Prediction Can Help Us Be Better Prepared for the Next Pandemic

abstract

Power of Ideas

Prediction Can Help Us Be Better Prepared for the Next Pandemic

Author(s)
Igor Tulchinsky
Igor Tulchinsky
(Founder, Chairman, and CEO, WorldQuant)
Christopher Mason
Christopher Mason
(Associate Professor, Weill Cornell Medicine; Director, WorldQuant Initiative for Quantitative Prediction)

The COVID-19 pandemic will be a force for technological progress in certain areas, much as World War II was in aeronautics, communications, computing, electronics, and nuclear physics. Techniques for rapid and accurate testing are undergoing a revolution in this extraordinary period, and the pandemic has unleashed powerful incentives to more effectively predict not only the ebb and flow of the virus itself, or the dynamics of the disease, but how it is reshaping human behavior.

The two of us have long shared a fascination with prediction; it’s a bridge between our separate worlds: quantitative finance and medical genomics. This is truly the Age of Prediction, driven by the exponential growth of data, increases in computing power, and advances in technologies such as artificial intelligence and machine learning, spanning both of our disciplines. We have seen the impact in our individual fields: the growth of quantitative finance with its predictive algorithms, the genomics revolution with its ability to rapidly sequence DNA and burgeoning ability to reveal risks of disease or health, and the explosion of machine learning in areas ranging from insurance and forensics to education and politics. 

The tools and technologies being created and harnessed for today’s response can be the foundation upon which we are better prepared in the future.

Prediction is not static; it responds to risk. Models are never perfect; however, many of the problems with the response to COVID-19 stem not from the models but from the data and how the public responds to that data.

We have seen many innovations tied to new technologies that provide an early warning system for COVID-19. Early in the pandemic, a so-called fever map generated by Kinsa, the maker of a popular digital thermometer, provided some of the first indications of community spread outside epicenters on the West Coast and Northeast. Kinsa was able to access thermometer readings from users across the US, suggesting that a surge of higher temperatures in the Southeast might indicate nascent COVID-19 spread.

A variation on that theme, known as wastewater-based epidemiology, is testing sewage for the presence of the COVID-19 virus. The project began on college campuses: The University of Arizona detected a dormitory outbreak through wastewater testing, allowing the school to locate the problem after two students tested positive. The technology has been adopted by other colleges eager to bring back students safely and by some water treatment systems in municipalities lacking large public health systems. 

More ambitious yet is the global MetaSUB (Metagenomics and Metadesign of Subways and Urban Biomes) program. Starting in New York City in 2016, the MetaSUB International Consortium—a group of genomics experts, microbiologists, and data scientists—launched a regular sampling of public transit systems around the world to establish baselines of DNA and RNA, particularly for various pathogens, in cities and neighborhoods. When the novel coronavirus hit, MetaSUB participants turned to specifically seeking traces of COVID-19 RNA, swabbing and taking air samples in more than 100 cities, creating a database that can be used to monitor the disease’s spread and contraction. (As part of MetaSUB, WorldQuant and Weill Cornell Medicine collaborated on a research project at the 2018 Milken Institute Global Conference, sequencing and analyzing the DNA on the smartphones of more than 1,000 attendees to further the study’s goal of developing new antibiotic-resistant bacteria treatments.)

But early detection and prediction is only one area of innovation in a rapidly evolving field. One of the challenges of the pandemic has been to understand how it’s reshaping perceptions and behavior. Using diverse data from a variety of public and private sources—epidemiology, macroeconomics, industry, and company statistics—building predictive algorithms, and running them through machine learning programs can broaden and deepen prediction dramatically. 

We are beginning to tackle questions about how the public is responding to social distancing and quarantines, and what effects that will have on businesses such as supermarkets, restaurants, and digitally driven delivery services. How does a crisis like this, with its isolation and lockdowns, alter economic and social behavior? And how does that reshaped behavior affect individuals, businesses, markets, and governments? In short, what lies beyond the next curve? 

The sheer explosion of data offers more opportunities. Another coronavirus-like event with a significant global impact will likely emerge, driven by a new virus or by some other cause. If the pandemic teaches us anything, it’s that better prediction of potential disruptive events and their cascading effects can strengthen our collective response across society, medicine, and economies. The tools and technologies being created and harnessed for today’s response can be the foundation upon which we are better prepared in the future.

Published September 30, 2020