RESEARCH ARTICLE
Rural Middle School Nutrition and Physical Activity Environments and the Change in Body Mass Index During Adolescence MARGARET DEMMENT, PhDa NANCY WELLS, PhDb CHRISTINE OLSON, PhDc
ABSTRACT BACKGROUND: For rural adolescents, schools are among the few places where environmental interventions can promote health outside of the home. The goal of this study was to assess the nutrition and physical activity (N&PA) environments of schools attended by a birth cohort and examine the association with change in body mass index (BMI) from sixth to eighth grade. METHODS: Using data from adolescents of a rural New York State birth cohort (N = 281), we used linear mixed models to identify N&PA environments associated with change in BMI. We also examined family income trajectory as a potential modifier to consider how the association between school environment and change in BMI might differ depending on income. RESULTS: We found considerable heterogeneity in environments within and between schools. Among students with low-income trajectories, reductions in BMI z-scores were associated with school environments that promote better physical education (PE) and general (non-PE, non-sport) physical activity. Schools with better sports environments were associated with reductions in BMI for some students, but not lower-income students. CONCLUSIONS: School environments may have differing effects on students depending on their socioeconomic status. Strategies are needed to identify and address barriers that impair low-income students’ access to health-promoting school resources. Keywords: BMI z-score; linear mixed models; longitudinal; middle school; nutrition environment; physical activity environment. Citation: Demment M, Wells N, Olson C. Rural middle school nutrition and physical activity environments and the change in body mass index during adolescence. J Sch Health. 2015; 85: 100-108. Received on July 31, 2013 Accepted on July 12, 2014
R
ural children are less physically active and more at-risk for overweight and obesity than their metropolitan counterparts.1-3 This relation remains even after accounting for disparities in income and access to health care, suggesting that the rural environment may be more obesogenic than urban settings. Obesogenic environments reduce healthy behaviors such as healthy eating and physical activity. Therefore, decisions that impact how environments are created or used are central to increasing healthy
behaviors. For rural adolescents schools may be one of the few places policy or environmental interventions could impact them outside of the home.4 Middle schools support the development of behaviors in adolescence that carry on into adulthood, and thus, have the potential to play a critical role in adult obesity prevention.5 In a study using a nationally representative sample, school-level disadvantage was significantly associated with obesity in adolescence and highlighted the need for future research to explore the
a
Postdoctoral Associate, ([email protected]), Cornell University, Division of Nutritional Sciences, 352 Martha Van Rensselaer Hall, Ithaca, NY 14850. Associate Professor, ([email protected]), Cornell University, Department of Design and Environmental Analysis, 2429 Martha Van Rensselaer Hall, Ithaca, NY 14850. c Professor, ([email protected]), Cornell University, 406 Savage Hall, Ithaca, NY 14850. b
Address correspondence to: Margaret Mochon Demment, Postdoctoral Associate, ([email protected]), Cornell University, Division of Nutritional Sciences, 352 Martha Van Rensselaer Hall, Ithaca, NY 14850. We are grateful to the mothers who originally participated in the Bassett Mothers Health Project 1. We are also grateful to the staff at the schools that participated in our surveys. We thank Stacy Carling for her assistance with the medical chart audits and Dr. John May, Dr. Chris Kjolhede, the medical records staff of Bassett Healthcare, and Dr. Kimberly O’Brien for their input and guidance. The funding for the Bassett Mothers Health Project was provided by the National Institute of Child Health and Human Development (Grant No. HD 29549). The school assessments were supported by the Robert Wood Johnson Foundation through a dissertation award from the national program Active Living Research. MMD was supported by the USDA National Needs Graduate Fellowship Competitive Grant No. 2008-38420-04825 from the National Institute of Food and Agriculture.
