This application addresses broad Challenge Area (01) Behavior, Behavioral Change, and Prevention and specific Challenge Topic, 01-GM-101: Individual-based model of social behavior. Obesity, which has reached crisis proportions in the United States, is clearly a multi-factorial problem, and describing causal mechanisms and developing effective programs and policies has been challenging. Our primary aim is to borrow intensive scientific computing modeling methods used in other disciplines to develop a robust and well-characterized model of individual-level behavior related to nutrition and physical activity, as influenced by social and physical environmental factors. In particular, we propose to develop an agent-based model that quantifies social-demographic and built-environmental mechanisms by which behavioral choices are made in young women, and calibrate it to existing data from the Berkeley center of the ten-year prospective NHLBI Growth and Health Study (NGHS) of obesity development in Black and White girls. Once calibrated, the model, which will simulate complex and dynamic individual-based behaviors within changing environments, will be used to evaluate the impact of environmental policies aiming to promote healthy behaviors. Development of such a model will address the need for model-based assessments of health impacts associated with environmental changes. The modeling work will be innovative, leveraging recent advancements in computing technology, including algorithms that run on inexpensive multi-core graphics processing units (GPUs). Further, this study will meet the intent of the American Recovery and Reinvestment Act by stimulating interest in the use of new research methods and computing technology.
Our specific aims are: 1. To develop an individual-level agent-based model of food choice and physical activity for young women. The model will be based on knowledge about childhood eating behaviors and family environment, and a geographic information system, thereby accounting for the influences of social factors that operate at the individual, family, and community-level, as well as the spatial distribution of physical environmental features that could influence behavior, such as green space and the food environment. The model will be calibrated and validated to existing data from the NGHS. 2. To explore hypothetical scenarios using the calibrated/validated model based on changes to social and physical environments. Our model will simulate various scenarios (e.g., rezoning, introducing more food stores, etc.) to determine the potential effect on our cohort. Based on these findings we will determine the sensitivity of various policy changes for improving eating and physical activity behaviors. 3. To make use of new advancements in low-cost commercial GPU technology to support the use of such technology for massive simulation studies that previously required expensive supercomputing technology. This research concerns the challenge of obesity, and our current inability to quantify the impact of environmental changes on food choice and physical activity behaviors. Quantifying the impact of environmental changes on food and physical activity behaviors will provide much needed information to guide the development of effective obesity prevention programs and policies. Our research will use advances in computing technology to develop and test an innovative model that will quantify the impact of social and built environments on the development of food choice and physical activity behaviors in young women. The successful completion of this project will allow this innovative methodology to be applied to other socio-demographic groups, advancing our knowledge of effective approaches for dealing with the obesity epidemic.
This research concerns the challenge of obesity, and our current inability to quantify the impact of environmental changes on food choice and physical activity behaviors. Quantifying the impact of environmental changes on food and physical activity behaviors will provide much needed information to guide the development of effective obesity prevention programs and policies. Our research will use advances in computing technology to develop and test an innovative model that will quantify the impact of social and built environments on the development of food choice and physical activity behaviors in young women. The successful completion of this project will allow this innovative methodology to be applied to other socio-demographic groups, advancing our knowledge of effective approaches for dealing with the obesity epidemic.
| Crespi, Catherine M; Wang, May C; Seto, Edmund et al. (2015) Associations of family and neighborhood socioeconomic characteristics with longitudinal adiposity patterns in a biracial cohort of adolescent girls. Biodemography Soc Biol 61:81-97 |
| Decker, Anna L; Hubbard, Alan; Crespi, Catherine M et al. (2014) Semiparametric Estimation of the Impacts of Longitudinal Interventions on Adolescent Obesity using Targeted Maximum-Likelihood: Accessible Estimation with the ltmle Package. J Causal Inference 2:95-108 |
