The Fueling Learning Through Exercise (FLEX) Study Cluster RCT: Impact on Children’s Moderate-to-Vigorous Physical Activity

Abstract

Introduction:

Most children do not meet recommendations for school-time and daily moderate-to-vigorous physical activity (MVPA), with significant demographic disparities and declines over the elementary school years. Investigators examined the impact of the Fueling Learning Through Exercise (FLEX) Study school-based PA programs on school-time (sMVPA) and total daily MVPA among lower-income schoolchildren.

Study Design, Participants, and Intervention:

Urban elementary schools (n=18) were cluster randomized to 100 Mile Club®, CHALK/Just Move^™, or control. Data collection and analyses occurred from 2015 to 2019 among third and fourth grade schoolchildren (n=1,008) across 2 academic years.

Main outcome measures:

Student MVPA was measured by 7-day accelerometry (Actigraph GT3X+) at baseline (pre-intervention), midpoint (6 months), and endpoint (18 months). Mixed effects linear regression models examined program impact on sMVPA and daily MVPA, adjusting for clustering, demographics, weight status, free/reduced-price lunch eligibility, school PA environment, wear time, and weather. Program reach by sex, weight status, race/ethnicity, and MVPA was explored.

Results:

Of the 979 participants analyzed (8.7 [SD=0.7] years, 44% male, 60% non-White, 40% overweight/obese, 55% eligible for free/reduced-price lunch), 8.4% (18.2 [SD=7.9] minutes/day) and 19.8% (45.6 [SD=19.4] minutes/day) fulfilled the 30-minute sMVPA and 60-minute daily MVPA recommendations at baseline, respectively. Overall, daily MVPA decreased from baseline to 18 months (p<0.001, −5.3 minutes, 95% CI= −8.2, −2.4) with no effect of programming. However, for sMVPA, intervention schools maintained sMVPA across the 2 academic years whereas sMVPA decreased in control schools (p=0.004, −2.3 minutes, 95% CI= −4.3, −0.4). Program reach on sMVPA appeared equitable by sex and weight status, but was different by race/ethnicity (p<0.001).

Conclusions:

Two different school-based PA programs were effective in preventing the decline in sMVPA that is typical across the elementary years, with similar reach by sex and weight status. Multiple opportunities for PA during school are needed to promote meeting sMVPA recommendations among diverse children.

Trial registration:

This study is registered at www.clinicaltrials.gov NCT02810834.

INTRODUCTION

The Physical Activity (PA) Guidelines for Americans recommend that school-aged children engage in 60 minutes of moderate-to-vigorous PA (MVPA) daily given its positive association with numerous physical and mental health benefits,¹ including obesity prevention² and academic achievement.³ Despite this, less than half of U.S. children meet this recommendation, of which 30 minutes should be accrued during the school day.^4,5 Competing demands such as standardized testing requirements and budget constraints are cited as reasons for the lack of physical education and PA opportunities in schools,⁶ factors that heighten the need to identify opportunities to increase PA outside of traditional PA times (recess and physical education) without jeopardizing academics. Furthermore, disparities in PA-supporting policies, activities, and environments in lower-SES schools compared with higher-SES schools may result in even fewer minutes of school-time MVPA (sMVPA) for underserved children and widen the gap in PA behaviors among certain groups of children already at risk for low PA, such as girls and overweight/obese children.⁷ These barriers create the need for PA programming that demonstrates broad reach among children and that is easy to implement, time-efficient, and sustainable.^7,8

Although several evidence-based intervention studies demonstrating positive effects on MVPA have traditionally focused more on physical education,^9–11 recent studies have implemented programming to promote active classroom breaks and active academic lessons.^12–15 There is a need to build upon this research with studies that consider additional opportunities to incorporate PA during the school day, with objective measures of PA and longitudinal assessments over more than one school year. Rigorous assessments of these programs over the longer term will help researchers and school administrators better understand if programmatic uptake is sustainable and effective over time. Furthermore, many studies have focused on assessments in changes in PA during school time but have not addressed the impact of school-based PA programming on daily MVPA.

