Projecting the seasonality of endemic COVID-19 (preprint)

Importance Successive waves of infection by SARS-CoV-2 have left little doubt that COVID-19 will transition to an endemic disease, yet the future seasonality of COVID-19 remains one of its most consequential unknowns. Foreknowledge of spatiotemporal surges would have immediate and long-term consequences for medical and public health decision-making. Objective To estimate the impending endemic seasonality of COVID-19 in temperate population centers via a phylogenetic ancestral and descendent states approach that leverages long-term data on the incidence of circulating coronaviruses. Design We performed a comparative evolutionary analysis on literature-based monthly verified cases of HCoV-NL63, HCoV-229E, HCoV-HKU1, and HCoV-OC43 infection within populations across the Northern Hemisphere. Ancestral and descendent states analyses on human-infecting coronaviruses provided projections of the impending seasonality of endemic COVID-19. Setting Quantitative projections of the endemic seasonality of COVID-19 were based on human endemic coronavirus infection incidence data from New York City (USA); Denver (USA); Tampere (Finland); Trøndelag (Norway); Gothenburg (Sweden); Stockholm (Sweden); Amsterdam (Netherlands); Beijing (China); South Korea (Nationwide); Yamagata (Japan); Hong Kong; Nakon Si Thammarat (Thailand); Guangzhou (China); and Sarlahi (Nepal). Main Outcome(s) and Measure(s) The primary projection was the monthly relative frequency of SARS-CoV-2 infections in each geographic locale. Four secondary outcomes consisted of empirical monthly relative frequencies of the endemic human-infecting coronaviruses HCoV-NL63, -229E, -HKU1, and -OC43. Results We project asynchronous surges of SARS-CoV-2 across locales in the Northern Hemisphere. In New York City, SARS-CoV-2 incidence is projected in late fall and winter months (Nov.–Jan.), In Tampere, Finland; Yamagata, Japan; and Sarlahi, Nepal incidence peaks in February. Gothenburg and Stockholm in Sweden reach peak incidence between November and February. Guangzhou, China; and South Korea. In Denver, incidence peaks in early Spring (Mar.). In Amsterdam, incidence rises in late fall (Dec.), and declines in late spring (Apr.). In Hong Kong, the projected apex of infection is in late fall (Nov.–Dec.), yet variation in incidence is muted across other seasons. Seasonal projections for Nakhon Si Thammarat, Thailand and for Beijing, China are muted compared to other locations. Conclusions and Relevance This knowledge of likely spatiotemporal surges of COVID-19 is fundamental to medical preparedness and expansions of public health interventions that anticipate the impending endemicity of this disease and mitigate COVID-19 transmission. These results provide crucial guidance for adaptive public health responses to this disease, and are vital to the long-term mitigation of COVID-19 transmission. Key Points Question Under endemic conditions, what are the projected spatiotemporal seasonal surges of COVID-19? Findings We applied a phylogenetic ancestral and descendent states approach, leveraging long-term data on the incidence of circulating coronaviruses. We found that seasonal surges are expected in or near the winter months; dependent on the specific population center, infections are forecasted to surge in the late fall, winter, or early spring. Meaning Globally, endemic COVID-19 surges should be expected to occur asynchronously, often coincident with local expected surges of other human-infecting respiratory viruses.

COVID-19 deaths and hospitalizations averted by rapid vaccination rollout in the United States (preprint)

Importance Randomized clinical trials have shown that the COVID-19 vaccines currently approved in the US are highly efficacious. However, more evidence is needed to understand the population-level impact of the US vaccination rollout in the face of the changing landscape of COVID-19 pandemic in the US, including variants with higher transmissibility and immune escape. Objective To quantify the population-level impact of the US vaccination campaign in averting cases, hospitalizations and deaths from December 12, 2020 to June 28, 2021. Design Age-stratified agent-based model which included transmission dynamics of the Alpha, Gamma and Delta variants in addition to the original Wuhan-1 variant. Setting Our model was calibrated to COVID-19 outbreak and vaccine rollout in the US. Model predictions were made at the country level. Participants Simulated age-stratified population representing US demographics. Main Outcomes and Measures Cases, hospitalizations and deaths averted by vaccination against COVID-19 in the US, compared to the counterfactuals of no vaccination and vaccination administered at half the actual pace. Results The swift vaccine rollout in the US curbed a potential resurgence of cases in April 2021, which would have been otherwise fuelled by the Alpha variant. Compared to the scenario without vaccines, we estimated that the actual vaccination program averted more than 26 million cases, 1.2 million hospitalizations and saved 279,000 lives. A vaccination campaign with half the actual rollout rate would have led to an additional 460,000 hospitalizations and 121,000 deaths. Conclusions and Relevance The COVID-19 vaccination campaign in the US has had an extraordinary impact on reducing disease burden despite the emergence of highly transmissible variants. These findings highlight that the pace of vaccination was essential for mitigating COVID-19 in the US, and underscore the urgent need to close the vaccine coverage gaps in communities across the country. Key Points Question How effective was the United States (US) vaccination campaign in suppressing COVID-19 burden? Findings The vaccination campaign was highly effective in curbing the COVID-19 outbreak in the US. We estimated that the vaccine rollout saved over 275,000 lives and averted 1.2 million hospitalizations. Meaning The swift vaccine rollout in the US averted a remarkable number of cases, hospitalizations and deaths despite the emergence of highly transmissible variants.

