Social Determinants of Childhood Obesity System Dynamics Model

Research Question
The purpose of the Washington University-Transtria LLC childhood obesity prevention system dynamics computer model will be to understand and forecast the childhood obesity trends at the community level. The emphasis of this project is on representing and studying the social determinants of childhood obesity (e.g., structural racism, poverty), and evaluating the effectiveness of childhood obesity prevention strategies by focusing on two or more specific communities in a multiple case-study design.

Social determinants are widely regarded as important drivers of childhood obesity and other health trends (Brennan Ramirez, Baker, & Metzler, 2008). Social determinants cover a wide variety of potential factors including poverty, structural racism, and built environment. However, understanding how social determinants interact at the local community level is challenging since the specific mechanisms can vary from one community to another community and across time. The sequence and timing of multiple community interventions can also lead to complex interactions that are generally difficult to anticipate and vary by local communities (Hovmand and Ford, 2009). This can influence how federal, state, and local policies interact with interventions to impact childhood obesity trends in a particular community.

Modeling Approach
The model is based on principles of system dynamics and participatory group model building techniques. The model simulates the growth of age-birth weight cohorts over time, and calculates the prevalence of childhood obesity by comparing the weight distribution of the simulated population against a standard reference population. It is designed to be easily adapted to a wide variety of communities where local trend data on childhood may be limited. The primary inputs to the model are standard growth curve tables and fractional growth rates for a reference population from ages 0 to 20 and for both boys and girls. The model generates growth curves for each group, distributions of growth curves for each age cohort, and prevalence of childhood obesity over time. Prevalence of childhood obesity is calculated from comparison of the simulation population's weight percentiles against a reference population.

The model will be able to study the interactions of federal, state, and local polices at the community level. The model will also place an emphasis on involving stakeholders in the modeling process to better identify and understand these interactions and the role of social determinants of childhood obesity.

Principal Investigator

Laura Brennan, PhD MPH
President and CEO
Transtria LLC

Peter Hovmand, PhD
Assistant Professor
Washington University in St. Louis


Sarah Castro, MPH
Project Director
Transtria LLC

Julie Claus, MPH
Chief Operating Officer
Transtria LLC

Nasim Sabounchi, PhD
Washington University in St. Louis