In 2012, the Active Schools Acceleration Project focused on scaling up innovative school-based PA programs and best practices across the U.S. The project led a nationwide competition among schools to identify scalable and sustainable school-based programs that promote MVPA for children throughout the school day. Two of the winning programs, 100 Mile Club^®, a walking and running program, and Choosing Healthy and Active Lifestyles for Kids (CHALK)/Just Move^™, a classroom-based PA break program, were successfully implemented nationwide and demonstrated potential for sustainability.¹⁶ This expansion identified the need for rigorous evaluation of each program’s impact on MVPA while also determining the reach of these programs.

To determine whether these programs can equitably increase children’s sMVPA, and to examine potential effects on daily MVPA, the Fueling Learning through Exercise (FLEX) Study was developed and implemented among lower-income districts across Massachusetts over 2 school years. The overall objective was to examine the impact of 100 Mile Club and CHALK/Just Move on school-time and daily MVPA among schoolchildren over the shorter term (1 school year) and longer term (2 school years). Child-level correlates, including demographics, weight status, and baseline MVPA, were also explored to understand program reach and uptake among these groups.

METHODS

The FLEX Study was a cluster RCT conducted in 18 diverse public elementary schools in Massachusetts. Children were followed over 2 school years (2015–2016 and 2016–2017). Third and fourth graders were consented at baseline in order to capture the upper elementary school years (Grades 3–4, Year 1; Grades 4–5, Year 2) and given that this age bracket is able to answer questionnaires with minimal assistance. Participants were measured at baseline (September 2015–February 2016), midpoint (April–June 2016; ~6 months of intervention) and post-intervention (April–June 2017; ~18 months of intervention). Figure 1 depicts the flow of study implementation and participation.

Figure 1.

FLEX study consort diagram.

FLEX, Fueling Learning Through Exercise.

Study Sample

Public elementary schools were recruited for participation if they had >40% of students who were eligible for free or reduced-price lunch or had >40% non-White students. Once enrolled, schools were randomized to 100 Mile Club (n=7), CHALK/Just Move (n=6), or control (n=5). Schools were block randomized in groups of 3 and stratified by district, to ensure comparable numbers of schools in each arm. Once a school agreed to participate, the study statistician informed recruiting staff of the school’s group assignment. All students were able to participate in programming regardless of whether they consented to participate in assessments. Control group schools were offered an intervention of their choice after completion of the evaluation in fall 2017.

Study staff conducted recruitment presentations in participating schools (n=18) to explain study and enrollment procedures and distribute consent materials to all third and fourth grade students (n=2,839 eligible) (Figure 1). Packets were sent home with each student and contained a flyer explaining the study, along with a parent consent form and child assent form. Documents were available in the primary languages spoken in targeted communities (English, Spanish, Portuguese, Arabic, Haitian–Creole, and Mandarin). The study was approved by the Tufts University IRB and individual school district research review boards where required. A detailed study protocol is published elsewhere.¹⁷

Measures

The 100 Mile Club is a walk/run program that encourages children to move 100 miles over the course of the school year (~3 miles/week). The program can be implemented before, during (physical education/recess), or after school and is led by 1 or 2 champions (e.g., physical education teachers) who log student miles (which could be marked off by tallying laps outside around a field/blacktop or indoors in the gymnasium/hallways). Champions were identified by school administrators, trained by study staff, and provided with materials, resources, and ongoing support (training manual, monthly e-mail communications, annual teacher/champion training) for implementation throughout the duration of the intervention. Schools were asked at a minimum to make the program available for ≥30 minutes/week to all children in Grades 3 and 4 (School Year 1) and Grades 4 and 5 (School Year 2) during the intervention and to track mileage during sessions.

The CHALK/Just Move program is composed of structured classroom-based PA breaks. Teachers were provided with a set of activity cards with various high- and low-intensity PA moves and suggested combinations of moves to group together to build movement breaks of 5–15 minutes. In addition, the activity cards offered ideas for integrating the moves with academic subjects (e.g., providing the answer to a multiplication problem by doing that number of jumping jacks). Teachers were encouraged to use the cards in most or all of the sessions and incorporate at least 1 break of 5 minutes/day. FLEX study staff trained classroom teachers, distributed cards and program guides with ideas for making breaks fun and engaging, and provided ongoing support and ideas for adaptations to classroom teachers throughout the intervention (i.e., monthly e-mails).

