Pickles Michael, Gregson Simon, Moorhouse Louisa, Maswera Rufurwokuda, Tsenesa Blessing, Dzamatira Freedom, Mandizvidza Phyllis, Bagnay Sophie, Dadirai Tawanda, Kumbirai Moyo Brian, Mugurungi Owen, Nyamukapa Constance, Hallett Timothy B
Medical Research Council Centre for Global Infectious Disease Analysis, School of Public Health, Imperial College London, London, UK.
Medical Research Council Centre for Global Infectious Disease Analysis, School of Public Health, Imperial College London, London, UK; Biomedical and Research Training Institute, Harare, Zimbabwe.
Lancet HIV. 2025 Mar;12(3):e214-e222. doi: 10.1016/S2352-3018(24)00317-5. Epub 2025 Feb 14.
An individual's HIV risk, and consequently their HIV prevention needs, change over time. In this study we aimed to quantify these changes, examine which life-course events were associated with them, and investigate the extent to which those life-course events were associated with HIV acquisition.
We used longitudinal data from eight rounds of a general population cohort in Manicaland province, eastern Zimbabwe, on sociodemographic and HIV risk behaviours, as well as HIV serostatus from the first seven rounds. We first visualised how HIV risk behaviours, comprised of having multiple, concurrent, non-regular, or transactional partners, condom non-use, drug use, and visiting bars, changed for individuals over time using Sankey diagrams. We then examined whether logistic regression models incorporating life-course events-namely, changes in marital or employment status, in-migration, or birth of a child-were more strongly associated with changes in HIV risk behaviour than models using only sociodemographic variables. Finally, we compared how well sociodemographic, HIV risk behaviour, and life-course events were associated with the person's risk of HIV acquisition as follows: we used logistic regression to identify which states (divided into sociodemographic, HIV risk behaviour, and life-course events) were most strongly associated with risk of HIV acquisition; based on this we use three models (corresponding to the three divisions) to identify the top 20% of individuals predicted to be at risk of acquiring HIV by each model, and computed what proportion of the actual HIV infection events occurred in that group.
Between 1998 and 2021, 21 213 individuals were interviewed at least twice, contributing a total of 34 212 participant observations. In this setting, individuals had periods of HIV risk lasting less than 3 years; only 12·3% (102 of 831) of those reporting transactional sex had also reported this in the previous round. We found that life-course events such as changes in marital status, employment status, and in-migrant status were associated with these changes in HIV risk behaviour. Using life-course events, particularly ones related to changes in marital status, 23% and 30% more HIV acquisitions were identified than using HIV risk behaviours or sociodemographic information, respectively.
HIV risk changes dynamically in this population, and life-course events could be a powerful way to understand changes in HIV risk behaviour and risk of HIV acquisition.
Bill and Melinda Gates Foundation, UK Medical Research Council, and Department for International Development.
个体的艾滋病毒感染风险以及相应的艾滋病毒预防需求会随时间变化。在本研究中，我们旨在量化这些变化，研究哪些人生历程事件与之相关，并调查这些人生历程事件与艾滋病毒感染的关联程度。
我们使用了来自津巴布韦东部马尼卡兰省普通人群队列八轮调查的纵向数据，内容包括社会人口统计学和艾滋病毒风险行为，以及前七轮的艾滋病毒血清学状态。我们首先使用桑基图直观展示了由拥有多个、同时存在、不固定或交易性伴侣、不使用避孕套、吸毒和光顾酒吧构成的艾滋病毒风险行为随时间的变化情况。然后我们检验了纳入人生历程事件（即婚姻或就业状况变化、迁入或生育子女）的逻辑回归模型是否比仅使用社会人口统计学变量的模型与艾滋病毒风险行为变化的关联更强。最后，我们比较了社会人口统计学、艾滋病毒风险行为和人生历程事件与个体艾滋病毒感染风险的关联程度，具体如下：我们使用逻辑回归来确定哪些状态（分为社会人口统计学、艾滋病毒风险行为和人生历程事件）与艾滋病毒感染风险关联最强；基于此，我们使用三个模型（对应三个分类）来确定每个模型预测有艾滋病毒感染风险的个体中排名前20%的人群，并计算实际艾滋病毒感染事件在该组中所占的比例。
1998年至2021年期间，21213人至少接受了两次访谈，共提供了34212条参与者观察数据。在这种情况下，个体的艾滋病毒感染风险期持续不到3年；报告有交易性行为的人中只有12.3%（831人中的102人）在上一轮也报告了这一情况。我们发现婚姻状况、就业状况和迁入状况等人生历程事件与艾滋病毒风险行为的这些变化有关。使用人生历程事件，尤其是与婚姻状况变化相关的事件，分别比使用艾滋病毒风险行为或社会人口统计学信息多识别出23%和30%的艾滋病毒感染情况。
该人群中的艾滋病毒感染风险动态变化，人生历程事件可能是理解艾滋病毒风险行为变化和艾滋病毒感染风险的有力方式。
比尔及梅琳达·盖茨基金会、英国医学研究理事会和国际发展部。
