Skip to main content
SearchLoginLogin or Signup

Review 1: "Critical Role of the Subways in the Initial Spread of SARS-CoV-2 in New York City"

Published onApr 20, 2022
Review 1: "Critical Role of the Subways in the Initial Spread of SARS-CoV-2 in New York City"
1 of 2
key-enterThis Pub is a Review of
Critical Role of the Subways in the Initial Spread of SARS-CoV-2 in New York City
Description

AbstractWe studied the possible role of the subways in the spread of SARS-CoV-2 in New York City during late February and March 2020. Data on cases and hospitalizations, along with phylogenetic analyses of viral isolates, demonstrate rapid community transmission throughout all five boroughs within days. The near collapse of subway ridership during the second week of March was followed within 1-2 weeks by the flattening of COVID-19 incidence curve. We observed persistently high entry into stations located along the subway line serving a principal hotspot of infection in Queens. We used smartphone tracking data to estimate the volume of subway visits originating from each zip code tabulation area (ZCTA). Across ZCTAs, the estimated volume of subway visits on March 16 was strongly predictive of subsequent COVID-19 incidence during April 1-8. In a spatial analysis, we distinguished between the conventional notion of geographic contiguity and a novel notion of contiguity along subway lines. We found that the March 16 subway-visit volume in subway-contiguous ZCTAs had an increasing effect on COVID-19 incidence during April 1-8 as we enlarged the radius of influence up to 5 connected subway stops. By contrast, the March 31 cumulative incidence of COVID-19 in geographically-contiguous ZCTAs had an increasing effect on subsequent COVID-19 incidence as we expanded the radius up to 3 connected ZCTAs. The combined evidence points to the initial citywide dissemination of SARS-CoV-2 via a subway-based network, followed by percolation of new infections within local hotspots.

RR:C19 Evidence Scale rating by reviewer:

  • Reliable. The main study claims are generally justified by its methods and data. The results and conclusions are likely to be similar to the hypothetical ideal study. There are some minor caveats or limitations, but they would/do not change the major claims of the study. The study provides sufficient strength of evidence on its own that its main claims should be considered actionable, with some room for future revision.

***************************************

Review:

The methodological rigor and visualization are good but I have concerns about the implications and values adding to the existing literature.

Using smartphone tracking data to estimate the ZCTA based subway visits, the authors found across ZCTAs, the estimated volume of subway visits on March 16 was strongly predictive of subsequent COVID19 incidence during April 1. This is different from the traditional notion of geographic contiguity along subway lines. The authors concluded that combined evidence points to the initial citywide dissemination of SARS-CoV-2 via a subway-based network, followed by the percolation of new infections within local hotspots.

The authors explained the value of the current paper beyond the existing efforts by saying its contribution to the comprehensiveness of data for the analysis and conclusion. However, I would recommend the author focus on why *subway contiguity* is important? Beyond the comprehensive efforts (which is important too), how subway contiguity adds value to the existing studies (as shown below in the reference)?

Besides, I also recommend the author further consider the differences of infections by racially segregated geographic areas. This might be more informative to policy and public health decision-makers, rather than treating all the geographic areas as the same. Lastly, it is also important to be critical about the smartphone tracking data. Due to the possibility that frequent users of the subway may/may not have a smartphone, or allows it to track the data, there is a high chance that some mobility is not captured, or the data itself is a bias towards frequent smartphone users. I suggest the author discusses this as a limitation.


Comments
0
comment

No comments here

Why not start the discussion?