KAVI-Institute of Clinical Research, University of Nairobi, Kenya; Center for Epidemiological Modelling and Analysis, University of Nairobi, Kenya; Paul G Allen School for Global Animal Health, Washington State University, United States; Ministry of Health Kenya, Kiambu County, Kenya.
Center for Epidemiological Modelling and Analysis, University of Nairobi, Kenya; Paul G Allen School for Global Animal Health, Washington State University, United States; Institute of Tropical and Infectious Diseases, University of Nairobi, Kenya.
Epidemics. 2022 Sep;40:100610. doi: 10.1016/j.epidem.2022.100610. Epub 2022 Jul 14.
Applied epidemiological models have played a critical role in understanding the transmission and control of disease outbreaks. Their utility and accuracy in decision-making on appropriate responses during public health emergencies is however a factor of their calibration to local data, evidence informing model assumptions, speed of obtaining and communicating their results, ease of understanding and willingness by policymakers to use their insights. We conducted a systematic review of infectious disease models focused on SARS-CoV-2 in Africa to determine: a) spatial and temporal patterns of SARS-CoV-2 modelling in Africa, b) use of local data to calibrate the models and local expertise in modelling activities, and c) key modelling questions and policy insights. We searched PubMed, Embase, Web of Science and MedRxiv databases following the PRISMA guidelines to obtain all SARS-CoV-2 dynamic modelling papers for one or multiple African countries. We extracted data on countries studied, authors and their affiliations, modelling questions addressed, type of models used, use of local data to calibrate the models, and model insights for guiding policy decisions. A total of 74 papers met the inclusion criteria, with nearly two-thirds of these coming from 6% (3) of the African countries. Initial papers were published 2 months after the first cases were reported in Africa, with most papers published after the first wave. More than half of all papers (53, 78%) and (48, 65%) had a first and last author affiliated to an African institution respectively, and only 12% (9) used local data for model calibration. A total of 60% (46) of the papers modelled assessment of control interventions. The transmission rate parameter was found to drive the most uncertainty in the sensitivity analysis for majority of the models. The use of dynamic models to draw policy insights was crucial and therefore there is need to increase modelling capacity in the continent.
应用流行病学模型在理解疾病暴发的传播和控制方面发挥了关键作用。然而，它们在公共卫生紧急情况下做出适当反应的决策中的效用和准确性，取决于其对本地数据的校准、为模型假设提供信息的证据、获取和传达结果的速度、政策制定者理解和愿意使用其见解的难易程度。我们对专门针对非洲的 SARS-CoV-2 传染病模型进行了系统评价，以确定：a）非洲 SARS-CoV-2 建模的时空模式，b）利用本地数据校准模型和建模活动中的本地专业知识，以及 c）关键建模问题和政策见解。我们按照 PRISMA 指南在 PubMed、Embase、Web of Science 和 MedRxiv 数据库中进行了搜索，以获取针对一个或多个非洲国家的所有 SARS-CoV-2 动态建模论文。我们提取了有关研究国家、作者及其所属机构、解决的建模问题、使用的模型类型、利用本地数据校准模型以及为指导政策决策提供模型见解的数据。共有 74 篇论文符合纳入标准，其中近三分之二（6 篇，占 6%）来自非洲的 6 个国家。最初的论文是在非洲报告首例病例后 2 个月发表的，大多数论文是在第一波疫情之后发表的。超过一半的论文（53 篇，占 78%）和（48 篇，占 65%）的第一作者和最后作者分别隶属于非洲机构，只有 12%（9 篇）的论文使用本地数据进行模型校准。共有 60%（46 篇）的论文对评估控制干预措施进行了建模。在大多数模型的敏感性分析中，发现传播率参数是最不确定的因素。因此，迫切需要利用动态模型来获取政策见解，从而需要在非洲增加建模能力。
