Move, Play, Learn! Creating Active Classrooms in Early Care and Education Centers

Overview

Increasing physical activity (PA) and limiting sedentary time are important for many aspects of a young child's wellbeing. However, young children's PA is often limited, and instead preschoolers (3-5 year olds) spend a large portion of their day inactive. Given that more than 7 million U.S. children younger than 5 attend center-based child care, the early care and education (ECE) setting is an important source of PA for young children. Thus, PA promotion efforts in ECE programs are crucial to increasing the activity levels of young children, although few interventions exist and most have limited success. Classroom teachers are important gatekeepers to physical activity in ECE classrooms, as they can determine how active children are in their care. However, teachers often hesitate to implement physical activity in their classrooms. Teachers' attitudes about PA, confidence in modifying children's PA, and their own physical abilities can influence the amount of activity they provide and how they interact with children to support PA. Few models exist to enhance teachers' skill in promoting children's activity, although such an approach could result in more sustainable outcomes. Innovative approaches to physical activity promotion are needed, which engage teachers through training, ongoing technical assistance, and easy-to-use resources. This project aims to improve children's physical activity through the Move, Play, Learn! intervention, an intervention designed to alter the behavior of classroom teachers to increase PA in children (3-5 years old) enrolled in ECE programs through a novel, behavioral coaching approach. To evaluate the 10-week intervention, 32 classrooms in ECE centers will be recruited and randomly assigned to the Move, Play, Learn! intervention or a waitlist control. Teachers in the intervention arm will attend training workshops to learn how children's activity can be increased using natural opportunities across the child care day and how teacher engagement with children can impact activity outcomes. Teachers will be asked to modify typical classroom activities to be more active, using strategies and resources (e.g., how-to videos, pocket activity cards) and will be asked to modify their interactions with children to support PA. Goal setting, self-monitoring, and tailored feedback will facilitate behavior change.

Study Overview The intervention program will be delivered as a randomized control study, with half of the 32 participants Move, Play, Learn intervention immediately and the other half receiving the intervention in a waitlist control format. Participants will begin the study first with baseline surveys and 5 days of accelerometry to assess current physical activity practices within participating classrooms. Then, if participants are randomized to the intervention arm, participants will proceed with the intervention as described elsewhere. If randomized to the waitlist control, participants will be encouraged to continue with normal classroom activities. At the end of the intervention, teachers will complete the same surveys and accelerometry at baseline as well as a participant satisfaction survey. This will end the teachers' participation in the project. At baseline, parents will be asked to provide demographic information about participating children. Children will be asked to wear accelerometers on the same 5 days as the teacher during baseline and follow-up periods. This data collection will be done during a school week (Monday-Friday), when children are attending center-based case. Recruitment Recruitment efforts will be targeted toward licensed child care centers in Orange, Chatham, Durham, Alamance, Wake, and Guilford Counties. A list of such facilities is publicly available through an online database maintained by the North Carolina (NC) Division of Child Development and Early Education's (http://ncchildcaresearch.dhhs.state.nc.us/search.asp). Study staff will contact directors at these child care centers and ask for their assistance distributing study flyers to their staff. Flyers will instruct interested providers and parents to contact the study office. Randomization Centers will be randomized in two sets, once halfway through baseline data collection and the other half at the end of baseline data collection. At these randomization time points, the Principal Investigator (PI) will generate a randomization table using a permuted block approach with block sizes of 2 to ensure balance between study arms at any given enrollment time point. The randomization table will be in excel format and include sequential numbers and randomization assignment (intervention or control). The PI will apply the randomization table to the list of centers to determine study arm assignment. The PI will notify centers of their assignment into either the intervention or control arm. Centers in the intervention arm will have immediate access to Move, Play, Learn! Intervention while the control arm will have delayed access to the program; after follow-up measures are collected. Power Calculation Investigators will compare the changes in child physical activity using the baseline and post-intervention accelerometer measures among children in classrooms randomized to Move, Play, Learn! intervention compared to children in classrooms randomized to the waitlist control group. The sample size (n=26) is sufficient to measure modest increases in physical activity; although more is expected, by measuring activity levels of 7 children per center, investigators can detect even a 7% increase (approximately 12 minutes per day) in non-sedentary physical activity, to put the effect size in terms of a measurable outcome. Based on previous work and other similar studies, investigators estimate a 0.55 effect size in the Move, Play, Learn! intervention. Because children are clustered within ECE classrooms, this needs to be accounted in the power calculation. Based on previous studies, the ECE center intra-class coefficient (ICC) estimate is 0.12 for nonsedentary time. This ICC is appropriate to use for the power calculation since only 1 classroom per center will be recruited; thus the center and classroom are interchangeable here. It is anticipated that an average of 7 children per classroom will be recruited. Using a two-sided test of significance at α=0.05, an ICC=0.12, and cluster size=7 children per classroom, a sample size of 13 centers per arm (91 children per arm) will provide at least 80% power to detect the anticipated change in physical activity (effect size=0.55. Accounting for 15% attrition of classroom teachers, investigators will recruit and randomize 32 ECE classrooms (16 per arm) with a minimum of 7 children per classroom for a total of 224 children. Statistical Analysis The primary analysis will test the hypothesis under the intent-to-treat principle (ITT) using longitudinal, mixed effects models that will account for the correlation of child physical activity by ECE center. As shown below, each model will include a random intercept (b0) and fixed effects for the intervention (B2) and baseline (B1) to test if the mean change in physical activity is different between the two arms. Child PA10 wk = β0 + β1Child PAbaseline + β2Intervention + β3Center + b0 + e Additional analyses will 1) adjust for baseline covariates considered to be relevant to change in physical activity; 2) adjust for baseline variables distributed differently between study arms; 3) test interaction terms between treatment group and other covariates; and 4) examine completers only. Missing data The study team will make every effort to minimize missing data; however, some missing data are expected to occur. Investigators will examine the potential for bias from non-ignorable, non-response using two methods. First, investigators will compare respondents and non-respondents to see if there are systematic differences in non-missing variables. Next, investigators will assess the type of missingness, whether missing completely at random (MCAR) or missing at random (MAR) or missing not at random (MNAR). If data are either MAR or MNAR investigators will employ multiple imputation to impute missing PA values. Sensitivity analyses will be conducted to determine the effect of these missing data on intervention effect estimates and robustness of results using different estimation methods.