AI-assisted tracking of worldwide non-pharmaceutical interventions for COVID-19

Abstract

The Coronavirus disease 2019 (COVID-19) global pandemic has transformed almost every facet of human society throughout the world. Against an emerging, highly transmissible disease, governments worldwide have implemented non-pharmaceutical interventions (NPIs) to slow the spread of the virus. Examples of such interventions include community actions, such as school closures or restrictions on mass gatherings, individual actions including mask wearing and self-quarantine, and environmental actions such as cleaning public facilities. We present the Worldwide Non-pharmaceutical Interventions Tracker for COVID-19 (WNTRAC), a comprehensive dataset consisting of over 6,000 NPIs implemented worldwide since the start of the pandemic. WNTRAC covers NPIs implemented across 261 countries and territories, and classifies NPIs into a taxonomy of 16 NPI types. NPIs are automatically extracted daily from Wikipedia articles using natural language processing techniques and then manually validated to ensure accuracy and veracity. We hope that the dataset will prove valuable for policymakers, public health leaders, and researchers in modeling and analysis efforts to control the spread of COVID-19.

Fig. 1

AI-assisted approach to build the Worldwide Non-pharmaceutical Interventions Tracker for COVID-19 (WNTRAC) dataset.

Fig. 2

An example of the NPI mentioned in the Wikipedia article of 15^th May 2020.

Fig. 3

The WNTRAC system consisting of a processing pipeline, WNTRAC Curator validation tool and NPI data browser.

Fig. 4

Updated NPI events from a Wikipedia article that are automatically extracted daily by the system are shown to a human volunteer for validation and correction in the WNTRAC Curator tool.

Fig. 5

NPI data browser for visualizing the Worldwide Non-pharmaceutical Interventions Tracker for COVID-19 dataset.

Fig. 6

Distribution of NPIs in the Worldwide Non-pharmaceutical Interventions Tracker for COVID-19 dataset.

Fig. 7

Number of regions implementing each NPI globally and within the United States.

Fig. 8

Distribution of NPI implemented in different geographies globally and within the United States.

Fig. 9

Number of imposed and lifted NPIs per month.

Fig. 10

Elapsed time (in days) between the introduction of NPIs and recording of first death (red) or 50 cases (blue) in countries that implemented travel-related vs community-related NPIs first. A negative value implies that the first travel or community NPI was imposed prior to the first death (or 50 confirmed cases) while a positive value means that the NPI was imposed after such an event. For example, Israel, Germany and France imposed a travel related NPI over a month before reaching 50 confirmed cases or first death, while in USA-NY such NPI was imposed only 10 days after the first death. In most countries the first NPI was imposed before reaching 50 confirmed cases and before the first death occurred.

Fig. 11

Example of the Compliance-Score for different levels of anticipated mobility, where a lower anticipated mobility (e.g., m_ant = −10%) is derived from a lower stringency index, and higher anticipated mobility (e.g., m_ant = −70%) is derived from a higher stringency index. A Compliance Score of 0 refers to a poor level of public compliance to the NPIs imposed for a given location, and a score of 1 implies that the public is complying to the NPIs imposed given the stringency index for a location.

Fig. 12

Trends in COVID-19 cases per 100,000 population since first 50 cases and the NPI-based indices in representative US states. At the beginning of the pandemic the 4 states shown have similar trends in the NPI index values. As the pandemic progressed, particularly by October 2020, there was as decrease in the NPI index values that appears to be associated with a subsequent increase in new COVID-19 cases.