Participant demographic information, including sex, race/ethnicity, grade, age, maternal and paternal education, and free or reduced-price lunch status, was collected at baseline via parent questionnaire returned with the consent form.

Data collection took place at each participating school during regular school hours at each time point (baseline, midpoint, and post-intervention). Height and weight were measured in triplicate in light clothing with shoes removed using a portable stadiometer and portable digital scale (Model 213 and 803, respectively, Seca Weighing and Measuring Systems). BMI was calculated (kg/m²) and converted into z-scores using the Centers for Disease Control and Prevention age- and sex-specific growth charts.¹⁸ BMI percentiles were classified accordingly as: <5th percentile as underweight, 5th–85th percentile as normal weight, 85th–95th percentile as overweight, and ≥95th percentile as obese.

Children were outfitted with waist-worn tri-axial accelerometers (models GT3X+ or wGT3X-BT, ActiGraph, LLC) to measure school-time and daily PA over 7 consecutive days. These devices have been validated and calibrated for use with children.¹⁹ Participants were asked to wear accelerometers during all waking hours except when bathing or swimming. To support wear time compliance, children were provided a paper calendar-style tracking log to record the time each day they put the accelerometer on and the time they removed it before bed. Weather data were collected from the weather station nearest each school from publicly available National Oceanic and Atmospheric Administration data,²⁰ and the high temperature (continuous) and precipitation (binary: yes/no) were recorded for each day accelerometers were worn.

Accelerometers were initialized to store activity counts beginning at 00:00:01 on the first day the child would wear the device. Data were processed using a 15-second epoch to more accurately record short, intermittent bursts of activity common among children.²¹ Accelerometer data were categorized into minutes of sedentary, light, moderate, and vigorous activity using thresholds developed specifically for children.²² In-school hours were calculated for each participant based on specific school start and end times for each day the accelerometer was worn and averaged across days. Weekday (sum of before school time, school time, and after school time) and weekend day daily activity was also calculated accounting for school hours and average awake time. Non-wear time was defined as 60 minutes of 0 activity counts (i.e., 240 consecutive epochs with 0 counts), allowing for 1 minute of light activity (i.e., 4 consecutive epochs with 1–99 counts) every hour. Wear time was estimated by subtracting non-wear time from the total daily monitoring time. A day was considered a “valid day” if daily wear time was ≥10 hours or school wear time was ≥6 hours. Participants with <3 valid total wear days or 2 full school days were excluded from the analysis. Child program participation was assessed at the short-term and post-intervention time points via child self-report (yes/no).

Champions/study liaisons at each school were asked to complete a 14-item school PA environment survey, adapted from the School PA Policy Assessment,^8,23 to assess PA-supporting policies and practices related to physical education, recess, classroom-based PA, and before- and after-school programming. In intervention schools, champions (100 Mile Club) and classroom teachers (CHALK/Just Move) were asked to complete brief surveys at baseline, short-term, and post-intervention timepoints to provide details of implementation including frequency and dose of programming as well as facilitators of and barriers to implementation.²⁴

Statistical Analysis

The dependent variables were daily minutes of MVPA: (1) during school time and (2) over the course of the entire day. The independent variables were program allocation (control, 100 Mile Club, and CHALK/Just Move), timepoint (baseline, midpoint, and post-intervention), and their interaction terms (program allocation × time point). Other covariates included participant demographics (sex, race/ethnicity, grade, and free or reduced-price lunch eligibility) as well as weight status at baseline. Potential school-level covariates including PA environment survey (score ranging from 0 to 24, with 24 being the most conducive) were also adjusted for. Further adjustments were made for covariates that could affect MVPA including weekly mean temperature during the data collection week, a binary weekly precipitation indicator, and total accelerometer wear time.

Results from an RCT targeting PA recess time demonstrated a 9.9% increase in recess PA in the intervention group (mean=53.4% [SD=25.6%]; 43.5% [SD=27.6%], Cohen’s d=0.372).²⁵ To ensure power for pairwise comparisons, t-tests were used. To detect such effect with 80% power and 5% type-I error rate, 115 cases/arm were required, with 7 schools allocated to each arm, and roughly 17 students/school were needed. To account for clustering due to group randomization and repeated measurements, a design effect of 1.88 was applied based on a mild-to-moderate intraclass correlation coefficient of 0.03 (17 × 1.88 = 32 students/school), and an attrition rate of 25% (32/0.75 = 43 students/school).²⁶ The final sample size was 903 (43 students/school × 7 schools × 3 arms).

