Big-data analytics to develop a precision public health approach to HIV prevention and treatment in a hyperendemic rural African population

尽管全民艾滋病毒预防和治疗工作取得了成功并付诸实施，但南非农村地区夸祖鲁-纳塔尔的艾滋病毒发病率仍然高得令人无法接受，发病率最高的是20至30岁的年轻人。虽然许多个人和结构性因素与艾滋病毒感染的风险有关，但传统的流行病学方法并不能充分解释艾滋病毒感染风险的总体情况。艾滋病毒预防工作要取得成功，就必须全面了解艾滋病毒风险和传播的基本机制。我们在这一提案中的总体目标是通过利用异质人口层面的数据和创新的大数据分析方法，揭示使年轻人处于艾滋病毒感染、艾滋病毒传播和治疗失败的最高风险的复杂关系和潜在机制。该项目将利用世界上最大的正在进行的基于人口的艾滋病毒队列研究之一-非洲卫生研究所在夸祖鲁-纳塔尔农村的人口队列研究，该研究所拥有个人一级的社会人口、生物和临床记录数据以及全面的基因组学数据。该项目将利用该研究所现有的大数据基础设施以及最近建立的研究平台，通过智能手机跟踪个人移动模式。最近在机器学习算法、基于智能手机的地理定位系统（GPS）跟踪软件应用程序和病毒基因测序技术方面的方法创新为解决关键知识差距提供了前所未有的机会，以确定预防艾滋病毒传播的最佳战略，并改善撒哈拉以南非洲贫困农村社区的艾滋病毒护理。具体而言，该项目将利用非洲卫生研究所完全集成的个人数据平台、基于智能手机的全球定位系统和创新的大数据技术：（20-30岁）艾滋病毒感染风险高; ii）使用机器学习算法确定使艾滋病毒感染者处于与艾滋病毒护理、治疗中断和病毒不抑制无关的高风险的因素的集合;以及iii）设计和试行智能手机干预，使用实时和精确的消息传递系统，针对艾滋病毒感染、传播和治疗失败的高风险人群。

Despite the successes and the implementation of population-wide HIV prevention and treatment efforts, HIV incidence rates remain unacceptably high in rural KwaZulu-Natal, South Africa, with the highest incidence peaks among young adults aged 20-30 years. While numerous individual and structural factors have been associated with the risk of HIV acquisition, traditional epidemiologic approaches do not sufficiently explain the totality of HIV acquisition risk. If HIV prevention efforts are to be successful, a comprehensive understanding of the underlying mechanisms for HIV risk and transmission will be critical. Our overarching goal in this proposal is to unravel the complex relationships and underlying mechanisms that place young adults at the highest risk of HIV acquisition, HIV transmission and treatment failure by harnessing heterogeneous population-level data and innovative big-data analytical approaches. The project will take advantage of one of the largest ongoing population-based HIV cohorts in the world - the Africa Health Research Institute’s population cohort in rural KwaZulu-Natal, with individual-level sociodemographic, biological, and clinical record data as well as comprehensive genomics data. The project will leverage the institute’s existing big data infrastructure as well as the recently established research platform for tracking individual mobility patterns via smartphones. Recent methodological innovations in machine-learning algorithms, smartphone-based geographic position system (GPS) tracking software applications, and viral gene sequencing technology provide an unparalleled opportunity to address key knowledge gaps to identify optimal strategies to prevent HIV transmission and improve HIV care in poor rural communities in sub-Saharan Africa. Specifically, the project will use the Africa Health Research Institute’s fully integrated individual data platform, smartphone-based GPS system and innovative big-data techniques: i) to elucidate the complex and interrelated factors that place young adults (20-30 years of age) at high risk of HIV acquisition; ii) to identify the constellation of factors that place HIV infected individuals at high risk of non-linkage to HIV care, treatment interruption, and viral non-suppression using machine learning algorithms; and iii) to design and pilot a smartphone intervention using a real-time and precision messaging system targeting those at high risk of acquisition of HIV infection, transmission, and treatment failure.