Social network dynamics of tobacco smoking and alcohol use among persons involved with the criminal legal system (PCLS): A modeling study.

Background: Tobacco smoking and alcohol use contribute to a synergy of epidemics (a "syndemic") that disproportionately affects persons involved with the criminal legal system (PCLS) and their social networks. An improved understanding of the complex interrelationships among the factors of the incarceration-tobacco-alcohol syndemic is essential to develop effective reform policies and interventions. However, collecting empirical data on these interrelationships is often hampered due to logistical and ethical challenges. Methods: We developed an agent-based network model (ABNM) to simulate the effects of the incarceration-tobacco-alcohol syndemic in the state of Rhode Island, USA. The model was validated and calibrated using empirical survey and demographic data. Outcomes included current smoking and heavy alcohol use rates in the first year after release among previously incarcerated agents and in their social networks. Results: The model successfully replicated demographic, substance use, and incarceration-related parameters. Simulation results suggest high rates of smoking (approximately 80% currently smoking persons in the first few weeks after release) and heavy alcohol use (approximately 40% current heavy alcohol use rate in the first few weeks after release) among PCLS, especially persons with multiple incarceration events. The model also estimated elevated rates of current smoking and current heavy alcohol use in the direct social contacts of PCLS. Discussion: This ABNM integrates biobehavioral health processes relating to incarceration and substance use. This model can be used as a platform to evaluate the potential impacts of interventions provided to PCLS and their networks.