Descriptive statistics by program allocation at baseline were derived and tabulated. Mixed effects linear regression models were used to examine the impact of program, expressed as the interaction between program by timepoint, on sMVPA and daily MVPA adjusting for covariates. To account for clustering introduced by the longitudinal and group randomized features of the study, investigators specified the personal identification number as a random intercept nested within the school identification number. The regression coefficients were tabulated, and the adjusted dependent variables by program and timepoint were plotted to facilitate interpretation. All analyses occurred following the intervention period (2018–2019).

To evaluate if the programmatic effect differed among prespecified subpopulations, reach analyses were conducted, which involved adding a 3-way interaction among program, timepoint, and the variable of concern into the model (sex, weight status category at baseline, race/ethnicity, and baseline PA levels as measured by daily MVPA). Stata, version 15 was used for data management and analysis. Statistical significance was based on p<0.05.

RESULTS

Eighteen schools were recruited and randomized to either 100 Mile Club (n=7), CHALK/Just Move (n=6), or control (n=5); a total of 1,008 children provided parental permission and 979 children were measured at baseline (Figure 1). Table 1 presents the descriptive characteristics and MVPA of the study sample stratified by intervention group. Slightly more female students (56%) and almost an equal distribution of third and fourth graders enrolled at baseline (mean age=8.7 [SD=0.7] years). Approximately 60% of children were non-White, 55% were eligible for free or reduced-price lunch, and 40% were overweight/obese.

Table 1.

Descriptive Characteristics of Sample at Baseline (n=979)

Characteristics	100 Mile Club (n=369)	CHALK/Just Move (n=311)	Control (n=299)
Age, mean (SD)	8.7 (0.7)	8.7 (0.7)	8.7 (0.6)
Sex, %
Male	41.2	44.7	46.8
Female	58.8	55.3	53.2
Grade, %
3rd grade	48.0	48.2	43.1
4th grade	52.0	51.8	56.9
Race/ethnicity, %
Non-Hispanic White	40.6	35.1	46.2
Hispanic	31.2	47.7	41.3
Black	14.2	6.0	7.5
Multiracial/othera	14.0	11.2	5.0
Weight status
Underweight	1.4	1.3	0.7
Normal weight	56.4	60.1	61.2
Overweight	19.2	19.0	19.4
Obese	23.0	19.6	18.7
Free/reduced lunch price eligible (%)
Yes	54.2	62.4	49.2
No	35.0	29.9	44.5
Missing	10.8	7.7	6.3
Accelerometry MVPA, minutes, mean (SD)
School-time	17.5 (7.4)	17.5 (7.9)	19.7 (8.2)
Daily	43.4 (17.7)	45.6 (19.1)	48.1 (21.4)
Valid days weekday (weekend),b mean	4.5 (0.8)	4.5 (0.8)	4.7 (0.9)

^aOther includes Native American and Asian Pacific Islander.

^bValid days: valid accelerometry wear days on weekdays and weekend days.

MVPA, moderate-to-vigorous physical activity.

In total, 538 students (1,614 of 2,940 people-time cases, 54.9%) and 548 (1,644 of 2,940, 55.9%) provided school-time and total MVPA data, respectively, at all 3 timepoints. At baseline, just 7.8% of children met the school-time 30-minute PA recommendation and only 18.4% of children met the daily 60-minute recommendation (mean=18.2 [SD=7.9] minutes/day for school and 45.6 [SD=19.4] minutes/day for total daily). Appendix Table 1 shows the baseline MVPA minutes broken down by sex, grade, race/ethnicity, and weight status. Compared with female students, male students accrued more MVPA minutes both in school and throughout the entire day, as did fourth graders compared with third graders. Black and non-Hispanic White children had the highest MVPA minutes both in school and throughout the day. A downward trend in MVPA minutes was also observed with increasing weight categories; no group difference was seen in sMVPA. School PA environment score ranges were highest in control schools (16–21, median=17), followed by 100 Mile Club schools (13–19, median=16), and CHALK/Just Move schools (13–17, median=14.5) (data not shown).

