To our knowledge, this is the first review that has attempted to identify potential strategies for increasing PA at each level of influence of the socio-ecological model in the case of childcare children (2-6-year-old). Despite the increasing interest in the promotion of PA among childcare children, the number of published studies remains low. Hence the inclusion of lower as well as high quality studies in this review enabled us to view things more comprehensively than would have been possible had the analysis been restricted to RCTs. However, the findings from the interventions were mixed and the level of evidence inconclusive.
Among the studies reviewed here, in addition to structured PA, the use of PA-specific in-service teacher training as intervention strategy was potentially fruitful [see Table 2]. Moreover, teachers’ experience and personal characteristics may play an important role in increasing PA among childcare children [34, 44, 57, 58]. However, due to the lack of mediating analysis to assess possible causal pathways between these strategies and increased PA in childcare children [30, 33, 34, 36], this task remains for future studies.
According to the socio-ecological model, children interact with others in their most immediate learning and development environment. This approach identifies the family as the most influential and proximal system. It also recognizes the importance of partnership between families and childcare [20, 59]. Together, parents and teachers have the best knowledge of the barriers children encounter when engaging in their routine daily PA and the potential that exists for PA in both the home and daycare contexts [60, 61]. Given the positive associations found earlier between parental support and children’s PA , parental support was expected to be a potential strategy. However, based on the present review, the influence of parents on their childcare children’s PA remains unclear. In the eight studies with a parental component, only one high quality intervention succeeded in significantly increasing PA. It may be that families need to be more strongly committed to the intervention, and that merely giving parents knowledge or materials is not enough in a center-based intervention .
On the other hand, the non-significant results of the studies on parental influence on children’s PA included in this review should be interpreted with caution. More methodologically sound research is needed in this area. Two studies by Fitzgibbons et al. (2005; 2006) used parent-proxies, which could have masked significant increases in PA [41, 42]. The study by Reilly et al.  was methodologically promising, but only potentially effective. The authors suggested that the home component was not intensive enough . In a childcare pilot, with rather intensive parent involvement (parent tip-sheets twice a week), parent involvement and school-parent communication during the intervention were found to be very important. Unfortunately, the study lacked the power to detect possible increases in children’s PA. Also fewer PA opportunities were scheduled postintervention than preintervention revealing that the study was not implemented as planned . Some of the studies with a home component limited their assessment to childcare attendance only [45, 48], omitting the possible impact of parental social support , and leaving a question mark over activity levels during time outside childcare.
In this review, intervention strategies extending to the community and policy levels of influence are considered to be large scale, as they deliver and assess interventions broadly and are also highly visible, reaching larger numbers of people (see Figure 1, ). In the present review, interventions of this kind were lacking. However, from a socio-ecological perspective and what can be inferred from the studies included in this review, future research might usefully focus on the upper levels of the model in seeking ways to lower the barriers to increased child engagement in PA. We could make an effort to influence the overall culture of childcare centers, and especially the status of PA in them. Teachers’ cultural beliefs about play and learning are translated into actions which, in turn, influence children’s play behavior (i.e. PA) . Copeland et al.  concluded that policies of childcare concerning children’s safety and school readiness may hinder children’s physical development . The question is could we enhance children’s PA by changing environmental policies without jeopardizing their safety and school readiness? Intervention at the community level, while requiring a lot of resources, may eventually prove a sustainable and a cost-effective strategy [64, 65]. Maintaining changes in health behavior is important, and to attain this goal calls, in particular, for long-term and post-intervention studies. Based on the lack of robust studies and conflicting results to date, further exploration of community interventions is warranted.
In this review only four studies had a follow-up of 6 months or longer [31, 41, 42, 47]. More long-term evaluations are needed. Despite being feasible and highly acceptable to both teachers and children, teachers may lose the motivation to continue with a program postintervention . Challenges reported by teachers in incorporating PA into the childcare curriculum included the weather, which could limit outdoor time and opportunities for active play, and lack of a designated gym area, which could restrict indoor physical activities [48, 49]. To be effective, intervention programs may need modification, and hence teachers should learn to customize the activity patterns of the program to fit their particular curriculum and physical environment . A solution for the resource problem could be found in integrated PA programs [37, 45], which have successfully been tested, although their long-term effects, e.g. on academic performance, have yet to be clarified.
The review findings also highlight other gaps in knowledge. First, only two high quality theory-based multilevel and multi-componential studies were included, and in both studies PA was not a primary outcome [34, 35]. Based on this review, the evidence is consistent with either an increase or no increase in PA as a result of the theory-based multilevel intervention. Overall, in this review theoretical-based studies were not more effective than non-theoretical studies. Although non-theoretical interventions were relatively more effective, it should be noted that in most of the theory-based studies PA was not a primary outcome. Also the extent to which the theory cited was in fact used in the intervention is unclear . Consequently, it is not possible to take a position on their superiority or inferiority compared to individual-level interventions or non-theoretical-based studies. More robust multilevel intervention research, which operates also at the community level, is needed before this can be attempted [67, 68].
Second, even where the intervention studies reported significant increases in PA levels, the results were nevertheless modest and the children’s post-intervention activity levels remained below the current PA recommendations [1–4]. However, it should be remembered that while recommendations are based on the best currently available research evidence, the optimal amounts and intensity levels of PA for children’s healthy growth and development remain unclear . Secondly, half of the studies included in this review only measured children’s PA during childcare attendance or recess, which fails to take into account possible PA outside of childcare day.
Third, in childcare settings, studies have focused on exploring outdoor playgrounds to the neglect of indoor facilities, despite the fact that children spend a considerable amount of time indoors during childcare day . That there is a role for teachers in enhancing children’s PA seems obvious, but none of the reviewed studies looked at peer influence on children’s PA; this should be examined in the future.
Strengths and limitations of the study
This study has limitations which must be taken into account when interpreting the results. First, the review process revealed that although interest in promoting children’s PA has increased in recent years (48% articles of the reviewed articles were published after 2011), research in this area remains scarce. Second, it is possible that potential articles were missed due to the search strategies and criteria used; only studies published in English were included, leaving potential studies written in other languages out of account. Third, the fact that measuring children’s PA is a complex task , may have also affected study results and hence also the results of the present review. Future studies need to utilize valid and reliable methods of measuring PA . Fourth, it should be noted that several studies in this review focused on subsamples of childcare children (i.e. low socioecomic or migrant background or specific ethnicity) [30, 35, 37, 39, 41, 42, 46]. When the study sample represents only a minority of the population generalizing their results to other groups must be done with caution.
Finally, we found that identifying studies with high methodological quality was challenging because of deficiencies in the reporting of studies. Concealment and blinding were reported in only a few articles [30, 34, 40, 43, 46]. Intention to treat was mentioned clearly in only three studies [34, 41, 43], and exposure to the intervention often remained unclear. However, in most studies the outcome measurements used and intervention itself were clearly described.
A major strength of our review was that the databases of several different disciplines – exercise science, general, psychology and pedagogy – were searched, enabling a wider range of intervention types to be found. Second, an advantage of present review compared to reviews that have examined factors associated to PA was that all the included studies were longitudinal and focused on a narrow target population. Children of childcare-age are in a developmentally different stage than older children. In fact, only a few PA-associated variables have been found to be the same in children and adolescents ; thus, childcare-age children may respond differently, for example, from primary school children to an intervention . Third, two researchers independently reviewed the article titles and abstracts to identify potentially relevant studies, and then assessed the quality of those included in the review.