Similarities and differences of the COVID-19 second surge in Europe and the Northeast United States

We estimated the aggregate prevalence over time for Europe and the Northeast US to characterize the COVID-19 second surge in these regions. We find a starting date as early as July 3 (95 %CI: July 2 - July 5) for Europe (Schengen Area) and August 19 (95 %CI: August 16 - August 22) for the Northeast US; subsequent infectious populations that, as of December 7, have always increased or remained stagnant; and the resurgences being the collective effect of each overall region with no one country or state dominating the dynamics by itself.


INTRODUCTION
The Northeast United States (US) and Europe are progressing through a second surge of the COVID-19 outbreak after experiencing a similar initial outburst and subsequent regression.
None of these resurgences was widely expected nor anticipated [1,2]. Both areas display high mobility among their locations and broad independence among locations to enact measures to mitigate the propagation of the outbreak. In the case of Europe, Schengen Area countries allow for unrestricted border crossings among them. The characterization of the similarities and differences of the outbreak progression in these two regions is needed to provide insights into the effectiveness of the actions taken, to ascertain the extent their results can be extrapolated from one region to another, and to informedly mitigate the current and potentially forthcoming resurgences.
We computed the aggregate values of the infectious populations and the corresponding time-varying reproduction numbers for the Northeast US and Europe's Schengen Area from the individual values of the locations of each region [3]. Reproduction numbers were computed from growth rates considering a gamma-distributed generation interval with a mean of 6.5 days and a standard deviation of 4.2 days (Methods) [4].
Aggregate values were computed from infectious-population-weighted individual location values (Methods). The individual location data was downloaded on December 31, 2020 from https://github.com/Covid19Dynamics/trajectories. All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

RESULTS
The initial progression of the overall infectious populations for both regions consisted of exponential growth followed by an exponential decay ( Figures 1A and 1B). Subsequently, there was a sharp transition to fast exponential growth in Europe on July 3 (95% Confidence Interval (CI): July 1-July 6) and a stagnant overall infectious population in the Northeast US, which started to grow slowly but with increasing speed on August 19 (95% CI: August 16 -August 23). Concomitantly, the time-varying reproduction numbers crossed above one on the resurgence dates less abruptly in the Northeast US than in Europe ( Figure 1C), reaching maximum values of 1.50 (95% CI: 1.48-1.51) in Europe and 1.30 (95% CI: 1.27-1.34) in the Northeast US. The sharp resurgence to exponential growth in Europe is coincidental with lifting major nonpharmaceutical interventions that curved the outbreak [5], including the coordinated end of travel bans in Schengen Area's countries on July 1, 2020 [6]. As of December 7, the estimated infectious population is still in a growing phase but slowing down in the Northeast US ( Figure 1A) and has remained in a stagnant state since early November in Europe ( Figure 1B).
The aggregated values for each region provide precise evidence of sustained growth of the outbreaks already over the summer, despite the uncertainty and variability present in each of the locations independently ( Figure 1D). There is also robust evidence that the resurgence was not driven by a unique location since any aggregate value of the starting date for each region leaving one of their locations out is within the confidence limits of that of the overall region (Methods, Figure 1D, and Supplementary Figure 1). Therefore, the resurgence was the collective effect of each overall region. All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

DISCUSSION
These results highlight the need to implement policies and surveillance approaches that include data at a supra-location level when there is high mobility among locations. In this regard, the Northeast US, as a region, is closely trailing Europe in the second surge of the outbreak, but with a markedly smaller growth and evidence of slowing down earlier in the growth phase than Europe. Key differences in the actions taken included more gradual lifting and swifter progressive reimplementation of measures in the Northeast US than in Europe [5]. The progression over time of the aggregate prevalence of Europe's Schengen Area countries shows, with high reliability, that Europe's initial acting upon the second surge in mid-late October [1] took place well after a three-month-long period of sustained growth of the COVID-19 infectious population in the overall region, which has resulted in a second surge deadlier than the first one [7]. Such a death toll has not taken place yet in the Northeast US and it is potentially avertable with the social restrictions currently in place [5] according to the observed trend in the progression of its estimated overall aggregate prevalence.

METHODS
The aggregate infectious population for a region is expressed as All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 5, 2021. The overall time-dependent growth rate is given by where ',# ( ) is the growth rate of the infectious population of the specific location with index . This expression follows from The corresponding upper and lower confidence intervals are computed as (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 5, 2021. Analogously, the upper bound of the confidence interval is computed as the first day after reaching the minimum in which the lower bound of the confidence interval of the reproduction number is above 1.
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 5, 2021. ; https://doi.org/10.1101/2021.01.02.21249125 doi: medRxiv preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 5, 2021. ; https://doi.org/10.1101/2021.01.02.21249125 doi: medRxiv preprint 9 resurgence dates correspond to the minimum infectious population reached for each location and region (red crosses) and for the whole region without the population of the location indicated (blue circles) (D). The intervals represent the 95% CI.
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 5, 2021. All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 5, 2021.  (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 5, 2021.  (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 5, 2021.  (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 5, 2021.  (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 5, 2021.

Infectious individuals
Supplementary Figure 1.6 All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 5, 2021.

Infectious individuals
Supplementary Figure 1.7 All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 5, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 5, 2021.

Infectious individuals
Supplementary Figure 1.9 All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 5, 2021.

Infectious individuals
Supplementary Figure 1.10 All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 5, 2021.

Infectious individuals
Supplementary Figure 1.11 All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 5, 2021.

Infectious individuals
Supplementary Figure 1.12 All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 5, 2021.

Infectious individuals
Supplementary Figure 1.13 All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.