Teacher Physical Education Practices and Student Outcomes in a Sample of Middle Schools Participating in the Presidential Youth Fitness Program

Obesity and lack of physical activity among children and adolescents are public health problems in the United States. This Presidential Youth Fitness Program (PYFP) evaluation measured program implementation in 13 middle schools and its effect on physical education practices, student fitness knowledge, and student physical activity and fitness levels. PYFP, a free program with the potential to positively affect student health and fitness outcomes, was designed to improve fitness education practices that are easily integrated into existing physical education programs. We used a 2-group (13 PYFP and 13 comparison schools) quasi-experimental design to collect FitnessGram assessments, accelerometry data, and surveys of students, physical education teachers, and administrators. Although the program was positively associated with student cardiovascular endurance and physical activity gains during the semester, schools underused professional development courses and fitness recognition resources.


Introduction
Obesity and lack of physical activity among children and adolescents are public health problems in the United States (1,2). The 2018 Physical Activity Guidelines Advisory Committee Scientific Report confirms a strong association between higher physical activity levels and better health outcomes, including cardiorespiratory and muscular fitness, bone health, and weight status (3). Because school-aged children spend more than half of their waking hours in school (4) and engage in 20% to 30% of their total physical activity at school (5), schools are ideal settings in which to implement interventions to increase physical activity. Multicomponent school-based physical education (PE) programs improve children's health and academic outcomes (6,7), and a standards-based PE curriculum helps students develop the knowledge and skills needed to achieve and maintain health-enhancing levels of physical activity and fitness (8).
fectiveness of PYFP on key outcomes was not examined. The objective of this study was to describe findings from a PYFP outcomes evaluation.
Purpose and Objectives PYFP has hypothesized 4 key components to increase health-related fitness and knowledge among students and improve the effectiveness of PE: 1) use of FitnessGram (www.cooperinstitute.org/fitnessgram), a criterion-based fitness assessment that compares student measurements with a set of health-related standards; 2) a focus on fitness education to promote cardiovascular and muscular health; 3) professional development for PE teachers; and 4) recognition for students who achieve Healthy Fitness Zone standards.
We conducted the evaluation in 26 middle schools in the United States from October 2017 through June 2018 with the purpose of addressing the following questions: To what degree was PYFP implemented? Did PYFP implementation lead to integration of fitness education into physical education, improve fitness testing practices, or have a positive effect on PE and physical activity policies, practices, or environments? Did PYFP affect fitness knowledge, physical activity levels, or fitness among students?

Intervention Approach
On the basis of evidence that fitness assessment and education might influence fitness levels (10), PYFP aims to improve teacher fitness education practices and student knowledge, physical activity levels, and fitness with no cost to schools. PYFP schools included in this evaluation voluntarily applied for a grant from the National Fitness Foundation in 2014 or 2015 to participate in PY-FP and, as part of the program, received FitnessGram software licenses, teacher textbooks and online training, and student recognition items.

Evaluation Methods
The evaluation was based on systems thinking theory, which focuses on linkages and interactions among system components (in this study, components of PYFP) and assesses intended and unintended outcomes (11). We used mixed-methods, a 2-group quasiexperimental design. A power analysis indicated that a sample of 22 schools (11 PYFP schools and 11 control schools) would be appropriate. We used the following data sources: ICF staff conducted them at comparison schools; ICF conducted all follow-up FitnessGram assessments. Students wore ActiGraph accelerometers (model GT3XP-BTLE), positioned on the waist, for 7 days at baseline and 7 days at follow-up.
Teacher-level data: degree of PYFP implementation and teacherspecific volume of PE. We measured the degree of PYFP implementation by calculating program dose scores for the following: the proportion of students who received FitnessGram assessments, the number of professional development courses completed by PE teachers (4 were offered), the number of fitness education activities (ie, integration of fitness education into physical education), and use of fitness recognition (certificates awarded to students who score in the Healthy Fitness Zone in at least 5 FitnessGram assessment categories). We developed a scoring algorithm for these data; possible dose scores for each program component ranged from 0 to 4 (for a maximum of 16); higher scores indicate a greater degree of implementation. We measured PE volume for each teacher as the number of PE minutes offered between baseline and follow-up to control for the effect of the program in the regression models.
School-level data: physical activity/physical education policies, practices, and environment. We calculated a score for the physical activity/physical education environment from the following items in the PE teacher and administrator surveys: 1) the number of physical activity opportunities outside of PE time, 2) school environmental supports for physical activity/physical education teacher practices, and 3) administrative support. Total possible scores ranged from 0 to 19; higher scores indicate more positive environments.
Student-level outcomes. Outcomes were fitness knowledge (as measured in the student survey), BMI percentile (12), PACER scores (20-meter laps were converted to 1-mile run or walk times to estimate aerobic capacity [maximum oxygen consumption, Vo 2max ]) (13), and intensity of physical activity (time in moderateto-vigorous physical activity [MVPA]), determined from accelerometer data and child-based cut points (14).

Analysis
We used Stata version 11 (StataCorp) and SAS version 9.4 (SAS Institute Inc) for all analyses. We calculated descriptive statistics and performed bivariate analysis for school-level, teacher-level, and student-level data. We used multilevel linear models for clustering of students within classrooms for average MVPA (during and outside of PE), Vo 2max , and BMI percentile. The regression models included students with complete baseline and follow-up data for each outcome: MVPA (n = 387), Vo 2max (n = 1,985), and BMI (n = 1,783). Because baseline Vo 2max differed between groups, we analyzed follow-up scores by using a group interaction term. We found no group differences at baseline for BMI and MVPA, so we examined change from baseline to follow-up. On the basis of a sensitivity analysis, we included in MVPA analyses data from students with accelerometry data for 3 or more days of 8 hours per day (55% of all observations). We excluded from BMI analyses students whose BMI was greater than 70 (n = 7) or whose height decreased from baseline to follow-up (n = 361). Vo 2max analyses excluded PE classes with documented deviations from the measurement protocol (9 classes; 210 students).