Detailed program implementation results have been published elsewhere.²⁴ All CHALK/Just Move schools implemented the program across 2 years. Program dose for 100 Mile Club schools was 34.9 minutes/week compared with 19.7 minutes/week for CHALK/Just Move. Dose of PA received per session was also greater in 100 Mile Club schools compared with CHALK/Just Move schools (13.6 minutes/session vs 2.7 minutes/session). Approximately 54% of eligible children reported participating in CHALK/Just Move compared with 31.2% for 100 Mile Club.

From baseline to endpoint, the percentage of children meeting the school-time PA recommendation slightly improved in intervention schools (100 Mile Club, 5.9% to 7.3%; CHALK/Just Move, 7.9% to 8.3%) whereas control schools remained almost steady (11.8% to 11.5%). Table 2 and Figure 2 present the unadjusted and adjusted models for the impact of 100 Mile Club and CHALK/Just Move on sMVPA and daily MVPA. As described in the statistical analysis plan, 4 indicators jointly modeled the program × time interaction, which captures if the trends of the dependent variables differ across programs; their joint significance would indicate at least 1 arm is different than the rest. Table 2 shows that for sMVPA, the trends were different across program in both the unadjusted (p=0.008) and adjusted model (p=0.002). For daily MVPA, no significant difference in trends were observed (p=0.296 unadjusted, 0.295 adjusted). To aid interpretation of models with interaction terms, the authors plotted the adjusted marginal means for the 2 outcomes, unadjusted and adjusted (Figure 2). For sMVPA (bottom line cluster), though both 100-Mile Club and CHALK/Just Move remained steady to slightly increasing, the main driver of the unparallel trend was due to the drop in sMVPA in the control arm from midpoint to endpoint. For daily MVPA (top line cluster), the parallel inverted “V” shape was more ubiquitous compared with sMVPA. Although some line crossing (indication of being unparallel) was observed, the overall variabilities across program and time were not statistically different.

Table 2.

Changes in School-Time and Daily Moderate-to-Vigorous Physical Activity by FLEX Program Among Schoolchildren

Variable	Unadjusted modela		Adjusted modelb
	ß (95% CI)	p-value	ß (95% CI)	p-value
School-time MPVA (minutes)
Time point (time)
Baseline	ref	0.004	ref	0.001
Mid-point	1.26 (0.27, 2.25)	0.012	0.76 (−0.26, 1.79)	0.144
End-point	−0.65 (−1.85, 0.55)	0.29	−1.59 (−2.87, −0.32)	0.014
Program (Prog)
Control	ref	0.839	ref	0.815
100-Mile Club	−0.99 (−5.51, 3.52)	0.666	−1.21 (−5.94, 3.52)	0.615
CHALK/Just Move	−1.39 (−6.08, 3.31)	0.563	−1.67 (−6.88, 3.54)	0.529
Prog × Time interaction		0.008		0.002
Mid-point × 100-Mile Club	−0.23 (−1.58, 1.13)	0.743	−0.11 (−1.49, 1.27)	0.875
Mid-point × CHALK/Just Move	−1.63 (−3.05, −0.22)	0.024	−1.33 (−2.76, 0.1)	0.068
End-point × 100-Mile Club	1.92 (0.31, 3.54)	0.019	1.98 (0.37, 3.6)	0.016
End-point × CHALK/Just Move	0.80 (−0.85, 2.46)	0.342	1.92 (0.19, 3.66)	0.029
Daily MVPA (minutes)
Time point (Time)
Baseline	ref	0.001	ref	<0.001
Mid-point	2.80 (0.39, 5.2)	0.023	1.01 (−1.48, 3.5)	0.427
End-point	−2.72 (−5.64, 0.2)	0.068	−5.80 (−8.88, −2.72)	<0.001
Program (Prog)
Control	ref	0.308	ref	0.158
100-Mile Club	−4.06 (−9.63, 1.5)	0.152	−4.61 (−9.5, 0.28)	0.065
CHALK/Just Move	−3.72 (−9.57, 2.14)	0.213	−4.54 (−10.11, 1.04)	0.111
Prog × Time interaction		0.296		0.295
Mid-point × 100-Mile Club	−2.06 (−5.37, 1.24)	0.221	−1.25 (−4.60, 2.11)	0.466
Mid-point × CHALK/Just Move	−0.79 (−4.24, 2.66)	0.654	0.49 (−2.98, 3.97)	0.781
End-point × 100-Mile Club	2.02 (−1.9, 5.95)	0.312	2.43 (−1.48, 6.35)	0.223
End-point × CHALK/Just Move	−0.02 (−4.05, 4.01)	0.991	3.50 (−0.70, 7.7)	0.102

