Abstract for: A Multi-scale Paradigm to Design Obesity Policy: Exploring the Integration of Individual- Based Modeling and System Dynamics

Complex adaptive systems-of-systems are inherently multi-scale across several dimensions, including temporal, geographical, and organizational. We present a multi-model paradigm integrating a localized community-scale individual-based model (IBM) with a population scale system dynamics (SD) model to analyze long term results of potential policy interventions for obesity prevention. The IBM uses virtual agents embedded in a social network to simulate the spread of opinions relating to nutrition and physical activity (N&PA) behaviors such as dieting and exercise, and the effects of these opinions on individual behaviors. The network structure uses a mixture of scale-free and uniformly random connections to represent a social network of relationships and interactions within a local community. The N&PA related health behaviors of individuals change dynamically relative to endogenous influences within their social network and exogenous influences from industry-based advertising and public health-related educational policies. The outputs of the IBM, seen as changes in obesogenic (N&PA unhealthy) behavior prevalences, can be used as inputs to a SD model to calculate the resulting changes in mortality and morbidity over the ensuing decades. We analyze and compare effects of possible policy interventions, and illustrate a policy cocktail that addresses multiple aspects of the obesity problem, resulting in amplification of desirable results and a strong uncertainty reduction.