基于大数据的MSM艾滋病管理和决策范式研究

The current HIV/AIDS strategies were failure to control HIV infection among men who have sex with men (MSM) because MSM is so much more sensitive, hidden and independent that we could not find source of infection among MSM early. The official report indicated that HIV following MSM behavior is spreading to attending student in China. However, the development of internet and MSM website provide the possibility and opportunity to learn characteristic of MSM while it is also providing convenient for MSM to find sexual partners. Focus on concern issues that we were lacking information to make effective HIV policies for MSM, this project intends to combine data both from internet and traditional public fields about MSM to map the social networks of MSM. By integrating methods and theories from system, big data and landscape epidemiology, we aim to set up a landscape identification system based on dynamic internet data related with MSM by investigating a series of algorithms in order to promote find infectious MSM early. Based knowledge of MSM, we will establish a coupled complex network of HIV dynamic transmission model to estimate effect of early detection HIV in MSM on transmission and trend of HIV and then to reach the selection of HIV management policies. In comparison with traditional management mode, the new mode integrating big data of website will provide detail and acute information of MSM which are helpful to make new and effective strategies to control HIV among MSM. We anticipate that the results from our project will promote finding of HIV infection, improve decision system of HIV prevention and control, and contribute to the new mode of management driving by big data in public health.

男性性接触（MSM）人群是一敏感而隐蔽的群体，当前艾滋病防控管理因缺乏对该群体的认识，不能及时发现该群体中的病人，艾滋病感染率持续上升并向青年学生蔓延。移动互联网的发展和MSM网站出现，为了解这个群体提供了机会和途径。本项目针对目前艾滋病防控管理瓶颈，拟通过整合来自网络空间和物理空间的MSM艾滋病相关数据，利用系统论、大数据和景观流行病等多学科理论和方法，建立动态MSM艾滋病全景信息辨识网络，掌握MSM人群社交特征，促进MSM艾滋病高危及感染者早期发现；构建MSM人群耦合复杂网络艾滋病传播动力学建模，评估病人早期发现对艾滋病流行规律和疾病负担的影响，实现艾滋病防控策略优选。与传统管理模式相比，基于融入网络空间动态数据驱动的MSM艾滋病管理与决策范式，能获得更及时、更完整的决策信息，预期能提高对MSM艾滋病高危和感染者早期发现能力，丰富艾滋病防控管理决策理论体系，推动新型公共卫生管理范式形

男性性接触（MSM）人群是我国艾滋病的高危人群，近年来MSM人群网络交友的兴起，给针对该人群的艾滋病防控带来了挑战，也使得MSM艾滋病相关数据呈现出多信息源、结构时变、强非线性的多体网络结构特征。项目研究紧紧围绕研究目标，完成面向千万级MSM艾滋病人群的、跨领域的多源融通大数据资源池的构建；构建了MSM人群社交行为特征辨识网络以及MSM人群艾滋病高危行为预警预测模型，包括头像信息特征辨识、情感特征辨识、数字足迹关联、行为轨迹评估、性别认同评估、舆情动力学网络等模型；建立起一套基于大数据的MSM艾滋病新型公共卫生管理与决策的初步范式，在黑龙江进行了预试验，实施效果良好，并为国家禁毒委、济南市公安局、成都市青羊区禁毒委等政府部门提供多项政策建议，产生了良好的社会效益。在项目资助下，研究成果完成并超额完成预期指标，共发表论文75篇，其中期刊论文33篇（SCI 19篇，EI 2篇），接收待发表2篇，会议论文42篇；撰写两本手册《药物滥用防治宣传教育手册》和《做自己健康的第一责任人——写给正在通过网络平台交友的MSM》 （在出版）；已获授权专利7项，在审11项；培养博士后1名，博士13名，硕士38名；课题负责人及参与人主办项目研讨会20次，在国内外主流学术会议上交流学术成果36次。