Sedda Luigi, Taylor Benjamin M, Cain Russell, Vajda Élodie A, Tatarsky Allison, Lobo Neil F
Lancaster Ecology and Epidemiology Group, Health Innovation One, Lancaster University, Sir John Fisher Drive, Lancaster, LA1 4AT UK.
School of Mathematical Sciences, University College Cork, Cork, T12 XF62 Ireland.
Ann Oper Res. 2024;342(3):1819-1835. doi: 10.1007/s10479-023-05491-3. Epub 2023 Jul 10.
Every day, hundreds of mosquito surveys are carried out around the world to inform policy and management decisions on how best to reduce or prevent the burden of mosquito-borne disease or mosquito nuisance. These surveys are usually time consuming and expensive. Mosquito surveillance is the essential component of vector management and control. However, surveillance is often carried out with a limited if not without a quantitative assessment of the sampling effort which can results in underpowered or overpowered studies, or certainly in overpowered studies when power analyses are carried out assuming independence in the measurements obtained from longitudinal and geographically proximal mosquito surveys. Many free, open-source and user-friendly tools to calculate statistical power are available, such as G*Power, glimmpse, powerandsamplesize.com website or R-cran packages (pwr and WebPower to name few of them). However, these tools may not be sufficient for powering mosquito surveys due to the additional properties of seasonal and spatially clustered repeated measurements required to reflect mosquito population dynamics. To facilitate power analysis for mosquito surveillance, we have developed TIMESS, a deployable browser-based Shiny app that estimates the number of repeated measurements and locations of mosquito surveys for a given effect size, power, significance level, seasonality and level of expected between-location clustering. In this article we describe TIMESS, its usage, strengths and limitations.
每天,全球都会开展数百次蚊虫调查,为有关如何最好地减轻或预防蚊媒疾病负担或蚊虫滋扰的政策和管理决策提供依据。这些调查通常既耗时又昂贵。蚊虫监测是病媒管理与控制的重要组成部分。然而,监测往往在对抽样工作量缺乏定量评估(即便不是完全没有评估)的情况下进行,这可能导致研究效能不足或效能过高,或者在对从纵向和地理上相近的蚊虫调查获得的测量值进行独立性假设的效能分析时,肯定会导致研究效能过高。有许多免费、开源且用户友好的统计效能计算工具可供使用,例如G*Power、glimmpse、powerandsamplesize.com网站或R语言的软件包(仅举几个例子,如pwr和WebPower)。然而,由于反映蚊虫种群动态所需的季节性和空间聚集重复测量的额外特性,这些工具可能不足以用于蚊虫调查的效能分析。为了便于进行蚊虫监测的效能分析,我们开发了TIMESS,这是一个基于浏览器的可部署Shiny应用程序,它可以针对给定的效应量、效能、显著性水平、季节性和预期的地点间聚集水平,估计蚊虫调查的重复测量次数和地点。在本文中,我们描述了TIMESS、其用法、优点和局限性。