Background
Commercial flights contributed to the early-stage international transmission of severe acute respiratory
syndrome coronavirus 2 (SARS-CoV-2). Understand the effect of international and inter-state flights on the
virus transmission is important to evaluate the initial response of the outbreak. This study investigated the
likely date of the emergence of the first COVID-19 case in Wuhan, China.
Methods
We constructed a geographical-structured model, including 9122 county-level geographical units in 250
different regions or countries, and 26,094,036 flight plans. Using the model, we estimated the date of the
number of deaths and the date of first death caused by COVID-19 in 155 different countries. We set a
certain trigger for country and county level lockdown, and a built-in flight randomizer, to assess the
different evidence that can suggest the possibility of different dates to be the emergence of the first
COVID-19 case.
Findings
We found the median number of global deaths caused by COVID-19 from 50 trails of our simulation is
between 853901 and 28432 when the emergence of the first case of COVID-19 case is within the time
interval between September 1, 2019 and November 1, 2020. Overall, the average and median R2 of 155
countries decrease as the date of the emergence postpone; however, a small increase of R2 is observed when
the date emergence of the first case of COVID-19 case is September 22. The deviation of the simulation
result of the deaths from the actual scenario in the eight major epicenters demonstrates a negative trend as
the first case emerges later. A similar pattern can be observed on the date of the first death in these eight
countries, which can be applied to a global scale.
Interpretations
We found that our simulation results can only include the actual number of global deaths caused by
COVID-19 on May 1,2020 inside the 95%PI when the date of the emergence of the first case is between
September 15, 2019 and October 15, 2020; The lowest deviation can be observed on September 22. We
also found a decrease of R2 when we deploy later emergence of the first case; unexpectedly, an abnormal
increase of R2 on September 22 is observed. When focusing on the eight major epicenters, we found that the
number of deaths in these eight countries can only be all included inside the 95%PI when the emergence is
on September 15 and September 22. The deviation of the first deaths demonstrates different patterns
depending on the continent, whereas the global average demonstrates a pattern of normal distribution with
a maximum of 0.2107 on September 22 after the removal of outliers. This could suggest a high likelihood
of the emergence of the first case of COVID-19 be around September 15, 2020 and September 22, 2020.