Abstract
OBJECTIVE: The aim of this study is to assess network-based weight loss interventions in the Chinese setting using agent-based simulation. METHODS: An agent-based model incorporating social, environmental and personal influence is developed to simulate the obesity epidemic through an interconnected social network among a population of 2197 individuals from the nationally representative survey. Model parameters are collected from literature and existing database. To ensure the robustness of our findings, the model is validated against empirical observations and sensitivity analyses are performed on calibrated parameters. RESULTS: When compared with the baseline model, significant weight difference is detected using paired samples t tests for network-based intervention strategies (p<0.05) but no difference is observed for the two conventional intervention strategies including choosing random or high-risk individuals (p>0.05). Targeting the most connected individuals minimizes the average population weight, average BMI, and generates a reduction of 2.70% and 1.38% in overweight and obesity prevalence. CONCLUSIONS: The simulations shows that targeting individuals on the basis of their social network attributes outperforms conventional targeting strategies. Future work needs to focus on how to further leverage social networks to curb obesity prevalence and enhance interventions for other chronic conditions using agent-based simulation.