Design
This study has previously been described in detail elsewhere (22). It followed the recommendations of the CONSORT statement and was registered with ClinicalTrials.gov (registration number NCT02954614). A matched-pair cluster-randomized controlled design was applied, with ASPs as clusters and first grade pupils as participants. The ASPs in the intervention group received the Active Play in ASP intervention, while the ASPs in the control group were given no follow-up. The ASPs randomized to control were invited to receive the intervention after the study was completed.
Recruitment and sample
The first step of the study recruitment process was to engage school physiotherapists (PTs) because the study relied on their assistance in the implementation of the intervention as well as in the data collection process. School health services in municipalities of three counties in Eastern Norway were approached and, within the time limit defined for this first phase of recruitment (August 2016), PTs from 14 municipalities volunteered to participate. They assisted in recruiting the ASPs in schools within their area of responsibility. All schools were eligible. School administrators, who accepted the invitation, provided written consent. The ASPs were paired based on available background information about the size (small, medium, large) and location (urban versus rural), as we assumed that the number of pupils, as well as space and access to nature areas, might have an impact on children’s PA level. The names of the ASPs were put in sealed envelopes and randomly allocated to either the intervention or control group. The person revealing the allocation was not involved in the study (22). Following the allocation, all parents of first graders (5–6 years of age) in the participating ASPs were asked to provide written consent on behalf of their child. There were no exclusion criteria. First graders were chosen as study subjects because a great majority of them attend ASPs in Norway (14). Moreover, in comparison to older children, there is less information on PA in this age group. Figure 1 shows the flow of the participants in the study. Sample size calculation was performed based on previous studies and on the pilot study, and we assumed that a 6-minute increase in MVPA during ASP opening hours in the intervention compared to the control (common SD=12) would be of clinical importance (18, 19, 22), which represents 10% of the 60 MVPA minutes recommended by the guidelines. We estimated N to be 121 in each group, without accounting for cluster effects, and planned to enroll a minimum of 200 children in each group to secure sufficient power for additional analysis at the cluster level. With a minimum of 25 first graders estimated in each ASP, a maximum of 16 ASPs would have to be included (22).
The intervention
The Active Play in ASP intervention consisted of a 7-month course program (October–May) directed at ASP staff. The intention was to enhance their knowledge of and skills in creating a PA-supportive environment by accommodating and gently encouraging activities instead of directing them in a controlling manner. The intervention was informed by experiences from practice-, theory-, and evidence-based knowledge. It was developed in a stepwise process, together with representatives from practice, and was built on principles of the self-determination theory (SDT) (22, 23). In line with the SDT principles, the intervention focused on the importance of retaining the self-chosen and child-managed character of play with an emphasis on supporting the children’s needs for competence, relatedness, and autonomy (24). Prior to starting in the ASPs, the local PTs in the intervention group were provided with an 8-hour course that included a presentation of the intervention study with an emphasis on their role and responsibility. They were also provided with intervention material (workbook). The intervention course program for the ASP staff consisted of two introductory sessions led by two of the researchers, KL (PhD and PE teacher) and KR (PhD and physiotherapist) focusing on the children’s physical activity play, the importance of friends, and the activity place. Furthermore, the sessions emphasized the ASP staff’s interaction styles, the children’s motivation for PA, and the ways in which the staff could support the children’s engagement in PA. The sessions took place in each participating intervention ASP and included short lectures, theme-based discussions, and group tasks. The staff were encouraged to provide examples from their own practical experience. Following the introductory sessions, one of the researchers met with the staff to outline how their ASP would combine new knowledge and previous experience into strategies for supporting children’s physical activity play during the time they spend at that ASP. The local PT participated in the introductory sessions and the meeting with researchers. Additionally, the local PT and an ASP representative were responsible for mapping the activity equipment as well as the indoor and outdoor facilities available at each ASP. This information provided an important supplement to the discussions that followed. The program continued during the school year, with monthly meetings for the ASP staff led by the local PT during which they worked together on predefined tasks related to physical activity play (Table 1).
Table 1 Intervention components and content
|
|
Component
|
Content
|
Introductory course for school physiotherapists
|
|
1-day course
|
Information about the intervention and the physiotherapists’ role and responsibilities. Presentation of intervention workbook.
|
Course program for ASP staff
|
|
3-hour session
|
Introduction to research-based knowledge on children's physical activity in play. Increasing the staff's awareness of how such play can be influenced and supported in ASPs
|
|
3-hour session
|
Basic theoretical principles of SDT applied to physical activity and physical activity play among children; how to be activity supportive.
|
|
Mapping
|
Thorough mapping of the ASP equipment and facilities.
|
|
Planning (1-2 hour meeting)
|
Summary of introductory sessions; how to make use of new knowledge.
|
|
5 meetings (monthly 1-2 hours) led by the local school physiotherapist
|
Discussions and practical tasks focusing on:
- Motor learning in children,
- Equipment and environment,
- Mapping of staff competencies,
- Inclusion/exclusion in play,
- Leading and supporting activity in groups.
|
The table has previously been published in the protocol article (22). No permissions were required to include it here.
