Physical Environment and Activity Relationships in Adolescents: The PEAR Project

High levels of adolescent obesity present a substantial public health problem, with a high burden of future health risk and substantial projected costs to the health services. Insufficient levels of physical activity and poor diet are believed to account for increases in obesity. The reasons for changes in health behaviours have not been clearly identified, but may be due to the environment in which we live. There is now substantial interest in understanding the influence of the physical environment on activity and dietary behaviours, and a number of environmental features have been associated with health in young people. However our understanding of the relationship between environment and diet and physical activity is limited by poor measurement. The major limitation is that most studies relate features of the physical environment to places they think adolescents visit for physical activity or to obtain food rather than places they actually visit. For example, the presence of a park near a school may be associated with the physical activity of adolescents in the school, but they may not actually use the park. This lack of specificity with respect to where physical activity occurs (or where food is obtained) is likely to contribute to an underestimation of the strength of real associations between the physical environment and health outcomes.
The aim of the research is thus to determine which measures of the environment are related to time spent being active in different locations or near to food outlets. To do this we will use objective measurement of the environment (GIS) combined with objective measures of where people actually go (GPS) linked with an objective measure of physical activity (accelerometer) and a valid self-report measure of diet. We will carry out these measures in 900 adolescents aged 13-15yrs from schools in different areas (urban/rural/suburban). This information will give us a precise estimate of which features of the environment differentiate between adolescents with high/low physical activity and healthy/unhealthy dietary profiles. We will use this information to generate similar self-report measures for wider use in public health research and policy. The study will be conducted by a team of established researchers from Universities of Bristol, East Anglia, London and Cardiff who have expertise in developing physical activity and diet measures (Page), objective measurement of physical activity in children (Cooper), objective measurement of the physical environment (Jones) and food environment (Cummins) and school based research methodology and statistics (Moore).

High rates of obesity in adolescents are a significant public health burden as they increase risk of current and future morbidity. Low physical activity is an important contributor to adolescent obesity. Manipulating the physical as well as social environment has the potential to increase
the reach and effectiveness of health interventions. However evidence relating the physical environment to physical activity has been severely limited by methodological problems. The major limitation is that most studies relate features of the physical environment to places they think adolescents visit for physical activity rather than places they actually visit. We will address this problem by using personal GPS receivers to measure where adolescents go in the environment to be physically active.
Specifically we will determine which features of the physical environment relate to time spent in physical activity in different locations. We will also determine the validity and reliability of subjective self-report exposure measures of the environment in relation to objective measures to
provide tools suitable for wider population health research.We will conduct this research in 2 phases. Phase 1 will identify from existing evidence and six adolescent focus groups candidate physical environment exposures for further investigation. In Phase 2 we will objectively measure these exposures (using GIS) in combination with an objective measure of location (using GPS). Physical activity will be measured by accelerometer. We will measures 900 adolescents aged 13-15yrs from 6 schools differing in geographical (urban/rural/suburban) and social
composition. Multivariate regression analyses adjusting for clustering and confounders (e.g. BMI, pubertal status, daylight) will be carried out to determine which objectively measured environmental exposures relate most strongly to physical activity in different locations. Further we will compare self-report to objective exposure measures to determine their utility for use in wider public health research and ameliorate the need for the complex measures in this research. Output will be methodologically the strongest research to date investigating the association between the physical environment and physical activity and new self-report tools for
use in observational and intervention population studies and to inform health policy. Thus this research directly addresses measurement needs within highlight notice (MRC005780: PSMB) for improved methods for health-related aspects of the built and natural environment.