(upbeat music) - Hospitals are producing so much data that it is extraordinarily difficult to provide the analytics of it so that we can actually benefit patients going forward. - We're collecting a lot more data on these patients, but we're only scratching the surface of how that can be used in the sense of decreasing critical illness. There are always better ways for us to improve the outcomes.
- And so, artificial intelligence allows us to take information from tens of thousands of patients in order to predict what's going to happen to this particular patient and to prevent it from happening. (lively electronic music) The Bronx is a very interesting community of people. It is the most vibrant urban environment, I think, in the United States.
But it also happens to be the poorest urban county in the country. For a system like Montefiore's, we have to be able to deliver care extraordinarily efficiently, because the amount that we get paid to do that is very, very low relative to the costs of delivering that care. - The reality is patients in the hospital are much sicker than 10, even 20 years ago.
The most severe type of respiratory failure is ARDS and upwards of 40% of the patients with ARDS may not survive the hospitalization. These patients often need ventilators. They often need to be in the ICU for long periods of time.
What we wanted to be able to do was to find a way of detecting patients who may be at high risk of developing respiratory failure. - The way to do that is the use of data that exists in thousands and thousands of patients to predict who's at risk and to intervene to prevent it from happening. - I'm finding patterns in the data, finding patterns that otherwise are much harder to find, but we're using the data to tell us where those objects are, instead of trying to look for them manually.
(upbeat rhythmic music) - We are using machine learning and artificial intelligence to build the system that's always watching. We call it PALM, Patient Centered Analytic Learning Machine. Our vision is to present PALM as the nervous system of this institution.
- So, it's tricky to detect respiratory failure. Respiratory failure can develop slowly and insidiously. - So what we have designed PALM to do is that it collects different kinds of information.
- Someone's heart rate, their blood pressure, their oxygen saturation, changes in their physiologic vital signs on a second-by-second or minute-by-minute basis has implications for where they are going. - PALM's job is to calculate the risk for the patients to have either respiratory failure or whether or not this patient is at risk of dying. - There is an overwhelming amount of data that we have to deal with in healthcare.
- The only way we're gonna be able to analyze it is through the use of machines. - So we have found a lot of reasons to work with Intel because they want to not only imagine the future but also shape the future of healthcare. We are not only using their engineering powers today, we are also using the cluster of servers as underpinnings of our technology.
It's also having a partner to imagine what is possible in the future. Our algorithm outperforms any similar model that exists. Not only we ended up with an algorithm that can predict respiratory failure, but also it can predict sepsis.
We have data that shows that is has decreased the length of stay overall as well as within the ICU. We see it as the future of healthcare. (lively music) Our work, it is all about providing the best care possible to the Bronx population.
- The thing that's gonna distinguish healthcare institutions that are extraordinary from other kinds of healthcare institutions are gonna be the ability to recognize the information that they're generating and to use that information productively for patient care. The ultimate goal of this kind of technology is going to be in, essentially, the cost and the efficiency and the safety of the care that we deliver. - Early treatment is really key in improving outcomes.
What excites me about this project is its possibilities in the sense of decreasing critical illness. What we wanna be able to see is patients continue to live their lives the way that they want to. - Our job at Montefiore is to pioneer so that others can take a look at what we've done and say yes, I can think of ways to apply to things that Montefiore hasn't even done yet.
This is the next evolution of where we're going. If you can provide the information to prevent an adverse outcome, you'd be foolish not to be pursuing that avenue.
