On Tuesday, August 11, Cleveland Clinic’s Transformation Tuesdays dove into the world of artificial intelligence (AI) and machine learning (ML) – detailing its development and uncovering some of its best use cases with the help of experts in the field. Kicking off August’s new topic, panelists Aziz Nazha, MD, Director, Center for Clinical Artificial Intelligence & Associate Medical Director for Artificial Intelligence, Enterprise Analytics, Cleveland Clinic, and Oliver Schabenberger, PhD, Executive Vice President, Chief Operating Officer & Chief Technology Officer, SAS were moderated by Anil Jain, MD, VP & Chief Health Informatics Officer, IBM Watson Health. In their riveting discussion of the AI and ML platforms of today, these thought leaders illuminated the tech’s long road to reality and its ability to help countless innovators understand complex situations – seeing problems and patients on a more comprehensive level.
Characteristic of any expert, our panelists began their discussion of AI with some level setting. Defining the concept in their own words, a base was laid for deeper discussion. Moderator Dr. Jain prefaced his request for definitions with a statement that few working with the technology would deny: the concept of AI has changed drastically over time. Dr. Nazha of Cleveland Clinic described the field as an umbrella, housing many subfields underneath it. Of these subfields, he highlighted machine learning, which he described as “teaching a computer with data.” Dr. Nazha dug even deeper to the concept of deep learning – a subset of machine learning that utilizes a neural network to find patterns in said data. When asked for his definition, Dr. Schabenberger of SAS began, “To me, AI is about building computerized systems that perform tasks or make decisions a human could make. It’s really a multi-disciplinary and multi-technology effort.” He continued, “It includes robotics, robotic process automation, expert systems, hand-crafted knowledge systems, and now, machine learning-based and data-driven methods, as well.” If nothing else, AI’s layered composition alludes to its incredibly diverse capabilities.
To tie the conversation to healthcare, Dr. Jain asked how clinicians at Cleveland Clinic are thinking about AI. As the Director of the Center for Clinical Artificial Intelligence at Cleveland Clinic, Dr. Nazha highlighted a point of pride in clinicians’ ability to see its true value. “We, as a society, tend to be driven by the media hype that AI is a miracle tool to fix our problems. But [clinicians], rather, approach AI as another helpful tool in the toolbox.” Dr. Nazha also made a point to say that in the past, we haven’t had all the pieces parts required to use AI to its full advantage. In our increasingly connected society, with increasing access to connected health devices, we now have appropriate amounts of data to feed these platforms – with more data comes more robust and informed machine outputs and human decisions. Dr. Shabenberger agreed stating, “It’s really a perfect storm. These new artificial intelligence techniques rely very heavily on data, and we have more data thanks to digital transformation. The availability of computing, the availability of advanced models, and rich data sources have really elevated the capabilities that we have.”
On the point of society perception, Dr. Schabenberger reflected, “It’s interesting to see how the conversation around AI has changed. In 2017 it was, ‘the machines are coming for us.’ In 2018: ‘okay that’s not going to happen, but the machines are coming for our jobs.’” Dr. Schabenberger continued with his recommended perception today in 2020, “It’s really about augmentation – how can this technology help us be better? But augmentation goes both ways. We, as professionals, have the ability to help artificial intelligence get better. This sort of coexistence improves the capabilities of the systems and, ultimately, our decision making.”
It came as no surprise that later in their discussion, the group referenced AI’s use in other industries. Healthcare, cautiously slow to integrate new technology into operations and practices, saw the emergence of AI much later than other workflows. While compiling data showed AI meeting efficiency milestones and pushing boundaries in varied trades, clinicians remained rightfully steadfast in their demand for more information. However, in our careful analysis of other industries, healthcare has learned some valuable lessons – namely the power of good data. The quality of insight is only as good as the quality of data used to train AI/ML models. The group shared support of the statement, “garbage in, garbage out.”
A natural conversation point, given the current environment, our panelists explored AI in the context of COVID-19. The field of AI, like other fields of research, has pivoted to find usefulness in the quest to cure and quell spread. Said Dr. Nazha, “We’ve seen a lot of attempts to use the technology to help in the pandemic – mainly to repurpose drugs and identify novel compounds to target this new coronavirus. Contact tracing represents another use. There have been a few companies using machine learning and artificial intelligence alongside health and mobility data to try and alert people who might have come in contact with COVID-19-positive patients.” Dr. Nazha followed with some research at Cleveland Clinic, including an ML model to predict length of hospitalization for insight into bed and resource utilization during COVID-19’s surge.
To view the full discussion and learn more about AI’s past, present, and future applications, click here. We want to thank IBM Watson Health and SAS for their contributions to this wonderful discussion – groups whose technical expertise keeps the healthcare industry and provider systems like us moving forward. In the month of September, we’ll be using our Transformation Tuesdays to talk augmented and virtual reality (AR/VR) in healthcare. Be sure to join us for our first webinar of this series from 1:00 – 2:00 pm (ET) on Tuesday, September 15.