100
•
Journal of School Health
•
February 2015, Vol. 85, No. 2
•
© 2015, American School Health Association
mechanisms between individual and school-level disadvantage and obesity risk.6 One possible mechanism is the obesogenity of the school environment. Despite the growing consensus that the environment influences nutrition and physical activity (N&PA) behavior, the science of measuring and evaluating this relationship is still in its infancy.7 Current methods for assessing the school environment are often used with a particular focus on one aspect of the environment, such as competitive food availability. In addition, previous studies of the school environment have focused heavily on self-reported data from students and administrators with no objective measures of the environment or input from key people who control selected aspects of the school environment. Several published reviews on the impact of the environment on childhood obesity revealed the following critical areas for future research: examining populations that are rural and vary in income, using longitudinal rather than cross-sectional data, and combining physical activity and nutrition variables.8-10 To address these gaps, we employ data from a rural, upstate New York birth cohort to investigate the links between income trajectories from birth to adolescence, the school N&PA environment, and measured body mass index (BMI) z-score during middle school, grades 6 through 8. The objectives of this study were to (1) characterize the school N&PA environment with a multidimensional assessment, (2) examine the association between the school environment and BMI z-score during middle school, and (3) determine whether family income trajectory modifies the relation between the school environment and BMI z-score during middle school. We hypothesized that a better middle school environment will be associated with a reduction in the BMI z-scores of students with poorer income trajectories.
METHODS Participants A longitudinal design was used to study the association between the school N&PA environment, family income trajectory, and adolescent BMI z-score in rural, upstate New York. Individual-level data were obtained from a population-based birth cohort whose mothers enrolled in the Bassett Mothers Health Project 1 (BMHP1). General health and growth data were obtained through a retrospective audit of the students’ medical charts, completed in August 2011. School-level data were collected from a contemporary assessment of 17 middle school N&PA environments using questionnaires administered to the principal, food service director, head PE teacher, and an observational checklist, completed in June 2011. At the school-level, there were 51 schools in the region attended by the students in the sample. The Journal of School Health
•
range in number of students per school was 1 to 40. The exclusion criterion for the school assessments was having fewer than 5 students from the cohort attending the school. Excluded schools were located at the periphery of the area served by Bassett Healthcare but all public middle schools in the 8-county region were included in this sample. Nineteen schools were eligible for the school assessment. All 19 schools participated in some aspect of the study. One school was missing the principal survey and another school had been closed in the process of the evaluation due to a natural disaster. The analysis sample consisted of the 17 schools (89%) that had complete data from all 3 respondents: principal, food service director, and PE teacher. At the individual-level, the original populationbased sample was comprised of 595 full-term, singlet children, born from June 1995 to July 1997. The sample was primarily white (96%), rural, and socioeconomically diverse adolescents (43% born
Rural Middle School Nutrition and Physical Activity Environments and the Change in Body Mass Index During Adolescence MARGARET DEMMENT, PhDa NANCY WELLS, PhDb CHRISTINE OLSON, PhDc
ABSTRACT BACKGROUND: For rural adolescents, schools are among the few places where environmental interventions can promote health outside of the home. The goal of this study was to assess the nutrition and physical activity (N&PA) environments of schools attended by a birth cohort and examine the association with change in body mass index (BMI) from sixth to eighth grade. METHODS: Using data from adolescents of a rural New York State birth cohort (N = 281), we used linear mixed models to identify N&PA environments associated with change in BMI. We also examined family income trajectory as a potential modifier to consider how the association between school environment and change in BMI might differ depending on income. RESULTS: We found considerable heterogeneity in environments within and between schools. Among students with low-income trajectories, reductions in BMI z-scores were associated with school environments that promote better physical education (PE) and general (non-PE, non-sport) physical activity. Schools with better sports environments were associated with reductions in BMI for some students, but not lower-income students. CONCLUSIONS: School environments may have differing effects on students depending on their socioeconomic status. Strategies are needed to identify and address barriers that impair low-income students’ access to health-promoting school resources. Keywords: BMI z-score; linear mixed models; longitudinal; middle school; nutrition environment; physical activity environment. Citation: Demment M, Wells N, Olson C. Rural middle school nutrition and physical activity environments and the change in body mass index during adolescence. J Sch Health. 2015; 85: 100-108. Received on July 31, 2013 Accepted on July 12, 2014
R
ural children are less physically active and more at-risk for overweight and obesity than their metropolitan counterparts.1-3 This relation remains even after accounting for disparities in income and access to health care, suggesting that the rural environment may be more obesogenic than urban settings. Obesogenic environments reduce healthy behaviors such as healthy eating and physical activity. Therefore, decisions that impact how environments are created or used are central to increasing healthy
