurate reporting of external and internal validity intervention components and to facilitate comparisons between studies. The use of such statements in reporting data has improved the quality of studies in the past several years²²; however, as evidenced by this review, in which only 1 article was framed by using the CONSORT statement and none by using the TREND statement, a large number of articles are not guided by these statements. Most studies provided some information about participants, the interventions, and some aspects of participant flow, including attendance (as a measure of intervention exposure) and retention rates; however, several other important CONSORT and TREND items were noticeably missing, such as blinding, randomization procedure, and baseline data. It is extremely difficult to determine the true treatment effect and compare and contrast findings when the CONSORT and TREND statements are not followed.

Strengths and Limitations

There are 3 main strengths to this review. First, published and unpublished studies were retrieved for a 20-year period. Second, to allow comparison between studies, we extracted extensive detailed information from each article. Finally, the inclusion criteria were kept broad, which allowed for the inclusion of studies from several countries with varying designs.

Several limitations existed in the studies reviewed. Only 2 (12%) recorded the height and weight of the participants, making it impossible to determine whether these results are confounded by growth.^24,28 Quality control measures were reported in only 2 of the studies,^24,28 meaning that we were unable to determine whether the delivery of the interventions was consistent throughout the individual studies. Finally, although we were able to compare the studies broadly, we were not able to determine whether the motor development component was differentially affected by the interventions. Direct comparison between quantitative data was not possible because a diverse range (eg, >10) of motor development instruments were used. Moreover, those that did use similar instruments reported different outcomes (eg, some reported both locomotor and object control scores and/or only locomotor and object control scores).

Recommendations for Practice and Future Research

The development of motor skills among children may be an important predictor of later physical activity levels.^8,50 School children with high levels of motor development are more likely to perform better on a number of physical and psychological health outcomes. Children have the potential to be proficient in most fundamental movement skills by approximately the age of 6 when they are given the encouragement and opportunity to learn and practice⁵¹; therefore, more high-quality interventions that target motor development are critical for enhancing the habitual physical activity of children. In light of this, several recommendations can be made from this review:

1. Physical activity motor skill interventions should be implemented through a combined or partnership approach, thereby sharing the responsibilities and burden between researchers and setting staff.

2. Program facilitators should be chosen carefully because their experience, competence, and confidence are likely to influence children's participation and enthusiasm for the program. Similarly, the program setting should be considered before implementation to ensure that children will be comfortable in their surroundings and with the equipment provided.

3. Parents should be involved, either directly or indirectly, in motor development interventions to complement and support the translation of knowledge and skills from the intervention setting to the home environment. Parental involvement is a key element for habitual and lifelong motor skill development.

4. Physical activity motor skill programs should be underpinned by a sound theoretical framework.

5. Sample size calculations should be completed before recruitment to ensure that appropriate statistical analysis can be conducted. Whole groups or child care centers may need to be recruited to maximize sample size. If whole groups or child care centers are recruited, then a cluster randomized, controlled trial would be required.

6. Physical activity intervention that focuses on motor skill outcomes should be methodologically sound and follow guidelines detailed in the CONSORT²¹ or TREND²² statement, which will ensure transparent reporting. Greater attention should be given to longer interventions and follow-up periods, randomization procedures, using assessors who are blind to group allocation, using validated measures of motor development, and comparing baseline characteristics.