As part of a new interview series with healthcare leaders across the country, the Olive team had the chance to interview Dr. Aziz Nazha about the biggest challenges facing healthcare today and the growing potential of AI. Previous to his current role as Director, Center of Clinical Artificial Intelligence at Cleveland Clinic, Dr. Nazha completed a hematology and medical oncology fellowship, also at Cleveland Clinic, and a leukemia fellowship at The University of Texas MD Anderson Cancer Center where he served as an instructor in the Leukemia Department.
Q: To begin, we know that you are currently leading the Center for Clinical Artificial Intelligence at the Cleveland Clinic. Can you give us a high-level overview of what you’re working on there?
We launched the center in March of 2019 and the whole mission of the center is to harness the power of artificial intelligence in healthcare, but really we envision the center becoming a hub of collaboration between academia and industry to bring the best of technology to healthcare to build models that can actually make a difference in our patients lives and in medical research. That was the whole purpose of the center – we built a platform for collaboration both inside and outside of Cleveland Clinic.
I’m a physician – I treat patients with cancer and my specialities are leukemia, particularly myelodysplastic syndromes (MDS), but also a programmer because we program all of our models in house to develop what we call physician data scientists – physicians who actually see patients in a clinical practice and are able to understand code and develop the machine learning and deep learning models. My team now consists of medical students, residents and fellows who we’ve taught to code and use the technologies of everything we develop. At the moment we have 24 ongoing projects, 12-13 in cancer space and others in medicine, medical operations, ICU, some genomic projects, as well as new conversations about cardiology projects – we’re continuously expanding and adding more projects.
Q: Can you tell us a little bit about the new course at Cleveland Clinic’s Lerner College of Medicine that’s focusing on integrating artificial intelligence into the curriculum?
As you know and experience yourself [at Olive], AI talent is difficult to find. And if you talk about people that understand the complexities of healthcare data compared to other industries – that’s extremely difficult to find. Why is healthcare behind in the adoption and application of AI? The simplest answer is you have individuals that don’t speak the same language trying to understand each other. So, the biggest problem for AI today in healthcare is that you have computer scientists and statisticians who can look at models, understand the models but don’t necessarily understand the clinical implications of that. And they’re speaking to healthcare providers, physicians or nurses for instance, and they understand the clinical implications but don’t understand the algorithms. The whole purpose of what we’re doing is to bridge this gap and have people speak the same language. And of course, the best way to find this talent is to build it – so, it’s why we put out these 3 courses.
[To build this talent] you really have to start from Medical School – we want to help students understand the technology and most importantly use it in their work. So, these courses teach students about python, machine learning, deep learning, and of course all applications for healthcare. The last module is emerging technologies, cloud computing, internet of things and their application in healthcare. And really we have brilliant students that continue to blow me away by what they can do, and say they’ll consider using this in their research throughout medical school.
Q: Although you’re focused in the clinical space, how do you see artificial intelligence playing a role in other functional areas?
The applications of AI are widely used, and the center lies under what we call enterprise analytics: the financial arm, and also the operational arm which is medical operations. I think there’s always an intersection between all these applications. When I think about AI, AI becomes a tool that gets me to where I need to go, and most of the time – or actually all of the time – if I’m posed with a question, and if I can’t answer that question using linear algebra or traditional statistics, our focus is solving these problems with AI.
One of the projects which will be helpful in the future is around no-shows, for example and building models around no-shows. Can we learn something from AI about why a patient doesn’t come back? If we take the clinical data, we can learn if a patient showed up, or did not show up. What we really need to do is take that data and compare it to other data like a baseball game, or a football game for instance, especially in Cleveland, and that will give you much better of an idea on why they didn’t show because previously you were missing other parts of information. The opportunity for AI is huge. The problem is people using this technology in the wrong way or overselling the technology, that’s a great example of IBM Watson, and we don’t want that. Taking these models and making the data explainable, focusing on explainable AI, so physicians can easily understand and adopt – I think that’s the whole key for the success of AI in healthcare.
Q: If you could eliminate one of the healthcare industry’s challenges overnight, which one would it be and why?
Anytime I give a talk about AI in healthcare, I always leave the audience on this note, “The lightbulb was not invented by continuous improvement of candles.” So, in order for us to really advance healthcare, we need to completely change the way we think about healthcare and how we do research today, and everytime we say that, people get excited. But to actually do that is really, really difficult. That’s the challenge – change gets people out of their comfort zone, and the way that they practice. It’s both the challenge and the opportunity. It takes a lot of time and patience.
Q: We saw that you recently rode in VeloSano to support cancer research, what motivates you to participate in that event?
I’m a cancer doctor and I’m also a researcher, so VeloSano funded my ride, and that fund was very helpful for us. Ultimately, I ride for my patients. I am not a biker, but I became one and love biking now because of VeloSano. I’m hoping next year I can do 25-50 miles.