Mathematical Approaches of Modeling Obesity Trends

Abstract

The prevalence of obesity has drastically increased over the past several decades and has caused strain within the healthcare system, as obesity puts individuals at an increased risk for a variety of diseases and conditions. This project develops multiple mathematical models for obesity trends in the United States. We first used linear regression to model how the overall trends of obesity have changed over time. Linear regressions enabled us to gain insight into the relationship between obesity and societal factors such as poverty and food insecurity and enabled us to gain insight into the relationships seen in the data. Further, the rise in obesity levels has been theorized to mimic the spread of an infectious diseases. Since infectious diseases are often studied using SIR-models, we next developed an SIR model to study and analyze their effectiveness in modeling obesity. This enabled us to gain an understanding of the population level dynamics however might be overly complex. Finally, we used agent-based modeling strategies to create a probabilistic model of obesity trends. The use of agent-based models is supported by the theory that one’s social community may also impact the likelihood of becoming obese. The agent-based model was relatively simple but modeled the population level dynamics well. Developing these and similar models could enable the investigation of various intervention strategies to reduce obesity levels within the United States.