Outcome measures
The data were collected at 3 time points—in September/October 2016 (before the beginning of the intervention; baseline, T0), 7 months later (immediately post intervention; follow-up 1, T1), and one year post intervention (follow-up 2; T2). All outcomes pertained to individual participants. The primary outcome was MVPA and secondary outcomes were VPA, LPA, and sedentary behavior. PA was measured during the time spent in an ASP during its opening hours for each ASP on weekdays over a period of one week. Objective PA assessments were obtained using the ActiGraph GT3X (ActigraphTM LLC, Pensacola, US). The ASP staff administered the children’s accelerometers following a standard procedure. The devices were fitted to each child’s hip at the time of arrival and removed before they left. Valid accelerometer data files consisted of at least 2 days of minimum 60 minutes of accelerometer wear time. All data registered before or after each ASP’s opening hours were regarded as non-wear time and were excluded from the analysis. Non-wear time was defined as 20 minutes of consecutive zeros, allowing for 2 minutes above zero. ActiLife version 6.13.3 was used to download raw accelerometer files and to reintegrate them in 10 second epochs to detect the intermittent and sporadic activity patterns of children. Subsequently, the files were loaded into Kinesoft v3.3.80 (Kinesoft, SK) for screening and classification of intensity level. The number of minutes spent in MVPA was estimated using cut points, as described and recommended in previous studies (25, 26). MVPA was defined as 2296–4011 counts per minute (CPM), while >4011 CPM was defined as VPA, 101–2295 CPM as LPA, and <101 CPM as sedentary behavior (25).
The local PTs measured the participating children’s height and weight using a stadiometer and a digital scale at the local school nurses’ office. The children wore light clothes and no shoes. The Body Mass Index (BMI = kg/m²) was calculated and converted into age and gender specific scores (27). The children’s gender was reported by the parents at the beginning of the study.
The data on average weekly outdoor temperature (degrees Celsius) were calculated for each school at baseline, follow-up 1, and follow-up 2, based on the daily mean temperatures that were estimated using the data recorded at local weather stations closest to each ASP during the ASP opening hours (at 14:00 PM). The average rainfall (in mm) was also documented for each ASP (between 12:00 PM and 17:00 PM).
Statistical analyses
Descriptive statistics were generated for the gender and BMI category variables for each group. The analyses of intervention effects were based on the between-group differences in MVPA (primary outcome), VPA, LPA and sedentary behavior (secondary outcomes) changes between T0, T1, and T2, respectively. There was a large variance in the participants’ average daily wear time at each measurement point. Thus, wear time would influence the MVPA minutes accumulated per day. To account for the variance in daily wear time, MVPA, VPA, LPA, and sedentary behavior were expressed as minutes/hour.
To account for clustering, we included random effects for schools and for pupils nested within schools. The analyses were conducted using mixed-effects models, and we employed the longitudinal data analysis (LDA) method to adjust for baseline value (28). In line with this method we included treatment, time (categorical), and their two-way interaction term as fixed factors in the main effects analyses of the pre-specified analysis. We controlled for the daily mean outdoor temperature at each school. We used an unstructured covariance pattern for level 1 residuals (or autoregressive patter if non-convergence), and an identity covariance structure for random effects (28).
Tests of effect modification were not specified in the protocol, and thus treated as exploratory (29). In separate models, we investigated gender, baseline BMI and baseline accelerometer counts per minute (CPM) as potential effect modifiers at each time point through inclusion in the three-way interaction terms (treatment x time x “effect modifier”) (30). We explored differential intervention effects for children with high vs. low baseline BMI using a binary measure; For the analyses, children with overweight and obesity (age- and gender adjusted BMI ≥ 25) were collapsed into one category. We used a binary measure of baseline CPM (CPM < median = 50% least physically active and CPM ≥ median = 50% most physically active) to explore the differential intervention effects for children with high vs. low levels of baseline physical activity.
As specified (22) the main analyses were conducted using an alpha level of 0.01, while in the exploratory analyses the alpha level was set at 0.05. Statistical analyses were performed using STATA 15.0 (31). The LDA method used all available participant data and were based on the missing at random (MAR) assumption (28, 32). We conducted multiple imputation by chained equations (MICE) as a sensitivity analysis (33). In addition to the variables in the analytic models, the auxiliary variables included in the imputation model were baseline BMI (continuous) and gender. In order to take into account the three-level structure of the data in the imputation model, we used the “ml.lmer” command from the R package “miceadds” R. We generated 25 imputed datasets and used the “lmer” command from the “lme4” R package to analyze them. The imputation procedure was performed using the RStudio version 1.1.423.
Ethical considerations
The study was reviewed and approved by the Data Protection Official for Research to ensure that the project was conducted in accordance with the Personal Data Act and the Personal Health Data Filing System Act (reference number 46008). Informed written consent to participate was obtained from the parents on behalf of their children. Adjusted oral information was given to the participating children at each measurement point.