Framework for Behavioral Risk Models of Alcohol Problems

摘要

ABSTRACT We propose to use an agent-based modeling approach to elucidate specific mechanisms underlying alcohol-related problems and provide a framework for developing comprehensive preventive interventions at the community level. Agent-based models are a tool to assess the health impacts of specific social mechanisms that support alcohol problems in community settings within a framework in which drinkers move through their environments (e.g., commute to home or school) and interact with each other and their environments in ecologically realistic ways (e.g., eat dinner with their families at home). This modeling process demands a high level of precision with regard to theoretical statements of social processes and makes explicit those processes in which we might creatively intervene to reduce problems, setting a high bar for scientific explanation. In this Center component, we will use the Framework for Reconstructing Epidemiologic Dynamics (FRED) platform, an agent-based model with realistic, open-access Census-based synthetic populations that capture the demographic and geographic heterogeneities of the population of the United States down to the Census Block level and allows agents to co-evolve with physical and social environments. We will: (1) Apply a behavioral risk model framework within FRED in order to help clarify specific social ecological mechanisms that underlie the etiology of alcohol-related problems in California communities; and (2) Develop a select set of scenarios to alter social ecological mechanisms related to drinking contexts and assess their effects on alcohol-related problems using the FRED platform. We will extend FRED’s capacities to include unique agent-environment dynamics typical of behavioral risk models related to alcohol use, alcohol use disorders, and related problems. A simulation platform must have a core set of capabilities in order to undertake the development of these behavioral risk models. These include the ability to model individual heterogeneities, heterogeneous agent interactions, and environmental heterogeneity, all of which are built into FRED. We will construct a set of models that accurately describe drinking agents, drinker movements, drinking environments, and specific risks within specific environments. Once the models have been constructed, we will change one element (e.g., increase the number of bars in a community by 10%) and measure how it impacts our outcomes of interest. We will apply this modeling approach to at least five kinds of problems based on current and prior Center Grant research: (a) Parents’ abuse and neglect of their children; (b) Early initiation of alcohol use and youth drinking problems (Component #3); (c) Health impacts related to hyper-availability of alcohol at the US-Mexico border (Component #4); (d) AUDs, heavy drinking and problems related to drinking outside the home (Component #5); and (e) The impact of social host ordinances on underage heavy drinking. BRMs allow for the integration of drinking agents and behaviors with drinking contexts—an essential step in designing and implementing effective strategies to prevent and reduce drinking problems within community systems. They make explicit those processes in which we might creatively intervene to reduce problems.