An Agent-based Model to Evaluate Positive Externality of Smoking Cessations 采用个体为本模型评估戒烟服务的正外部性


  • Donglan Zhang US Centers for Disease Control and Prevention
  • Lu Shi


agent-based model, smoking cessation


Introduction: Cigarette smoking can be viewed as a contagious disease whereby an active smoker will turn nonsmokers into passive smokers. Agent-based models (ABM) have been shown to have the advantage of exploring heterogeneity and inter-agent interaction, as compared with more aggregate models. In this study, we use an ABM framework and simulate a hypothetical tobacco control program in a multiunit dwelling, to examine the program’s “return on investment” in terms of passive smoking reduction. Method: We assume that in a multiunit building of 121 people there are 30 active smokers, with their neighbors as passive smokers. We simulate different spatial distributions of these 30 active smokers. Results: Helping the last active smoker quit smoking gave us a net reduction of four passive smoking cases, revealing a pattern of marginal increase in return to smoking cessation efforts. For population segments where active smokers are more likely to be clustered together (in households, work sites, residential units, etc.) this pattern of “increasing returns to health investments” will be even stronger. Discussion: This hypothetical intervention experiment provides an insight for the potential impact of reducing active smoking prevalence on reducing passive smoking prevalence. A model-based discussion can help public health stakeholders strategize their approaches to design tobacco control programs.






Research Article