Sufficient strategies for travel quarantine and testing

Wells, C. R.; Townsend, J. P.; Pandey, A.; Fitzpatrick, M. C.; Crystal, W.; Galvani, A. P.

A major policy tool used by governments during the COVID-19 pandemic has been the imposition of quarantine and testing requirements on international travel, extending even to the extreme approach of closing national borders. However, both approaches have far-reaching socioeconomic implications. Therefore, we evaluated the implications of quarantine and testing compared to the reductions achieved by border closure on in-country imminent transmission based on country-specific prevalence, daily incidence, vaccine coverage, immunity, age-demographics, and travel flow. We considered travel quarantines of 0-14 days with and without RT-PCR or antigen testing for COVID-19 in 31 European countries. Our analysis demonstrates that for the vast majority of origin-destination country pairs, there are combinations of short-duration quarantine and testing that are as effective as border closure. Furthermore, for most origin-destination country pairs, travel with testing and no quarantine will reduce imminent in-country transmission more than would border closure. We find that the duration of quarantine is predominantly influenced by country prevalence and quantity of travel. With higher prevalence in the origin country compared to the destination country, the minimum duration of quarantine increases from zero to beyond fourteen days (at which point border closure would likely be practical). Asymmetry in travel flow can also cause an increase or decrease of infections within the destination country, respectively resulting in longer and shorter quarantines. We apply the same framework to genetic variants of concern to limit their spread as a consequence of travel; the widespread variant of concern B.1.1.7 yields similar sufficient quarantine and testing regimes, whereas the lower-frequency, geographically heterogeneous 501Y.V2 variant requires longer, more specific quarantines. We show that adaptation of our analytical framework to the European Union traffic-light country risk stratification provides a simplified policy tool. Lastly, we examined the effect of travel quarantine and testing or border closure policy on hospitalization in the destination country. We find that hospitalization rates in the destination country are far more sensitive to changes in the vaccine coverage than on the duration of the travel quarantine or the country of origin. Our analysis provides rigorous guidelines enabling travel between most countries during early and late phases of pandemic disease.