How NewYork–Presbyterian plans to use AI in 2024

The new year is approaching, and that means everyone is writing out their 2024 goals and priorities—including health systems.

NewYork–Presbyterian, one of the largest health systems in New York, has prioritized implementing its artificial intelligence (AI) strategy in 2024, Peter Fleischut, group SVP and chief transformation and information officer at the health system, told Healthcare Brew.

At the core of the health system’s AI strategy is streamlining workflows to make it easier for providers to care for patients, Fleischut said. “It is important for us to make sure that we don’t just implement technology for technology’s sake,” Fleischut said. “A big part of our role is to really understand the business, understand how we’re trying to take care of patients, and make sure we’re applying technology where technology should be applied.”

AI use cases

NewYork–Presbyterian has around 120 AI initiatives across clinical and nonclinical areas, Fleischut said.

One clinical initiative aims to identify diseases earlier.

“There is an unbelievable promise in being able to use data […] and AI to identify patients that may have illness earlier in the course of their treatment,” Fleischut said.

“This algorithm […] can identify people who may be at high-risk, who need an ultrasound,” Fleischut said. “If we can use something like that, we can get people that ultrasound sooner, and ultimately, if they have a heart problem, we can treat them when they’re healthier—which leads to them having a better outcome.”

The goal is to be able to develop similar algorithms for other diseases as well, he added.

One of the nonclinical use cases the health system is experimenting with is automated timekeeping, Fleischut said.

With roughly 40,000 employees, it can be a complex process for NewYork–Presbyterian to track when each employee clocks in and out, he said. Using robotic process automation streamlines the timekeeping process, he said.

“That’s a very basic example, but it’s very impactful,” Fleischut said.

Implementing AI ethically

As NewYork–Presbyterian rolls out its AI strategy, ethics are top of mind. The health system has put in place a process to review all new technologies for cybersecurity concerns and algorithmic biases, Fleischut said.

“Technology is not value-neutral; it can possibly worsen disparities,” he said. “We’ve removed algorithms from our health system where we feel that they were biased in the way that they would treat patients.”

For example, the health system found bias in an algorithm used by some hospitals to identify eligible patients for a lung transplant. The algorithm was trained with data from 100 years ago, when, according to Fleischut, the lung capacity of Black and white patients was “drastically different” due to historic disenfranchisement and slavery.

Using that data could lead the algorithm to have a bias when determining which patients should be prioritized for a transplant, he said, adding that “when we look at AI, we want to make sure it’s being built off of diverse datasets.”

Part of implementing AI ethically is also making sure patients understand when the technology is used during care.

“We want to be transparent about where we are and aren’t using technology,” Fleischut said.

NewYork–Presbyterian has also added training to help providers communicate with patients about how AI may be being used in clinical care, Fleischut said. The health system also test new technology and collects patient feedback before the tech is officially introduced to patients, he added.

“We’re educating people on what we’re doing, why we’re doing it, and having them [be] a part of the process,” Fleischut said.