Developing TranStat: A user-friendly R package for the analysis of infectious disease transmission and control among close contacts

项目概要/摘要 家庭、教室、医院、工作场所和其他密切接触环境是感染的主要场所。 许多传染性病原体的传播。因为它们允许流行病学家遵循一个定义明确的 在有感染风险的人群中，对这些环境中传染病传播的纵向研究可以 产生独特的见解传染性和易感性的决定因素，在- 受感染个体随时间推移的繁殖力（传染性特征），以及控制的有效性 策略（例如，接种或掩蔽）。然而，这样的研究很少做，往往是分析我们- 为慢性病或人口水平的监测数据设计的统计方法，可以重新 导致严重偏见。为了实现这些研究的巨大潜力，为公共卫生应对提供信息， 传染病，开发用户友好且多功能的软件工具至关重要，可以提供访问权限 为密切接触环境设计的统计方法。该软件还必须支持正确的计算- 统计功效和样本量的确定，以帮助设计观察性研究和干预性研究， 在这种情况下的审判。基于我们在方法研究和代码开发方面的丰富经验， 密切接触人群中的各种传染病（包括埃博拉病毒，诺如病毒， 霍乱、SARS-CoV-2等），我们建议开发一个用户友好的，多功能的，计算效率高的 R软件包名为TranStat。我们的流行病学家、生物统计学家和计算生物学家团队 将实现以下具体目标：（1）集成离散时间的独立实现 将二项模型和连续时间成对生存模型链接到单个R包中。这一目标 将统一两个软件包的数据输入、模型规范和输出格式，同时提高用户 友好性、计算效率、功能和文档。(2)开发模拟工具， 计算密切接触环境中观察性研究和干预性试验的把握度和样本量。 这一目标将支持设计传染病在室内传播的流行病学研究， 货舱、教室、集体住房设施、工作场所等，可以为控制策略提供信息。(3)到 建立处理结果和协变量中缺失数据的能力，并考虑未观察到的异质性 在可传递性方面（例如，超扩散）。这一目标将使TranStat的用户能够部分保留- 在他们的分析中观察到的数据，以最大限度地提高统计能力，同时避免偏见和准确的定量， 不确定性。集成的、扩展的和免费提供的TranStat包将允许在 科学家通过研究传染病传播产生详细和可靠的科学见解， 密切接触群体。通过这些见解，TranStat将帮助政策制定者，官员的公共卫生， 和公众共同努力，更有效地控制流行病。

Project summary/abstract Households, classrooms, hospitals, workplaces, and other close contact settings are major venues for the spread of many infectious pathogens. Because they allow epidemiologists to follow a well-defined population at risk of infection, longitudinal studies of infectious disease transmission in these settings can generate unique insights into the determinants of infectiousness and susceptibility, the evolution of in- fectiousness over time in infected individuals (the infectiousness profile), and the effectiveness of control strategies (e.g., vaccination or masking). However, such studies are rarely done and are often analyzed us- ing statistical methods designed for chronic diseases or population-level surveillance data, which can re- sult in severe bias. To realize the enormous potential of these studies to inform public health responses to infectious diseases, it is critical to develop user-friendly and versatile software tools that provide access to statistical methods designed for close contact settings. This software must also support the proper calcu- lation of statistical power and sample size in order to aid the design of observational studies and interven- tion trials in these settings. Based on our extensive experience in methodological research and code devel- opment for a variety of infectious diseases in close contact groups (including influenza, Ebola, norovirus, cholera, SARS-CoV-2, etc.), we propose to develop a user-friendly, versatile, and computationally efficient R package called TranStat. Our team of epidemiologists, biostatistician, and computational biologists will achieve the following Specific Aims: (1) To integrate independent implementations of discrete-time chain binomial models and continuous-time pairwise survival models into a single R package. This aim will unify data input, model specification, and output formats for the two packages while improving user- friendliness, computational efficiency, functionality, and documentation. (2) To develop simulation tools to calculate power and sample size for observational studies and intervention trials in close contact settings. This aim will support the design of epidemiological studies of infectious disease transmission in house- holds, classrooms, congregate housing facilities, workplaces, etc., that can inform control strategies. (3) To build capacity to handle missing data in outcomes and covariates and to account for unobserved hetero- geneity in transmissibility (e.g., superspreading). This aim will allow users of TranStat to retain partially- observed data in their analyses to maximize statistical power while avoiding bias and accurately quanti- fying uncertainty. The integrated, expanded, and freely available TranStat package will allow epidemi- ologists to generate detailed and reliable scientific insights by studying infectious disease transmission in close contact groups. Through these insights, TranStat will help policy-makers, public health officials, and the public work together to control epidemics more effectively.