The Effectiveness of National Border Closures During the COVID-19 Pandemic

Key Points

1. Targeted border closures, applied in early February 2020, did not slow the spread of the COVID-19 pandemic globally but total border closures, applied in March 2020, did temporarily slow the spread of the pandemic globally.
2. Total border closures were more likely than targeted border closures to slow transmission within countries. Early implementation and targeted closures so extensive that they resemble total border closures were the most likely to be effective.
3. While border closures can temporarily slow pandemic spread, uncertain results and negative social and economic impacts make border closures unlikely to be the best policy response for most countries. They should only be deployed in rare circumstances and with great caution.

Introduction

Each year over 200 pandemic threats emerge, and national governments must evaluate how to respond. One measure used by most countries around the world during the COVID-19 pandemic was the adoption of national border closures. In some cases, these border closures were targeted, banning travel from countries known to have had outbreaks of COVID-19. In other cases, total border closures were adopted, banning non-essential travel from all countries. At the time there was little evidence on the effectiveness of border closures and the evidence that existed mainly relied on mathematical modeling.

This briefing note includes findings and policy implications from a recent study that used real-world data from 166 countries to measure the effectiveness of national border closures on the global transmission of COVID-19. The study provides important information for responding to future pandemic threats, especially for policy makers weighing the benefits of border closures against the significant economic and social costs involved.

Research Findings

After thorough analysis of the data available, we found that the global sum of targeted border closures implemented by February 5, 2020 was not sufficient to slow global COVID-19 transmission, but the sum of total border closures implemented by March 19, 2020 did achieve this effect. These findings can be seen in the graphics and description below.

_{In late January February, a series of targeted border closures began, the figure on the left shows that they did not slow the spread of the pandemic globally (measured by Global Rt). In mid-March, total border closures were enacted and there was a steady reduction in Global Rt which can be clearly seen in the figure on the right.}

While there was a great deal of variation from country to country, early and extensive border closures worked to slow national epidemics temporarily in some countries. Maps on page 3 illustrate the effects of targeted and total border closures at the country level.

The study also found that the larger the number of countries included in the targeted border closures (approaching total border closures) the more likely they were to work. Which countries were targeted (eg. those with reported outbreaks versus those with no reported outbreak) was less important than the number of people targeted overall.

Data Analysis and Methods

Data used in the study:

• Population and known cases targeted by border closures from Oxford COVID-19 Government Response Tracker
• Calculated global & national R_t for 166 countries during the first 22 weeks of the pandemic using Johns Hopkins University’s CSSE

Methods used in the study:

• Interrupted time series analysis evaluated changes in level and slope of 5-day lagged R_t after border closures
• Global intervention points on Feb 5 for targeted and Mar 19, 2020 for total closures
• Country-level analysis performed for 34 targeted closures and 103 total closures
• Meta-regression to identify factors associated with effective closures
• Coarsened exact matching reduces covariate imbalance between countries with and without border closures to allow for regression analysis

Policy Implications

While total border closures can be effective, uncertain results and negative social and economic impacts make border closures unlikely to be the best policy response for most countries. They should only be deployed in rare circumstances and with great caution.

Countries that do not have the resources to implement other less restrictive measures, like mandatory quarantines, may have fewer options. In those cases, completely closing their borders early on can be an effective way to slow down the spread of pandemic threats to buy time to set up stronger public health measures.

The usefulness of national border closures depends on what others do. All countries would benefit from global mechanisms to coordinate national decisions on border closures during pandemics. In the future, the World Health Organization could help national decision makers to coordinate the timing and extent of border closures.

The International Health Regulations (IHR) prevent countries from using travel restrictions that are not supported by science or if there are less restrictive options available. The results of this study imply that most border closures were illegal since most countries had better options than border closures.

Conclusion

This research provides new evidence and guidance for policy makers considering measures to respond to pandemic threats. It shows that total border closures, if applied early, can be an effective tool in delaying the global and national spread of pandemic threats.

Context matters when evaluating the usefulness of border closures. It is important to weigh if the time bought by the introduction of these restrictive measures will be put to good use to prepare the domestic public health response. Countries should also consider if other less restrictive measures, such as mandatory travel quarantines, can be just as effective while imposing fewer social and economic costs. For countries with financial and administrative resources, border closures are unlikely to be the best policy response.