Dr. Bruce Darrow, chief medical information officer and interim chief digital and information officer at Recent York’s Mount Sinai Health System offered some thoughts this week on why AI is playing such an crucial moment in healthcare and how AI may one day be taking some (emphasis on “some”) cases away from doctors.
In this Q&A, Darrow takes a closer look at how Mount Sinai is using artificial intelligence and plans to expand its exploit. It discusses how long the health care system has been using AI in clinical care, what principles its clinical and IT leaders are guided by when considering clinical exploit cases for AI, and discusses AI implementations currently in exploit at Mount Sinai. In the video accompanying this article, Darrow also describes what determines whether an AI initiative at Mount Sinai is likely to succeed.
Q. How long has Mount Sinai been using AI in clinical care, and where is the health system using it?
AND. You might think that the exploit of artificial intelligence is a newer development. It depends on where you draw the line and where you make the definition.
We have been using algorithmic care and methods of using computer decision support for many, many years. The first real exploit of AI was in 2013. That was over 10 years ago at Mount Sinai, where at that point the first exploit case we reported or published was using AI algorithms to find patients in a hospital who were likely to get very diseased before they got to this point and their results were worse.
By using artificial intelligence to get them into care earlier, we were able to significantly boost their likelihood of surviving hospitalization. So it’s been over 10 years.
A lot of the work that we’ve done at Mount Sinai over the last 10 or 12 years has been in the area of what I would call predictive AI, which is finding patients who are likely to get the disease, finding patients who are likely to have this condition that would benefit from having this knowledge, the right skill set, the right expertise and providing the patient with the right treatments earlier in the process.
Over the last year or two, we’ve been looking at ways to exploit AI to improve care, not necessarily related to clinical care specifically, but ways to make care easier for our patients, streamline the operational parts, and also start to automate some of the things that physicians and other members of the healthcare team do , and which take a lot of time and can be developed and the preparatory work done for them.
Q. What principles does Mount Sinai exploit when considering clinical exploit cases for AI?
AND. It is very crucial. As I said, we’ve been using AI for over 10 years, and we’ve found over the last two or three years that as it’s become clear that AI is going to be an increasing part of our patient care portfolio, we need to be intentional about it. how we would exploit it.
At Mount Sinai, we have held to the principles that the exploit of artificial intelligence in clinical care should be safe and sound, effective, fair and ethical. Sheltered and effective Of course, we need to have tools that will transform patient care. They have to work. They must serve a purpose that will improve care.
Ethical and equitable in how we provide these tools to all of our patients in a way that is consistent with our mission as an organization.
Q. What AI exploit cases currently exist at Mount Sinai?
AND. Most of the AI we exploit comes from basically three different pipelines. We are fortunate that at Mount Sinai we have a very talented and committed team of data scientists, implementation scientists, artists, and other team members who can exploit the learning platform, data pipeline to create, test, and leverage their own AI algorithms over our patients.
They published many publications and were appreciated for it. David Wealthy, president of Mount Sinai Hospital, and Robbie Freeman, who is our chief nursing information officer and vice president of Digital and Technology Partners for Innovation, have been very dynamic with their teams.
Some examples include: finding patients before they become ill enough to need ICU care, determining with greater accuracy than existing tools whether a hospital patient is at risk of falls, identifying patients who are at risk of malnutrition or pressure ulcers so that we can bring it to the attention of the appropriate members of the care team.
These are a perfect complement to the care that our nurses, doctors, social workers and registered dietitians already provide to our patients in the hospital setting.
We have a lot of our own knowledge and experience and we have been doing this basically since around 2016. In the last five years we have observed growing amount of imaging artificial intelligence. These are all FDA-approved software tools and algorithms that we can exploit for our patients.
Many of them do not replace, as I said in yesterday’s discussion, the radiologist or clinician, but they make the radiologist’s work more correct, more capable and faster. One example is if you imagine that there may be 20 patients with CTS of the head, which is a computed tomography scan of the head to detect abnormalities that may include a stroke or bleeding within the head.
If a doctor goes through a list of 20 of them, he may not know. They can be ordered by the date the photos were acquired. However, if you have AI running in the background and it says: out of these 20, look at these two first because the algorithm says there will probably be something in them that looks abnormal. This is good for clinicians.
They pay attention to the right research first, and that’s good for patients because they get care faster when we think it could make a difference in their care. There’s a lot of artificial intelligence available for imaging, both for diagnostic accuracy and just making sure we have the right selection of places to look at.
The third area where I see a lot of AI is tools provided by our existing software or other software vendors in the community. Almost any software we exploit at Mount Sinai, if it doesn’t already have AI built in, I can expect to have AI built in within the next 3-5 years.
Technology is just going this way. Our electronic health record system has artificial intelligence built into it, which we consider, approve and decide whether to place a patient in care. Quite simply, everything from emails to presentation documents to video collaboration that we exploit will have elements of artificial intelligence.
BONUS CONTENT: Click here to watch a video of the interview, in which Dr. Bruce Darrow also discusses the factors that determine whether Mount Sinai’s artificial intelligence initiative is likely to succeed and what his colleagues in other hospitals and health systems can learn from it.