While the results of the analysis by Tariq et al. are important, the recent developments in Singapore highlight the fact that heterogeneity in an outbreak is not just a possibility, but an inevitability. It appears at first glance that Singapore is experiencing three outbreaks, which can be confirmed by Singapore Ministry of Health data [2] and demonstrated by multiple distinct epidemic curves. The first was an epidemic of imported cases that peaked on March 23rd with less than 50 cases per day; this epidemic curve was abruptly halted by travel restrictions for short-term visitors [4]. Next came an epidemic curve of community cases that peaked on April 8th with less than 60 cases reported per day; this curve saw a sharp decline after a nationwide “circuit breaker” was announced, which closed all non-essential businesses and included social distancing mandates and restrictions on social gatherings, as well as strict penalties for those not complying [5].
However, travel restrictions and social distancing mandates did nothing to prevent a new epidemic of COVID-19 cases from the dormitories, which ballooned just as the strict “circuit breaker” measures were introduced on April 7th, and which peaked on April 20th with 1371 cases from the dormitories reported in a single day [2]. Even this idea of three separate epidemics in Singapore does not tell the whole story. The dormitories themselves are heterogeneous, with each dormitory, each building and each floor undergoing its own small epidemic and revealing its own specific transmission dynamics. Countries will increasingly see that epidemic curves, reproduction numbers and outbreak interventions will vary between contexts within a country.
In the COVID-19 pandemic, heterogeneity in transmission patterns are not the exception, they are the rule, and we will soon see this increasingly identified across the world. Communities may manage to reduce transmission by asking residents to stay at home and comply with social distancing measures, but broadly implemented interventions will not be effective or practical in many settings. COVID-19 outbreaks in the USA in prisons [6] and nursing homes [7] and some highly publicised outbreaks on international cruise ships [8] have already shown us that asymptomatic viral shedding may be the major driver of transmission in crowded conditions where social distancing is not a practical intervention. Outbreaks in three clusters are driving Australian figures even during their lockdown: a hospital in Tasmania, an abattoir in Victoria and a nursing Home in New South Wales [9].
Transmission dynamics and containment strategies will look very different in selected settings, and this cannot be seen any more clearly than in facilities housing hundreds or thousands of individuals. Community engagement and other public health measures will need to be targeted and adaptable to the needs of these settings. Singapore’s case and contact finding measures paired with isolation and quarantine were extremely effective at stopping community spread. The same principles are now being adapted and tailored to fit the needs of the dormitory residents, their living circumstances and the scale of the issue.