behaviors. For rural adolescents schools may be one of the few places policy or environmental interventions could impact them outside of the home.4 Middle schools support the development of behaviors in adolescence that carry on into adulthood, and thus, have the potential to play a critical role in adult obesity prevention.5 In a study using a nationally representative sample, school-level disadvantage was significantly associated with obesity in adolescence and highlighted the need for future research to explore the
a
Postdoctoral Associate, ([email protected]), Cornell University, Division of Nutritional Sciences, 352 Martha Van Rensselaer Hall, Ithaca, NY 14850. Associate Professor, ([email protected]), Cornell University, Department of Design and Environmental Analysis, 2429 Martha Van Rensselaer Hall, Ithaca, NY 14850. c Professor, ([email protected]), Cornell University, 406 Savage Hall, Ithaca, NY 14850. b
Address correspondence to: Margaret Mochon Demment, Postdoctoral Associate, ([email protected]), Cornell University, Division of Nutritional Sciences, 352 Martha Van Rensselaer Hall, Ithaca, NY 14850. We are grateful to the mothers who originally participated in the Bassett Mothers Health Project 1. We are also grateful to the staff at the schools that participated in our surveys. We thank Stacy Carling for her assistance with the medical chart audits and Dr. John May, Dr. Chris Kjolhede, the medical records staff of Bassett Healthcare, and Dr. Kimberly O’Brien for their input and guidance. The funding for the Bassett Mothers Health Project was provided by the National Institute of Child Health and Human Development (Grant No. HD 29549). The school assessments were supported by the Robert Wood Johnson Foundation through a dissertation award from the national program Active Living Research. MMD was supported by the USDA National Needs Graduate Fellowship Competitive Grant No. 2008-38420-04825 from the National Institute of Food and Agriculture.
100
•
Journal of School Health
•
February 2015, Vol. 85, No. 2
•
© 2015, American School Health Association
mechanisms between individual and school-level disadvantage and obesity risk.6 One possible mechanism is the obesogenity of the school environment. Despite the growing consensus that the environment influences nutrition and physical activity (N&PA) behavior, the science of measuring and evaluating this relationship is still in its infancy.7 Current methods for assessing the school environment are often used with a particular focus on one aspect of the environment, such as competitive food availability. In addition, previous studies of the school environment have focused heavily on self-reported data from students and administrators with no objective measures of the environment or input from key people who control selected aspects of the school environment. Several published reviews on the impact of the environment on childhood obesity revealed the following critical areas for future research: examining populations that are rural and vary in income, using longitudinal rather than cross-sectional data, and combining physical activity and nutrition variables.8-10 To address these gaps, we employ data from a rural, upstate New York birth cohort to investigate the links between income trajectories from birth to adolescence, the school N&PA environment, and measured body mass index (BMI) z-score during middle school, grades 6 through 8. The objectives of this study were to (1) characterize the school N&PA environment with a multidimensional assessment, (2) examine the association between the school environment and BMI z-score during middle school, and (3) determine whether family income trajectory modifies the relation between the school environment and BMI z-score during middle school. We hypothesized that a better middle school environment will be associated with a reduction in the BMI z-scores of students with poorer income trajectories.
METHODS Participants A longitudinal design was used to study the association between the school N&PA environment, family income trajectory, and adolescent BMI z-score in rural, upstate New York. Individual-level data were obtained from a population-based birth cohort whose mothers enrolled in the Bassett Mothers Health Project 1 (BMHP1). General health and growth data were obtained through a retrospective audit of the students’ medical charts, completed in August 2011. School-level data were collected from a contemporary assessment of 17 middle school N&PA environments using questionnaires administered to the principal, food service director, head PE teacher, and an observational checklist, completed in June 2011. At the school-level, there were 51 schools in the region attended by the students in the sample. The Journal of School Health
•
range in number of students per school was 1 to 40. The exclusion criterion for the school assessments was having fewer than 5 students from the cohort attending the school. Excluded schools were located at the periphery of the area served by Bassett Healthcare but all public middle schools in the 8-county region were included in this sample. Nineteen schools were eligible for the school assessment. All 19 schools participated in some aspect of the study. One school was missing the principal survey and another school had been closed in the process of the evaluation due to a natural disaster. The analysis sample consisted of the 17 schools (89%) that had complete data from all 3 respondents: principal, food service director, and PE teacher. At the individual-level, the original populationbased sample was comprised of 595 full-term, singlet children, born from June 1995 to July 1997. The sample was primarily white (96%), rural, and socioeconomically diverse adolescents (43% born
