“You get an adrenaline rush racing against the clock,” says Alicia Cowley ’15, Res.’18, MBA ’15 (STERN), about how she spent a weekend last May.
Dr. Cowley didn’t run a marathon or take a timed exam, however. The hospitalist participated in New York City’s first-ever Digital Health Datathon, hosted by NYU Langone Health.
The intense 48-hour competition started by defining and building the new interdisciplinary team for the digital era: clinicians and data scientists working side by side. Colleagues from NYU Langone, Massachusetts Institute of Technology (MIT), and Google worked in groups.
Once the teams were established, providers posed clinical questions that the data scientists turned into computer queries on a large electronic health record (EHR) dataset. Dr. Cowley’s team asked: Do higher daily diuresis (urination) goals decrease ICU patients’ length of stay for heart failure exacerbation?
“Our team broke down what the question meant and the variables that led to the diuresis goal,” Dr. Cowley explains. “Then, we came up with assumptions and criteria. The pharmacists chose patients from the dataset based on medications, and the hospitalist physicians utilized admitting diagnoses. Data scientists then wrote queries to input the values we wanted and used statistics to better determine which patients were more ill or less ill.”
According to Dr. Cowley, the Datathon was a unique opportunity to develop knowledge and skills at the growing interface between technology and medicine.
“I got a 10,000-foot view and started asking why we do something the way we do it, by analyzing massive amounts of patient data,” she says.
According to Sarah MacArthur, MD, Res. ’11, ’15, director of digital health innovation for Medical Center Information Technology (MCIT) and the Center for Healthcare Innovation and Delivery Science (CHIDS) at NYU Langone, the benefits Dr. Cowley reaped from the Datathon are exactly what NYU Langone aims for when exposing medical professionals and students to data science and digital health.
“Physicians increasingly receive diagnosis and management suggestions from automatically triggered clinical decision support tools,” says Dr. MacArthur.
“While these suggestions are often enormously helpful, it’s essential to understand how these tools are built to fully appreciate their promise and pitfalls. In medical school and residency, we’re taught how to tear apart a New England Journal of Medicine article, look at its methodology, and understand where it might be flawed—but we haven’t yet been taught how to analyze the new knowledge created by data scientists’ analyses of EHRs.”
Enter events like Datathon and, immediately following that event, the NYU Langone Health Tech Symposium supported by MCIT and CHIDS. The symposium provided a survey course on artificial intelligence and machine learning in healthcare to more than 500 NYU Langone Health clinicians and medical professionals.
“At NYU Langone, we’re enormously fortunate to have an abundance of national experts in healthcare delivery design, predictive analytics, and clinical informatics—most dual trained,” Dr. MacArthur notes. “How do we diffuse their knowledge and skills throughout our entire institution? We’re empowering the average physician to partner with data scientists to generate new knowledge and think about it critically.”
Understanding Virtual Health
The school’s commitment to teaching innovation can also be found in a two-week elective at NYU Grossman School of Medicine that exposes medical students to the burgeoning field of virtual health and empowers them to educate NYU Langone’s leadership about new ideas for potential implementation.
As one of the nation’s first electives in virtual health, Telehealth—Technology Transforming Healthcare Delivery introduces undergraduate and graduate medical students to NYU Langone’s current virtual health practices, including video urgent care visits and online follow-up patient appointments for the Medical Weight Management Program.
But Dr. MacArthur, who created the course, notes that its core purpose is not merely to observe; ultimately, it’s to help drive a technology-enabled redesign of healthcare delivery.
“We give the learners several questions that come straight from administrative leaders, or they may choose their own question—and they present their findings and recommendations directly to hospital leaders,” she says.
One such leader, Viraj Lakdawala, MD, clinical associate professor in the Ronald O. Perelman Department of Emergency Medicine, medical director of Virtual Urgent Care, and system chief for Emergency Medicine Telehealth at NYU Langone, recalls how his team harnessed information that virtual health elective students shared.
“One of the presentations we acted upon was about determining provider competency when it comes to telemedicine visits,” he explains. “Seeing a patient via telemedicine is a much different skill set from seeing a patient in person. We are working with our colleagues in internal medicine who have expertise in using simulations with standardized patients [actors] to develop a program to measure and determine competency in certain core telemedicine skills.”
In this way, the virtual health elective ends up benefiting both the student and the institution.
“The students’ efforts get embedded because they’ve worked on something that the institution would like to understand more deeply,” Dr. MacArthur says. “This is all about novel delivery design. We are driving forward and we want to create new models to generate knowledge.”