Keeling Matt J, Datta Samik, Franklin Daniel N, Flatman Ivor, Wattam Andy, Brown Mike, Budge Giles E
Zeeman Institute: SBIDER, University of Warwick, Coventry CV4 7AL, UK
Mathematics Institute, University of Warwick, Coventry CV4 7AL, UK.
J R Soc Interface. 2017 Apr;14(129). doi: 10.1098/rsif.2016.0908.
Sentinel sites, where problems can be identified early or investigated in detail, form an important part of planning for exotic disease outbreaks in humans, livestock and plants. Key questions are: how many sentinels are required, where should they be positioned and how effective are they at rapidly identifying new invasions? The sentinel apiary system for invasive honeybee pests and diseases illustrates the costs and benefits of such approaches. Here, we address these issues with two mathematical modelling approaches. The first approach is generic and uses probabilistic arguments to calculate the average number of affected sites when an outbreak is first detected, providing rapid and general insights that we have applied to a range of infectious diseases. The second approach uses a computationally intensive, stochastic, spatial model to simulate multiple outbreaks and to determine appropriate sentinel locations for UK apiaries. Both models quantify the anticipated increase in success of sentinel sites as their number increases and as non-sentinel sites become worse at detection; however, unexpectedly sentinels perform relatively better for faster growing outbreaks. Additionally, the spatial model allows us to quantify the substantial role that carefully positioned sentinels can play in the rapid detection of exotic invasions.
哨点监测点能够早期发现问题或进行详细调查，是人类、牲畜和植物外来疾病爆发应对计划的重要组成部分。关键问题包括：需要多少个监测点？它们应设置在何处？以及它们在快速识别新的入侵方面效果如何？用于监测入侵性蜜蜂病虫害的哨点蜂场系统说明了此类方法的成本和收益。在此，我们用两种数学建模方法来解决这些问题。第一种方法具有通用性，使用概率论证来计算首次检测到疫情时受影响地点的平均数量，提供了快速且通用的见解，我们已将其应用于一系列传染病。第二种方法使用计算密集型的随机空间模型来模拟多次疫情爆发，并为英国的蜂场确定合适的哨点位置。两种模型都量化了随着监测点数量增加以及非监测点的检测能力变差，监测点成功检测到疫情的预期增加情况；然而，出乎意料的是，对于增长较快的疫情爆发，监测点的表现相对更好。此外，空间模型使我们能够量化精心设置的监测点在快速检测外来入侵方面可发挥的重要作用。
