Data Collection
In the present study, to calculate the importation risk of 2019-nCoV infection we collected the number of passengers for given airports of an origin country travelling to various international airports in India. We also accounted the total number of 2019-nCoV infected cases in the country of origin. Information about the total number of infected cases in the country of origin was obtained by extensive online search of more than five different sources, which mainly included the government sources and were rigorously verified and validated for any inconsistencies. To obtain the count of air passengers from a origin airport in given country, we relied on a web-based analytical tool called FLIRT (Huff, Allen, Whiting, Breit, & Arnold, 2016). The major application of FLIRT is to predict the air travel network flow between origin and destination airports throughout the world, where user can choose the origin and destination pair. Note that the FLIRT does not take an account of transit passenger is making between origin to destination and would always assign the last boarding airport as an origin, once entered in final destination. This web application allows user perform passenger simulations based on the air traffic between source and destination. To simulate the results, it uses a database of flight schedules of more than 800 airlines (Andrew Huff et al., 2016; A. Huff et al., 2016). Researchers in the past (Haider et al., 2020; Andrew Huff et al., 2016; A. Huff et al., 2016) have used this web interface to estimate the number of infected travelers originating from different countries to the respective destinations. For India, in this study, the countries selected for importation risk index calculation were China, France, Germany, Iran, Italy, Kuwait, Oman, Qatar, Singapore, Spain, Thailand, United Arab Emirates (UAE), United Kingdom (UK), and United States of America (USA), which constituted ~70% of the total infected cases. For simplicity in analysis, the countries Iran, Kuwait, Oman, Qatar, and UAE were grouped together as Middle Eastern countries. The major international airports from all the above-mentioned countries were selected to simulate the air traffic flow. The IATA codes for all the selected airports are listed in Table 1.
The FLIRT web application allows two different types of flight data simulation. The first mode includes the simulation of total seats offered by airlines between origin and destination airports, and the other mode simulates the total passengers travelled between airports. The second mode, however, has a significant disadvantage that it allows only 20,000 passengers to be simulated at a time, which is an upper bound on the total number of passengers that can be simulated. Such an upper bound can have a strong bias in the analysis, particularly for this kind of studies. Therefore, the first simulation mode was preferred here. This selection would not affect the analysis, as FLIRT assumes that the number of seats between two airports is directly proportional to the number of passengers between them (Andrew Huff et al., 2016). Thus, for each airport mentioned in Table 1, the simulation was performed using FLIRT as per the first criteria mentioned above. This simulation was carried out for the period of 4th March 2020 till 24th March 2020. This period was selected because the number of 2019-nCoV cases in India started to grow significantly from 4th of March, and India imposed an international travel ban from 24th of March. During this period India has observed a screening at all international airports across the country. After each simulation, the result was filtered for all the India bound flights for the respect chosen airports, and thus the total number of passenger seats and the number of Indian bound passenger seats were estimated from each airport of the origin. This data along with the number of infected persons in the origin airport country, from the various authenticated sources was then used to estimate an importation risk index.