CONCERN Early Warning System (EWS)
The CONCERN Early Warning System (EWS) is a powerful tool that uses Artificial Intelligence
(AI) to help
doctors and nurses detect when hospitalized patients might be getting worse—
two days earlier than other
warning systems. This early prediction gives healthcare teams more time to take action,
preventing serious
complications. What makes CONCERN special is that it's built on nurses' expertise and real-life
observation
patterns in both regular hospital units and intensive care units. Because of this, it has been
widely
accepted by doctors, nurses, and other care providers. As part of eight years of funding by the
National
Institutes of Health (NIH), CONCERN EWS was developed and evaluated at Columbia University
Irving Medical
Center (CUIMC) and tested in a large clinical trial at CUIMC and Mass General Brigham, which found
that
patients whose care teams used CONCERN had:
✔ 36% lower risk of death
✔ 11% shorter hospital stays
✔ 7.5% lower risk of developing sepsis (a life-threatening infection)
✔ 25% more timely transfers to the ICU when needed
These improvements are largely due to CONCERN's ability to predict problems two days in
advance, allowing
care teams to step in sooner. Unlike many AI-based healthcare tools, CONCERN has been successfully
tested in
real hospitals and proven to work across multiple locations. An early version of the predictive
model was
even independently validated by researchers at the University of Utah, who found it worked
well on data from
over 200 hospitals. Current efforts were expanding the system to pediatric hospitals
and emergency departments, in collaboration with several major medical centers,
including the University of Colorado, Mass
General Brigham, Washington University in St. Louis, and Vanderbilt University Medical Center. During their
eight
years of NIH funding, the CONCERN research team was able to bring their vision from a novel idea
based
on nursing practice to a validated AI-based tool in the hospital helping healthcare teams save lives by
predicting problems before they become critical.
We encourage interested hospitals to join the CONCERN Initiative to receive the toolkit, training, and implementation materials. Please fill out our registration form to receive access to the complete toolkit; there is no charge to join our initiative.
If you have questions, please reach out to the CONCERN team at CONCERN-EWS@cumc.columbia.edu.




