Effect of hot zone infection outbreaks on the dynamics of SARS-CoV-2 spread in the community at large

Transmission of SARS-CoV-2 appears especially effective in "hot zone" locations where individuals interact in close proximity. We present mathematical models describing two types of hot zones. First, we consider a metapopulation model of infection spread where transmission hot zones are explicitly described by independent demes in which the same people repeatedly interact (referred to as "static" hot zones, e.g. nursing homes, food processing plants, prisons, etc.). These are assumed to exists in addition to a "community at large" compartment in which virus transmission is less effective. This model yields a number of predictions that are relevant to interpreting epidemiological patterns in COVID19 data. Even if the rate of community virus spread is assumed to be relatively slow, outbreaks in hot zones can temporarily accelerate initial community virus growth, which can lead to an overestimation of the viral reproduction number in the general population. Further, the model suggests that hot zones are a reservoir enabling the prolonged persistence of the virus at "infection plateaus" following implementation of non-pharmaceutical interventions, which has been frequently observed in data. The second model considers "dynamic" hot zones, which can repeatedly form by drawing random individuals from the community, and subsequently dissolve (e.g. restaurants, bars, movie theaters). While dynamic hot zones can accelerate the average rate of community virus spread and can provide opportunities for targeted interventions, they do not predict the occurrence of infection plateaus or other atypical epidemiological dynamics. The models therefore identify two types of transmission hot zones with very different effects on the infection dynamics, which warrants further epidemiological investigations.

新冠病毒（SARS-CoV - 2）的传播在个体近距离互动的“热点区域”似乎尤为高效。我们提出了描述两类热点区域的数学模型。首先，我们考虑一种感染传播的集合种群模型，其中传播热点区域由一些独立的同类群明确描述，相同的人群在这些同类群中反复互动（被称为“静态”热点区域，例如养老院、食品加工厂、监狱等）。假设除了病毒传播效率较低的“广大社区”部分外，还存在这些热点区域。该模型得出了一些与解读新冠疫情（COVID - 19）数据中的流行病学模式相关的预测。即使假定社区病毒传播速度相对较慢，热点区域的疫情爆发也可能暂时加速社区病毒的初始增长，这可能导致对普通人群中病毒繁殖数的高估。此外，该模型表明，热点区域是一个病毒库，使得在实施非药物干预措施后，病毒能在“感染平台期”长期持续存在，这在数据中经常被观察到。第二个模型考虑“动态”热点区域，这些区域可以通过从社区随机抽取个体反复形成，随后又解散（例如餐厅、酒吧、电影院）。虽然动态热点区域可以加快社区病毒的平均传播速度，并为有针对性的干预措施提供机会，但它们无法预测感染平台期或其他非典型流行病学动态的出现。因此，这些模型确定了两类对感染动态影响截然不同的传播热点区域，这值得进一步进行流行病学研究。