Results
Student demographic characteristics did not differ significantly between groups for school enrollment, percentage of students who receive free or reduced-price lunch, or race (Non-Hispanic black and non-Hispanic white), but PYFP schools had a significantly greater percentage of Hispanic students than comparison schools (11% vs 7%; P = .01) (Table).
Degree of PYFP implementation. Of the program dose scores, the highest scores were received for FitnessGram assessments (3.9 of 4 points), followed by integration of fitness into PE (2.9 of 4 points). The lowest score was for completion of professional development courses (1.2 of 4 points); only 6 teachers completed 2 or more courses. Almost 40% of teachers reported time devoted to fitness education increased after PYFP implementation, and PY-FP teachers reported greater use of student physical activity logs (44% vs 16%) and individual feedback on students' physical activity plans (52% vs 32%) than comparison teachers.
Most administrators (92%) reported that PYFP had a positive effect on school climate; 85% agreed that PYFP added value to PE, physical activity programs, and students by improving PE quality. However, only 22% of PE teachers reported that PYFP had increased opportunities for physical activity breaks during school, and only 17% indicated that physical activity increased during PE.
Student outcomes. Student surveys showed no significant differences in knowledge between groups. Most students in both groups knew the importance of exercising 5 days or more per week, knew that 60 minutes of daily exercise is needed for good health, learned how to be fit in PE classes, and learned about setting fitness goals to improve fitness scores.
Student BMI percentiles were not significantly different between groups at baseline or follow-up, and change from baseline to follow-up was not significantly different between groups. MVPA levels were not significantly different between groups at baseline or follow-up (Figure), but the MVPA of PYFP students increased significantly more than the MVPA of comparison students (P = .04). In multivariate models, changes in MVPA and BMI from baseline to follow-up did not differ significantly by group after adjusting for age, sex, teacher-specific volume of PE, baseline values, and physical activity/physical education environment score. Younger students (P = .03) and students who were offered higher volumes (frequency and length) of PE (P =.03) had significantly lower BMI than older students and those with lower PE volumes. No predictors were significantly associated with the MVPA model. In the bivariate analysis, baseline Vo 2max was modestly but significantly higher among PYFP students than comparison students, whereas change in Vo 2max from baseline to follow-up was significantly higher among comparison students. After we adjusted for student age, student sex, teacher-specific PE volume, baseline Vo 2max , and physical activity/physical education environment, the regression model for Vo 2max at follow-up showed significant group effects, with higher scores at follow-up for PYFP students than comparison students (P < .001), but no group differences for the change-over-time model. Being younger (P = .01) and having higher baseline Vo 2max (P < .001) were significant predictors for follow-up Vo 2max .

Implications for Public Health
This evaluation was the first to assess the effect of PYFP on student health and fitness and to use comparison schools. Findings indicated school administrators and teachers strongly supported PY-FP and attributed substantial improvements in PE courses and PE/ PA environments to the program. Moreover, the positive associations between PYFP and student cardiovascular endurance at follow-up provided evidence for the health benefits of the program.
Because PYFP's components are consistent with the recommendations and evidence described in the National Physical Activity Plan (15) and the Comprehensive School Physical Activity Program (16), and they involve little or no cost to participating schools, PYFP should be considered a strong and practical strategy to improve student physical activity levels. Interestingly, degree of PYFP implementation was not significantly associated with student outcomes, and professional development courses and fitness recognition resources were underused, which suggests more research is needed to determine the amount of training required for teachers and the role of student recognition in promoting student fitness achievements.
Our study has several limitations. We did not randomly assign schools to PYFP or comparison conditions. PYFP schools had voluntarily started the program 2 or 3 years before the evaluation, and the evaluation was designed to examine PYFP as implemented.
Thus, a selection bias may have been present. The use of matched comparison schools and statistical controls was intended to minimize the influence of factors known to affect student fitness (eg, race/ethnicity, sex, age), but they might not have eliminated the influence of known and unknown factors. In addition, our study was retrospective, so reports by school personnel might have been influenced by memory bias. Lack of random assignment and the retrospective design preclude the ability to determine cause and effect. Because of the time required to obtain approvals and recruit schools, the study period was limited to 1 semester. A longer study period might have produced different findings.
PYFP is a free program with the potential to positively affect student health and fitness outcomes. Strategies to support greater and more consistent use of PYFP resources, such as professional development, and program enhancements to address implementation barriers should be considered.  Non-Hispanic black, % 9 9 .50 Hispanic, %   and by 2-sample t test for student BMI percentile and student Vo 2max . b Teacher-level variables; online surveys were completed by teachers once during semester. c Fitness education covers such concepts as the importance of health-related fitness and physical activity for good health. d Student-level variables; paper-and-pencil surveys were completed by students once during semester; BMI and Vo 2max were measured at beginning and end of semester. e Determined by bivariate analysis of PACER scores; 20-m laps were converted to 1-mile run/walk times to estimate aerobic capacity (maximum oxygen consumption, Vo 2max ) (12). Vo 2max is measured in mL of oxygen used in 1 minute per kg of body weight (mL/kg/min).