(Editor's note: This is the second of a two-part interview. To read the first part, click here.)
Dr. Bruce Darrow, chief medical information officer and interim chief digital information officer at the Mount Sinai Health System in New York This week I shared some thoughts (emphasis on “some”) about why artificial intelligence is getting so much attention in healthcare, and why AI may one day ease the burden on doctors.
In this Q&A, Darrow shares more about how Mount Sinai uses AI and how they plan to expand their use. He discusses how long the health system has been using AI in clinical care, the principles clinical and IT leaders follow when considering clinical AI use cases, and Mount Sinai's current AI deployments. The video that accompanies this article also discusses the factors that will determine whether AI initiatives at Mount Sinai will be successful.
Q. How long has Mount Sinai been using AI in clinical care and where is it being used in the health system?
A. You might think of the use of AI as a relatively recent development. It depends on where you draw the line and define it.
We've been using algorithmic care and computer-based decision support for many years, but the first real-world application of AI was in 2013. At Mount Sinai, over 10 years ago, the first use case we reported or published at that time was using AI algorithms to identify patients in the hospital who were likely to become seriously ill and have poor outcomes.
And with the help of AI, early detection has significantly increased the chances of surviving hospital stays for over a decade now.
A lot of the work that we've done at Mount Sinai over the last 10 to 12 years has been in the area of what's called predictive AI: finding patients who are likely to get sick, who are likely to have conditions that would benefit from having that knowledge, and then bringing in the right skillset and expertise and getting that patient the right treatment earlier in the process.
Over the last year or two, we've been looking at how we can use AI to streamline care, not necessarily specifically in clinical care, but to make patient care easier, to make the operational parts more efficient, and also to automate and pre-plan and draft some of the time-consuming tasks that physicians and other members of the care team do.
Q. What principles does Mount Sinai follow when considering clinical AI use cases?
A. This is really important because, as I mentioned, we've been using AI for over 10 years, but over the last two or three years, it's become clear that AI is going to be an increasingly large part of our patient care portfolio, and we knew we needed to be purposeful about how we were using it.
The principles we stuck to at Mount Sinai were that the use of AI in clinical care should be safe, effective, equitable, and ethical. We need tools that are not only safe and effective, but that make a difference in patient care. They have to work. They have to serve some goal: to move care forward.
Being ethical and fair in how we ensure we provide these tools to all patients in a way that is consistent with our mission as an organization.
Q. What AI use cases is Mount Sinai currently implementing?
A. Most of the AI we use essentially comes from three different pipelines: At Mount Sinai, we have an incredibly talented and passionate team of data scientists, implementation scientists, artists, and other team members who use our learning platform, data pipelines to create and test our own AI algorithms, which can then be used in patient care.
They have published extensively and been recognized for it. David Rich, president of Mount Sinai Hospital, and Robbie Freeman, chief nursing information officer and vice president of Digital and Technology Partners for Innovation, are very active in their respective teams.
Examples include identifying patients before they become sick enough to require ICU treatment, identifying whether hospitalized patients are at risk of falling with greater accuracy than existing tools, and identifying patients at risk of malnutrition or pressure ulcers and notifying the appropriate members of their care team.
These are a wonderful complement to the care our nurses, doctors, social workers and dietitians already provide to patients within the hospital.
We have a lot of homegrown knowledge and expertise and have been doing it basically since about 2016. Over the last five years or so, Imaging AI is on the rise. These are all FDA-approved tools and software algorithms that can be used with patients.
As we discussed yesterday, many of these do not replace radiologists or clinicians, but they do make radiologists more accurate, efficient, and fast. To give you an example, imagine you have 20 patients who have had a head CTS (computed tomography of the head) to look for abnormalities such as stroke or bleeding in the head.
If a doctor sees a list of 20, they might not understand it. They might be in the order that the images were taken. But if there's some AI running in the background that's saying, out of these 20, look at these two first because the algorithm says these two are most likely to be abnormal, that's good for the doctor.
It's also better for the patient because the right tests are brought to our attention first, and if we determine that they may make a difference in treatment, they can get treated sooner. There's a fair amount of imaging AI being used to improve the accuracy of our diagnoses and to help us make the right choices about where to focus our attention.
The third area where we see a lot of AI is in existing software and tools provided by other software providers in the community. Nearly all of the software we use at Mount Sinai has AI built into it, if it isn't already, and we expect it will be built into it over the next three to five years.
This is the way technology is advancing. Our electronic health record systems will have AI built into them, we will review it, validate it, and decide whether to incorporate it into our care. Everything we use, from emails to presentation documents to video collaborations, will have elements of AI built into them.
Bonus content: Click here to watch a video of this interview, which also features Dr. Bruce Darrow discussing what will determine whether AI initiatives at Mount Sinai will be successful and what his colleagues at other hospitals and health systems can learn from this.
Editor's note: This is the ninth installment in a series featuring leading voices from the healthcare IT world discussing the use of artificial intelligence in healthcare. To read the first installment, featuring Dr. John Halamka of Mayo Clinic, Click here To read the second interview with Dr. Aalpen Patel from Geisinger, click here To read the third interview with Helen Waters from Meditech, click here
In the fourth installment, we'll read with Sumit Rana from Epic. Click here to read part five with Dr. Rebecca G. Misulis of Massachusetts General Hospital Brigham; click here to read part six with Dr. Melek Somai of the Froedert School of Medicine Health Network of Wisconsin; click here to read part seven with Dr. Brian Hasselfeld of Johns Hopkins Medicine; and click here to read part eight with Craig Kwiatkowski, senior vice president and CIO at Cedars-Sinai.
The HIMSS AI in Healthcare Forum is scheduled to take place in Boston from September 5 to 6. Learn more and register.
Follow Bill's HIT articles on LinkedIn: Bill Siwicki
Email: bsiwicki@himss.org
Healthcare IT News is a publication from HIMSS Media.
