As the world struggles to contain the coronavirus (COVID-19), parallels are inevitably drawn to one of the last major global epidemics: the SARS outbreak in 2003. SARS was a different virus with different impacts, but something else is also very different today: We are all much more interconnected than in 2003.

Almost everyone has a smartphone; many people have personal health-monitoring devices such as Fitbits or smartwatches able to measure their temperature and heart rate. People communicate all the time via social media channels about what they are doing, where they are going and how they are feeling.

This hyper-connectivity creates numerous opportunities to improve the monitoring and control of pandemics like the coronavirus, and there are already some compelling initiatives. In Australia, the Commonwealth Scientific and Industrial Research Organisation (CSIRO) has demonstrated its ability to get very early warning of an epidemic, simply by monitoring Twitter posts.

Back in November 2016, environmental conditions caused by a local thunderstorm created a huge spike in asthma attacks in Melbourne, overwhelming emergency services and hospitals. By 6:00 p.m. that day, more than 8,000 asthma sufferers had been admitted to the hospital; some died.

Using anonymised and publicly available Twitter data, an artificial intelligence (AI) tool developed by CSIRO analysed more than 3 million tweets containing keywords related to asthma, such as “breath” and “coughing” and was able to detect the outbreak up to 9 hours before it was officially reported and before the first news story broke.

One Canadian company, BlueDot, already offers a commercial service providing clients — which include governmental agencies, hospitals and businesses — with early warnings of emerging epidemics, and it flagged COVID-19 before it became a global news item.

 

BlueDot beat WHO on COVID-19 alert

On December 31, 2019, BlueDot alerted clients to the outbreak of a flu-like virus in Wuhan, China, 9 days before the World Health Organization (WHO) released a statement about it.

BlueDot says its global early warning system delivers critical insights on the spread of infectious diseases by combining more than 100 data sets with proprietary algorithms.

These initiatives show there is enormous potential to use real-time personal and other data to track and counter a pandemic, but to be really useful, that information needs to be married intelligently with public health information.

What is needed is a platform able to take data from all these sources and create insights out of it — an endeavour that is challenging due to the heavily regulated nature of the healthcare industry.

Taiwan, however, is credited with leveraging its public-health infrastructure and data analytics to combat COVID-19.

 

Taiwan integrated data sources to counter COVID-19

In an article published in the Journal of the American Medical Association, Jason Wang, an associate professor of paediatrics at Stanford University, described how Taiwan integrated its national health insurance database with its immigration and customs database and applied data analytics to identify likely carriers of COVID-19.

In real time, when citizens reported to a clinic, Taiwan was able to classify travellers’ infection risks based on the point of origin of their flight and subsequent travels over the next 14 days. Those deemed to be most at risk were quarantined at home and tracked through their mobile phones to ensure they remained at home during the incubation period, Wang reported.

Also, he said the Taiwanese government was able to find 113 people suffering from severe respiratory symptoms by searching its National Health Insurance (NHI) database for people who had recently tested negative for influenza and then retesting them for COVID-19.

 

Mining a trove of data

Singapore is also at the forefront of nations that have been most effective in controlling COVID-19. According to a report from the U.S. Centers for Disease Control and Prevention, their government employed multiple surveillance methods that complemented each other, enabling it to identify infected persons. These overlapping detection methods enhanced the success rate: None of the methods alone would have detected all of the patients.

The South Korean government is using a smartphone app developed by the Ministry of the Interior and Safety that tracks the location of people in quarantine using the phone’s GPS capability to make sure they do not break quarantine. Officials said it was intended to help manage the increasing caseload and prevent “super spreaders”, who have been blamed for significant numbers of infections.

South Korea is leveraging data gathered from smart city technologies to aid in its fight against COVID-19. This helps health investigators trying to trace the movements of COVID-19 patients by giving them access to credit card transactions and footage from closed-circuit television cameras.

In the United States, Zencity, a company that runs an artificial intelligence-driven platform to help local governments better understand the needs of their citizens, has analysed the public online conversations of people in more than 100 U.S. cities to highlight their concerns and priorities about COVID-19.

In a post on Medium, Eyal Feder-Levy, the chief executive officer of Zencity, said: “Using topic modelling and clustering, we were able to identify what people are specifically talking about, and by extension, most interested in or concerned with. This information can help local government officials shape both policy and messaging.”

School closures topped the list, followed by cancellation of public events and public transportation.

These are just a few examples of the many ways in which the rich trove of data created by technologies such as smartphones is being mined, communicated and manipulated to better track, understand and counter what is almost certainly the biggest threat to health the world has seen since the Spanish Flu of 1918.