发表机构
School of Computing and Information Systems, The University of Melbourne; Department of Infectious Diseases, The University of Melbourne, at the Peter Doherty Institute for Infection and Immunity（墨尔本大学计算与信息学学院; 墨尔本大学传染病系，彼得·多赫蒂感染与免疫研究所）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究构建基于智能体的动态网络模型，利用NATSAL-3数据校准，探究伴侣动态与人群异质性对性接触网络结构及性传播感染传播的影响。
AI 中文摘要
人群层面的异质性，加之性伴侣关系的时序波动，塑造了性接触网络的结构，并可能显著影响性传播感染（STI）的传播。传统的静态网络模型假设伴侣关系具有固定属性（如数量和持续时间），可能无法充分捕捉伴侣关系对STI传播的影响。相比之下，基于智能体的动态网络模型为纳入个体和人群层面的异质性提供了灵活的框架。我们开发了一个基于智能体的动态网络模型，其中伴侣关系的形成与解散概率按年龄、性别和性取向（包括双性恋个体）分层，控制一夫一妻制及并发伴侣关系的形成，并通过持续时间相关的风险函数控制其解散。来自《全国性态度与生活方式调查》（NATSAL-3）的伴侣关系统计数据被用作模型校准目标，并采用拉丁超立方抽样（LHS）生成候选参数组合。参数估计通过选择使模型输出与校准目标之间均方误差（MSE）最低的组合来完成。本研究探讨三个问题：（1）基于智能体的模型能在多大程度上再现NATSAL-3中观察到的性伴侣关系特征；（2）并发性如何塑造动态性接触网络的结构；（3）并发伴侣关系如何影响STI传播的动态。在本研究中，我们发现年龄、性别和性取向等个体特征与伴侣数量、持续时间和并发性等伴侣属性之间的相互作用，在塑造人群层面的性接触网络以及进而影响STI传播动态方面起着关键作用。
英文摘要
Population-level heterogeneities, combined with temporal fluctuations in sexual partnerships, shape the structure of sexual contact networks and can substantially influence the spread of sexually transmitted infections (STIs). Traditional static network models, which assume fixed attributes of partnerships, such as count and duration, may not adequately capture the effects of partnerships on STI transmission. In contrast, agent-based dynamic network models offer a flexible framework for incorporating individual and population-level heterogeneities. We developed an agent-based dynamic network model in which partnership formation and dissolution probabilities, stratified by age, sex, and sexual orientation (including bisexual individuals), govern the formation of monogamous and concurrent partnerships and their dissolution via a duration-dependent hazard. Partnership statistics from the National Survey of Sexual Attitudes and Lifestyles (NATSAL-3) were used as model calibration targets, and Latin Hypercube Sampling (LHS) was used to generate candidate parameter combinations. Parameter estimation was performed by selecting the combination that produced the lowest Mean Squared Error (MSE) between the model outputs and the calibration targets. Our study addresses three questions: (1) how well can the observed characteristics of sexual partnerships in NATSAL-3 be reproduced using an agent-based model; (2) how does concurrency shape the structure of dynamic sexual contact networks; and (3) how do concurrent partnerships affect the dynamics of STI transmission. In this study, we find that interactions between individual characteristics such as age, sex, and sexual orientation, and partnership attributes such as count, duration, and concurrency play a critical role in shaping the population-level sexual contact network and, in turn, the dynamics of STI transmission.
