Exploring Untapped Data in COVID-19

April 13, 2020

Nicholas Tatonetti, PhD, and collaborators are uncovering new insights in COVID-19 gleaned from untapped data.

Dr. Tatonetti, a member of the Herbert Irving Comprehensive Cancer Center (HICCC) at NewYork-Presbyterian/Columbia University Irving Medical Center and associate professor of biomedical informatics at Columbia's Vagelos College of Physicians and Surgeons, has teamed up with colleagues in bioinformatics, urban policy and affairs, and infectious disease on a new data-collection app and website for tracking the impact of the COVID-19 pandemic in New York City.

CovidWatcher, which launched on April 12, works as a crowd-sourcing research platform that surveys users--both healthy and potentially exposed to COVID-19--about their exposure to the virus, symptoms, and access to medical care. The data will be used not only as a decision tool to advise patients if they should seek help but also to enable healthcare officials to deploy resources at New York City “hot spots”, or where they are needed most. The ultimate goal is to prevent future spread.

Since the COVID-19 outbreak, New York City has been labeled an epicenter, hit hard with mounting confirmed cases and patient deaths due to the novel coronavirus. The City of New York, notes Dr. Tatonetti, and Columbia researchers have been partnering to help problem-solve the lack of available COVID-19 related data.

“There has been no ability yet to track and obtain this kind of data from New Yorkers,” says Dr. Tatonetti, who also is a researcher in the HICCC's Precision Oncology and Systems Biology program and directs clinical informatics at Columbia’s Institute for Genomic Medicine. “Our live surveys, for instance, could help find out where New Yorkers are in the city that are having symptoms, even before they may or not may not become infected with the virus. Then, we can redistribute this data to hospital sites to help them prepare for what’s coming their way.”

CovidWatcher principal investigators (PIs) include Noémie Elhadad, PhD, (biomedical informatics); Olena Mamykina, PhD, (biomedical informatics); and Jason Zucker, MD, (medicine), with co-PIs Suzanne Bakken, PhD, RN, a member of the HICCC (nursing and biomedical informatics) and Ester Fuchs (public affairs and policy).

Dr. Tatonetti’s research uses advanced data science methods, including artificial intelligence and machine learning, to better understand complex biological problems. One of the main interests of the Tatonetti lab is making drugs safer, which has an incredible impact on patient outcomes, including cancer patients. Dr. Tatonetti and his team use electronic health records and genomics databases to assess drug toxicities and investigate molecular pathways involved in the adverse effects of drugs.

Like many scientists, Dr. Tatonetti has pivoted his research efforts to address the rising stakes of COVID-19. In addition to co-developing CovidWatcher, Dr. Tatonetti has also participated in the COVID-19 virtual study-a-thon to help inform healthcare decision-making in response to the global pandemic. Hosted by the Observational Health Data Sciences and Informatics (OHDSI) international community, more than 330 people from 30 nations registered for the virtual event. The group now has begun studies on an international set of observational health databases (including insurance claims and electronic health records) to aid decision-making during the pandemic. One project launched from the study-a-thon is looking into the overall safety profile of hydroxychloroquine, a somewhat controversial drug currently being evaluated as a potential treatment for COVID-19. The study is being executed across a database of 130,000 patients from the U.S., England, Germany, and South Korea.

Dr. Tatonetti and other faculty across the HICCC and the University have been quickly transforming their day-to-day research to combat COVID-19. The collaborative energy from the scientific community has been “incredible to witness,” he says, underscoring that right now, research is moving quickly but there is still very much a sense of a shared vision and goals. Indeed, big data is key.

“Data science has to play a supporting role in all of this; it’s not going to be the hero of this story,” says Dr. Tatonetti. “The real heroes of COVID-19 are those who are risking their lives to take care of patients. But data is helping us point physicians and clinicians in the right direction. Instead of shooting in the dark, we’ll have the data to help them make more informed decisions. That is where data can be extremely powerful.”