Should international borders re-open? The impact of travel restrictions on COVID-19 importation risk

Abstract

Background: Novel coronavirus disease (COVID-19) has spread across the world at an unprecedented pace, reaching over 200 countries and territories in less than three months. In response, many governments denied entry to travellers arriving from various countries affected by the virus. While several industries continue to experience economic losses due to the imposed interventions, it is unclear whether the different travel restrictions were successful in reducing COVID-19 importations.

Methods: Here we develop a comprehensive probabilistic framework to model daily COVID-19 importations, considering different travel bans. We quantify the temporal effects of the restrictions and elucidate the relationship between incidence rates in other countries, travel flows and the expected number of importations into the country under investigation.

Results: As a cases study, we evaluate the travel bans enforced by the Australian government. We find that international travel bans in Australia lowered COVID-19 importations by 87.68% (83.39 - 91.35) between January and June 2020. The presented framework can further be used to gain insights into how many importations to expect should borders re-open.

Conclusions: While travel bans lowered the number of COVID-19 importations overall, the effectiveness of bans on individual countries varies widely and directly depends on the change in behaviour in returning residents and citizens. Authorities may consider the presented information when planning a phased re-opening of international borders.

Keywords: COVID-19; Disease modelling; Infectious disease spread; Travel restrictions.

Fig. 1

Estimated daily COVID-19 importations. Our model estimates that a total of 6,003 COVID-19 cases were importations into Australia between 1^st January and 30^th May 2020, considering the current travel restrictions (solid line). Without any travel restrictions a total of 48,715 cases would have been imported during the same time period (dashed line). The shaded are indicated the 95% confidence interval of our estimations that was obtained by averaging over 100 model runs

Fig. 2

Comparison of importations by citizens/residents and visitors before and after the implementation of travel bans. The stacked bar chart shows the estimated number of importations by Australian residents and citizens before (dark blue) and after (dark green) the date of the travel ban. The light blue and light green bars show the estimated importations by visitors before and after the date of the travel ban, respectively

Fig. 3

Estimated percentage reduction of imported COVID-19 cases. The estimated cumulative number of importations by Australian citizens/residents (dashed curve) and visitors (solid curve), assuming no travel bans are implemented. The vertical dashed line indicates the date when the cumulative number of visitor importations reached one. The corresponding label shows the expected percentage reduction in the total number of importations over the studied period if a travel ban had been implemented on the same day. The solid vertical line indicates the implementation date of the actual travel ban and the corresponding percentage reduction in imported cases

Fig. 4

The relationship between daily incidence rates, the number of arrivals and the expected number of importations. We assume that arrivals spent an average of 15 days in the source country. Darker areas of the contour plot indicate a higher number of expected importations. Stars mark the expected number of importations from several countries during October 2020 if borders were to re-open at the start of the month. The results are averaged over 1,000 model runs