Quantifying the potential dominance of immune-evading SARS-CoV-2 variants in the United States (preprint)

Recent evidence suggests that some new SARS-CoV-2 variants with spike mutations, such as P.1 (Gamma) and B.1.617.2 (Delta), exhibit partial immune evasion to antibodies generated by natural infection or vaccination. By considering the Gamma and Delta variants in a multi-variant transmission dynamic model, we evaluated the dominance of these variants in the United States (US) despite mounting vaccination coverage and other circulating variants. Our results suggest that while the dominance of the Gamma variant is improbable, the Delta variant would become the most prevalent variant in the US, driving a surge in infections and hospitalizations. Our study highlights the urgency for accelerated vaccination and continued adherence to non-pharmaceutical measures until viral circulation is driven low.

Quarantine and testing strategies to ameliorate transmission due to travel during the COVID-19 pandemic: a modelling study (preprint)

Background Numerous countries imposed strict travel restrictions, contributing to the large socioeconomic burden during the COVID-19 pandemic. The long quarantines that apply to contacts of cases may be excessive for travel policy. Methods We developed an approach to evaluate imminent countrywide COVID-19 infections after 0–14-day quarantine and testing. We identified the minimum travel quarantine duration such that the infection rate within the destination country did not increase compared to a travel ban, defining this minimum quarantine as “sufficient.” Findings We present a generalised analytical framework and a specific case study of the epidemic situation on November 21, 2021, for application to 26 European countries. For most origin-destination country pairs, a three-day or shorter quarantine with RT-PCR or antigen testing on exit suffices. Adaptation to the European Union traffic-light risk stratification provided a simplified policy tool. Our analytical approach provides guidance for travel policy during all phases of pandemic diseases. Interpretation For nearly half of origin-destination country pairs analysed, travel can be permitted in the absence of quarantine and testing. For the majority of pairs requiring controls, a short quarantine with testing could be as effective as a complete travel ban. The estimated travel quarantine durations are substantially shorter than those specified for traced contacts. Funding EasyJet (JPT and APG), the Elihu endowment (JPT), the Burnett and Stender families’ endowment (APG), the Notsew Orm Sands Foundation (JPT and APG), the National Institutes of Health (MCF), Canadian Institutes of Health Research (SMM) and Natural Sciences and Engineering Research Council of Canada EIDM-MfPH (SMM). Research in context Evidence before this study Evidence from early in the pandemic indicates that border closures at the epicentre slowed global dissemination of COVID-19. As community transmission became established in many nations, studies have suggested that the benefit of strict border closures in mitigating the transmission of disease from travellers diminished. Research for community settings has shown that testing later during quarantine, rather than upon entry into quarantine, can substantially shorten the duration of quarantine needed to reduce post-quarantine transmission. In particular for international air travellers, a 14-day quarantine can effectively be shortened to five or seven days. The number of infectious COVID-19 cases that escape from these quarantines depends on the prevalence of disease in the country the traveller originated as well as the travel volume into the country. Added value of this study We developed a framework to identify quarantine and testing strategies that enable travel from specific origins without increasing their infection rates per capita within destinations. No prior study has evaluated the appropriate duration of quarantine necessary to prevent any rise in infection rates per capita in the destination countries as a result of travel. By accounting for prevalence, daily incidence, vaccine coverage, immunity, age demographics, and travel flow between countries, we quantified the contribution of travel towards within-country the imminent infections in the destination country under different quarantine and testing strategies. For travel between 26 European countries, our results for the pandemic situation observed on November 21, 2021 demonstrate that there are often less burdensome quarantine and testing strategies that can serve as effective alternatives to strict border closure. Specifically, these estimated sufficient quarantine durations are especially dependent on COVID-19 prevalence and immunity within the two countries. We also found that asymmetry in the travel flow, just not the volume of travel flow, can also influence the estimated sufficient quarantine durations. Using data on variants of concern, including Omicron, we found that the adequacy of a border control strategy to limit variant spread depends strongly on the geographical distribution of the variant. While our results pertain to European countries, we also provide an interactive spreadsheet that can be used to determine appropriate quarantine durations between any two countries. Moreover, our framework can also be applied at any spatial or population scale within which movement restrictions could feasibly be implemented. Implications of all available evidence Travel quarantine and testing strategies can effectively mitigate importation and onward transmission within a country. Identifying sufficient strategies can allow countries to permit travel to and from other countries, without risking a short-term increase in infection rates. As long as the community transmission is occurring, the long-term epidemic trend within the destination country is more apt to be determined by other disease control measures, e.g., contact tracing, vaccination, and non-pharmaceutical interventions. Together, travel quarantine and other related control measures can mitigate the risk of transmission between countries, limiting the threat of variants of concern.

Cocooning is essential to relaxing social distancing (preprint)

As the first wave of COVID-19 recedes, policymakers are contemplating the relaxation of shelter-in-place orders. Using a model capturing high-risk populations and transmission rates estimated from hospitalization data, we find that postponing relaxation will only delay a second wave and cocooning vulnerable populations is needed to prevent overwhelming medical surges.