Notes: Changes assessed at mid-point and follow-up compared to controls (n=2,067 observations from 979 participants). All models adjusted for student’s unique identification number as random effect, nested within school. Boldface indicates statistical significance (p<0.05). Italicized p-values are based on extra sum of squares F-test.

^aAdjusted for wear time.

^bAdjusted for wear time, sex, race/ethnicity, grade, reduced priced/free school lunch status, baseline weight category, precipitation (yes/no), weekly mean temperature, and school physical activity environment score.

FLEX, Fueling Learning Through Exercise.

Figure 2.

School-time MVPA and daily MVPA by program over 2 school years.

Notes: Left figure unadjusted for covariates; right figure adjusted for sex, race, grade, eligibility for free/reduced price lunch, BMI category, accelerometer wear-time, precipitation (yes/no), average daily high temperature, physical activity environment score, baseline daily MVPA minutes, and clustering within schools.

MVPA, moderate-to-vigorous physical activity.

The PA programming did not differentially impact sMVPA by sex, baseline weight status, or baseline daily MVPA (data not shown). However, there was a significant effect of program impact on sMVPA by race/ethnicity (p=0.001), whereby Black children appeared to be less exposed to interventions/impacted by programming. PA program reach did not differ among any of the select subpopulations for daily MVPA.

DISCUSSION

Delivery of simple and accessible PA programs in lower-income schools over 2 school years was successful in preventing a decline in sMVPA among upper elementary schoolchildren, but these individual programs alone did not offer a large enough dose of activity needed to help children meet daily recommendations. Less than 9% of participating third and fourth grade children met the school-time PA recommendation at the beginning of the study period, with notable disparities among female and male students and children with obesity, versus normal weight/overweight children, consistent with prior research.^4,5 Given these significant sex and weight status differences at baseline, it is also noteworthy that the CHALK/Just Move and 100 Mile Club programs impacted these key demographic groups equally.

Many school-based PA interventions to date have not examined the impact of programming over more than one school year.^15,27–29 The current study was specifically designed to determine if PA program delivery and uptake (dose received) varied over the short-term (~6 months) versus longer-term (18 months; across 2 school years). There was no significant impact of programming on increasing school-time MVPA in the first year of the intervention, but there was a significant effect in the prevention of the decline in school-time MVPA that was observed in control schools across the second school year. Children in control schools decreased school-time MVPA by approximately 2 minutes over this period, whereas children in the intervention schools maintained school-time MVPA levels and somewhat improved the percentage of those meeting the school-time recommendation. Though the prevention of decline appears small, this is significant given how little time children engage in MVPA at school (~18 minutes/day) and the benefits of short bursts of activity on attention and cognition.^{12,13,30–35} The International Children’s Accelerometry Database, which includes data from 10 countries including the U.S., demonstrated a 4.2% decline in daily PA with each additional year of age³⁶; this equates to approximately 2 minutes/day, the difference observed in the present study.

Even though it would have been beneficial to improve daily MVPA, it is not surprising that implementing just 1 of these PA programs had no effect on daily MVPA. Although the goal of the present study was not to examine the combined effects of multiple programs, other studies have demonstrated the benefit of more comprehensive, multipronged approaches to increasing PA.²⁹ The Australian PA 4 Everyone trial among adolescents included 7 school-based PA strategies and demonstrated an increase in total daily MVPA of 14 minutes after 12 months of the intervention and 7 minutes after 24 months (the authors did not report on sMVPA).^37,38 In the Australian Supporting Children’s Outcomes using Rewards, Exercise and Skills (SCORES) trial,³⁹ children from low-SES schools randomized to a multicomponent intervention for 6 months increased children’s sMVPA by 2.9 minutes, but not daily MVPA, similar to the present findings. These findings illustrate the important contribution of individual programs implemented in FLEX, but also reinforce the need to provide a variety of opportunities to be active throughout the school day to increase the chances of meeting PA recommendations.

