Tracking the Spread of COVID-19 Through Air Travel
Jun 5, 2020
A New Tool Helps Analyze Commercial Air Travel Involving Infected Passengers
Cases of coronavirus disease 2019 (COVID-19), which was first reported in Wuhan, China, have been confirmed in all 50 U.S. states. In this report—the first of several from a RAND Corporation team examining the role of commercial air travel in the COVID-19 pandemic—we quantify potential vectors of virus transmission to the United States as a result of commercial air travel. Understanding the COVID-19 propagation patterns, regionally and globally, will help policymakers mitigate the resulting threats to public health.
The COVID-19 Air Traffic Visualization (CAT-V) tool combines case data from Johns Hopkins University with detailed air travel data from the International Air Transport Association. Together, these data sets make it possible to visualize how COVID-19 infections and commercial air travel have interacted to export infection risk across the world. Given air travel data and reported infection rates, our tool can also be used to estimate future patterns of COVID-19 transmission. As a result, policymakers, analysts, and others can estimate the impact of travel-related policy interventions, such as restricting air travel from various countries.
Using reported COVID-19 case data, we calculate case rates for every country. We combine the case rates with the numbers of air passengers originating from each country, along with the passenger destinations. These values combine to generate estimates over time of the numbers of infected passengers who travel from one place to another.
Specifically, we characterize the risk to country or region j, Rj, as
where Ai is the number of active (i.e., currently infected) cases in country or region i, while Pi is the population in country or region i. Nij is the number of passengers who travel from country or region i to country or region j.
Using our tool, we identified the eight countries with the highest risk of importing COVID-19 from China on January 24, 2020. Those countries, in descending order of risk, were Japan, Thailand, South Korea, the United States, Taiwan, Australia, Singapore, and Malaysia. This relative risk ranking is independent of the actual caseload in China, and, one week later, these same eight countries indeed had the most confirmed COVID-19 cases outside of China.
The tool also allows the user to visualize the risk associated with individual air travel routes. The routes can be depicted as either importation risks or exportation risks. The figure below shows the routes of importation risk to the United States on January 29, 2020, with the thickest red line beginning in China and ending in the United States.
As of the publication of this report, we continue to develop the CAT-V tool. The most important recent development is the ability to use real-time air travel data to drill down from the monthly data to daily data. Future tool developments will include the ability to examine province- or state-level data and the ability to visualize routes from and to specific airports.
We will continue to develop and use the CAT-V tool in new ways, and we will continue to release a stream of derivative findings on topics of interest to policymakers. The first set of findings released includes an estimate of the rate at which reported cases were being exported from China by the end of January 2020, an estimate of the rate at which cases were being underreported in China that same month, an estimate of the greatest source of importation risk to the United States in late February 2020, and the probable origin of cases in Gulf Cooperation Council countries.
In accordance with RAND's quality assurance standards, this analysis is based on the best available data. However, COVID-19 is an evolving threat, and even the best available data being used by government agencies and research institutes have significant limitations. Here, we outline several caveats regarding country-level data, passenger risk profiles, inaccurate country caseload reports, and other data limitations.
International air travelers from all parts of a country make up a relatively small subgroup of that country's population. This subgroup, including individuals who could be visitors to the country of origin, might be at greater or lesser risk of infection, relative to the national population as a whole, during any given stage of a pandemic outbreak.
Consistent with most academic research and public policy tools, our tool uses confirmed COVID-19 cases, as reported by individual countries. The true number of infected individuals in any given country is certainly higher than reported. Our tool does not correct for inaccurate or misleading case reporting.
Systemic differences between actual cases and reported cases could have material effects on our results. For example, a relatively high ratio of confirmed to actual cases in a country with strong testing protocols could falsely elevate the perceived risk of total infected passengers from that country.
To the extent possible, we framed the questions and conclusions of this analysis in a way that would reduce the sensitivity of our findings to these known differences.
Finally, the CAT-V tool simulates the movement of the coronavirus through air travel but does not assume or imply that the virus is transmitted on airplanes. The tool is about the movement of people from one country to another and the likelihood that those people are infected with the virus.
As of May 2020, the most important caveat of the CAT-V tool is the use of country-level data and thus the assumption of equal passenger risk profiles. Future versions of the tool will incorporate province- and state-level data.
To exhibit one limitation of country-level data, we consider Hubei province, for example. Hubei initially had a much higher case rate than did other Chinese provinces. This limitation is particularly relevant if the travel patterns of air travelers from Hubei were different from the patterns of air travelers from China more broadly. Nonetheless, in our initial analysis, we assume that the frequency and pattern of international air travel from Hubei province were roughly similar to those from the rest of China.
As of the publication of this report, our tool does not account for the lag time associated with the infection. In reality, the number of active COVID-19 positive cases today could reflect the rate of infection up to two weeks ago. Such lag times could be incorporated into future uses of the tool.
This analysis and associated reports describing findings from using the CAT-V tool will be further updated and released as warranted. When feasible, we will adjust the assumptions and correct the conclusions in accordance with the caveats discussed here.
This report and the CAT-V tool will be of interest to U.S. Department of Defense policymakers and planners at the strategic and operational levels. At the strategic level, projections of how COVID-19 is likely to propagate across the world can inform defense-related decisions on global deployment pathways and timing, force health protection measures, and emerging risks that competitors may exploit infection patterns for strategic advantage. At the operational level, the projections can inform
These issues will be explored in future reports. The tool and associated reports will also be of interest to U.S. policymakers and planners in other fields, such as foreign policy, public health, transportation, and homeland security.
The COVID-19 Air Traffic Visualization tool combines confirmed COVID-19 case data from the Johns Hopkins University Center for Systems Science and Engineering's COVID-19 Dashboard with detailed air passenger data from the International Air Transport Association's Nationality Traffic Report program. Together, these data sets allow the researchers to visualize and analyze the estimated transmission of the novel coronavirus via air travel, outline the resulting implications, and offer suggestions for minimizing the most-dangerous potential vectors.
This research was sponsored by the Office of the Secretary of Defense and the U.S. Air Force and conducted jointly within the Acquisition and Technology Policy Center of the RAND National Security Research Division and the Strategy and Doctrine Program of RAND Project AIR FORCE.
This report is part of the RAND Corporation Research report series. RAND reports present research findings and objective analysis that address the challenges facing the public and private sectors. All RAND reports undergo rigorous peer review to ensure high standards for research quality and objectivity.
Our mission to help improve policy and decisionmaking through research and analysis is enabled through our core values of quality and objectivity and our unwavering commitment to the highest level of integrity and ethical behavior. To help ensure our research and analysis are rigorous, objective, and nonpartisan, we subject our research publications to a robust and exacting quality-assurance process; avoid both the appearance and reality of financial and other conflicts of interest through staff training, project screening, and a policy of mandatory disclosure; and pursue transparency in our research engagements through our commitment to the open publication of our research findings and recommendations, disclosure of the source of funding of published research, and policies to ensure intellectual independence. For more information, visit www.rand.org/about/research-integrity.
This document and trademark(s) contained herein are protected by law. This representation of RAND intellectual property is provided for noncommercial use only. Unauthorized posting of this publication online is prohibited; linking directly to this product page is encouraged. Permission is required from RAND to reproduce, or reuse in another form, any of its research documents for commercial purposes. For information on reprint and reuse permissions, please visit www.rand.org/pubs/permissions.
The RAND Corporation is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND's publications do not necessarily reflect the opinions of its research clients and sponsors.