From interactive real-time maps to advanced prediction algorithms, scientists are using the technology to predict and track the spread of the new coronavirus. Solutions include the use of artificial intelligence to process a large amount of data from multiple sources in different countries.
“There is a tremendous amount of data online and discussions often do not come from government channels, but through news and chat rooms and blogs and social media,” said John Brownstein, head of innovation and professor at Harvard Medical School, who is researching the predicting patterns of flu epidemics and pandemics. “We need to sift through that information, classify it, sift through the noise, geocode it, and then picture emerging diseases.”
Several companies and organizations have put together tools to detect the coronavirus. Scientists say that data management platforms and artificial intelligence methods together can lead to a mass analysis of infectious diseases, which in turn can support government decisions. Still, this poses challenges to scientists as global data standards have not yet been developed.
“Modeling and outbreak data analysis efforts typically occur in silos with limited method and data communication between model developers and end users,” said the authors of a Nature study on epidemic modeling. “Modeling cross talk between stakeholders within and between countries is also typically limited, often in a landscape of legal and ethical uncertainty.”
Nine days before the World Health Organization warned about the new coronavirus, a Toronto-based startup saw the threat of COVID-19, the disease caused by the virus. BlueDot, a $ 9.5 million company, used data from hundreds of thousands of sources, such as statements from official public health organizations, media, health reports, and airline demographics. The company relied on its co-founder's expertise in treating SARS patients as an epidemiologist and physician.
“Our world is changing rapidly, and as a result, outbreaks are increasing that threaten human health, safety and prosperity,” Kamran Khan, founder and CEO of BlueDot, said last August.
Meanwhile, Brownstein's HealthMap, a Harvard Medical School-developed artificial intelligence technology that follows infectious diseases, also picked up early signs of the coronavirus spread in Wuhan, China, in December.
“(The System) provides an early identification but we are now using the technology to build that situation image of what emerges, the mining data sources and trying to build a global map of cases that can be used to model and to predict, spread impact on different demographic tools, “he said in a conversation.
Other startups, such as WeBank in China, have taken AI even further and are using it to predict the upcoming recovery of the Chinese economy from the crisis caused by COVID-19.
In particular, technology and artificial intelligence can be applied to emergency medical services, using algorithms to predict the spread of a particular disease, monitor patients, and emergency department operations. Scientists around the world can use algorithms to classify and cluster information, search the Internet and interpret language related to the disease being followed, analyze images, and deploy robotics for a range of medical tasks that physicians do not would feel safe to run.
“In this era, there is a lot of discussion about innovations to develop new solutions, especially from developing countries, AI is likely to benefit from EM (emergency medicine) through its digitization capability and information storage,” wrote two medical researchers from Pakistan and Oman in a report.
Still, using AI in healthcare and health alerts for the coronavirus isn't free from challenges, Brownstein says. The main challenge is to produce a compelling and reliable global dataset on the virus and its spread drawing from multiple national datasets that can have different metrics.
“There are different structures and taxonomies, ways in which people refer to the disease, different languages, cultural context, terms people use around the disease that could be used for other types of things, so a lot of filtering is needed and it requires considerable amount of system training to get to the point where these tools become valuable “.