Two other factors should be considered when evaluating school-time PA opportunities. First, well-controlled trials that have shown improvements in school-time PA have not negatively impacted out-of-school PA.²⁷ In the present study, improvements in sMVPA did not result in a decrease in total daily activity (i.e., no negative “compensation” effect). Second, prior research has demonstrated the even greater importance of school-time PA opportunities for underserved children.⁴⁰ Lower-income schoolchildren accrue the majority of their daily MVPA minutes at school compared with children from middle-income households, likely due to fewer opportunities to be involved with organized PA and sports outside of school hours.⁴¹ Thus, continued efforts to increase school-time activity is warranted, especially in lower-income schools.

An important goal of this study was to understand whether 2 distinct types of PA programming, which were intended to be implemented outside of more traditional physical education or recess time, had differential “reach” among subgroups more at risk for low PA levels—female students, children with overweight/obesity,^36,42 and considering race/ethnicity. In addition, given the importance of a school’s support and resources for PA on children’s activity levels, investigators also considered the school’s PA environment (e.g., including pre-existing PA-supporting policies and programs, indoor and outdoor physical space).⁴¹ In this study, there was not a differential impact of programming on MVPA among these groups of children, nor among those with higher versus lower baseline levels of PA. However, significant differences were observed in reach by race/ethnicity that cannot be fully explained. Black children had the highest baseline levels of MVPA, yet appeared to be the least impacted by PA programming during school time. There could be several reasons for this finding. For example, this group may already have been more active or there may have been differences in program implementation that could not be fully documented. Further research is needed to explore this finding. In addition, culturally targeted interventions are needed to better reach this subgroup.

The FLEX Study targeted improving MVPA among schoolchildren. It would be logical to expect that the 100 Mile Club more effectively delivered MVPA because it is a walking/running program, whereas the CHALK/Just Move program incorporates less vigorous activity and specifically targets breaking up classroom sedentary time. However, both programs were effective in maintenance of sMVPA at the second year of follow-up. The effect of CHALK/Just Move in mitigating sMVPA decline may also partially be due to the broader implementation and reach of this program.²⁴ Future studies are also warranted to examine the impact of programming on shifting sedentary time to light bouts of activity given the negative impacts that extended sedentary time may have on attention and cognition in the classroom.⁴³

The notable strengths of this study include the collection of objectively measured PA for school time and total day over 2 academic years, the delivery of 2 simple PA interventions that can be implemented throughout the school day, and the ability to deliver these programs among lower-income schools serving diverse schoolchildren who likely have less access to PA opportunities.

Limitations

This study has several limitations worth noting. First, schools agreed to participate in this RCT given their interest in improving the PA of the children in their schools. However, schools did not necessarily receive the program that they thought might be most ideal for their school and available resources (i.e., limited indoor/outdoor space for 100 Mile Club or teachers not excited about implementing CHALK/Just Move active classroom breaks). Furthermore, once the schools were given the materials and training for their respective program, the delivery and uptake of the program by champions, teachers, and students happened with continued encouragement (e-mails, refresher trainings, periodic check-ins for possible questions and tips), but with minimal pressure from the study team to ensure implementation.²⁴

CONCLUSIONS

Randomization and implementation of 2 innovative school-based PA programs among lower-income diverse schools demonstrated an impact on mitigating the decline in children’s school-time MVPA typically observed across elementary school years. The effect was most significant in the second year of implementation and the reach of the programs did not differentially impact 2 groups at risk for lower PA levels, females and overweight/obese children. This study provides evidence that infusing new PA programs into schools can attenuate decreases in PA levels across the elementary school years. However, small effect sizes and implementation challenges highlight the need for multiple PA opportunities across the school day and attention to matching schools with programs that best fit their needs in order to move the needle on children meeting the 30-minute school-time recommendation.