Print Friendly, PDF & Email

We asked top clinicians what they really think about the future of AI in healthcare. Here’s what they said

 Despite the debate, most say AI’s potential to transform healthcare is great…and it’s closer to reality than we expected

 

Artificial Intelligence is among healthcare’s most widely discussed and hotly contested technologies. It has been invested in by industry players and startups alike, and hospitals are evolving from asking “why should I adopt it?” to “when?”

With that, it’s no surprise that the topic is taking center stage at the largest gathering of radiology professionals, the Radiological Society of North America (RSNA), in Chicago this week. The Pulse asked six top clinicians taking part in the conference how they view AI and it’s potential to transform healthcare. Here’s what they said.

Why is AI so widely talked about? How is it different from other digital health solutions that have come to market?

“We get excited about AI because we’re inundated with data today at unprecedented scales. At UPMC, we have 27 petabytes of data that’s doubling every 18 months. AI offers a glimmer of hope – these algorithms can tame that data beast and decipher insights in ways that have not been possible before.”

– Dr. Rasu Shrestha, Chief Innovation Officer, UPMC

 

“The power of Artificial Intelligence technology is that there’s not a boundary in its capability. The more data we can show it, the better it performs. And that flexibility is really what’s so exciting.”

– Dr. Mark Michalski, ‎Executive Director, MGH & BWH Center for Clinical Data Science

“People see smart computers all around them – Apple’s Siri, Amazon’s Alexa, Tesla’s self-driving car – and they think healthcare should be the same. Obviously, healthcare is far more complex, requires much higher accuracy, and has less margin for error. Thanks to advances in computing power and data science, we have entered a new era of medicine – we now have a tremendous opportunity to improve the quality and efficiency of care, and prevention and prediction for an individual are finally going to be possible.”

– Dr. Michael Blum, Associate Vice Chancellor for Informatics, UCSF

“We’re being asked as clinicians to integrate a tremendous and ever growing amount of data into our clinical practices. Until recent times, we really haven’t had the tools to harness that vast quantity of information in a meaningful way to help our patients. AI helps solve that problem.”

– Dr. Rachael Callcut, Director of Data Science, Center for Digital Health Innovation, Associate Professor of Surgery, UCSF

What’s the one year, two year, five year trajectory?

“In one year you’re going to see the first proof of concept where we’re actually doing care fundamentally better and saving lives. Those are already queued up, and we’re working on them now. In two years, we’ll define what this means to the system – the payers, the administrators. In five years, you will see paradigm shifts in pathology, radiology – where different kinds of data are making our predictions and planning.”

– Dr. Scott Hammond, UCSF

“Today the technology is really good at a narrow set of applications – things like characterizing an image, segmenting a tumor or pointing out that there is a pulmonary nodule in a CT scan. What it can do in the future is start to give us more contextual awareness, start to find features in those images that maybe we don’t even see as humans.”

– Dr. Mark Michalski, ‎Executive Director, MGH & BWH Center for Clinical Data Science

“People ask me when will AI happen, and I ask when did the Internet happen? It will be a continuous process with incremental advancements. And those advancements will require a substantial commitment of capital, expertise, personnel and collaboration – to not just create the solutions but to test them, bring them to market, and scale them globally.”

– Dr. Keith Dreyer, Vice Chairman of Radiology, Massachusetts General Hospital

What’s the key to successful AI development and implementation?

“We want to see these technologies work for radiologists, and what that implies is that they have to work within the workflows that exist today. AI needs to be not only an interesting tool in an academic pursuit but also something that really works and helps us practice our profession better.”

– Dr. Mark Michalski, ‎Executive Director, MGH & BWH Center for Clinical Data Science

“Partnership is critical. As clinicians and researchers, we can interact with people who are at the forefront of technology and can bring us tools that we only dreamed of having. Our partners have the opportunity to gain important clinical context which is vital as we try to bring these AI digital solutions into the healthcare marketplace.”

– Dr. Rachael Callcut, Director of Data Science, Center for Digital Health Innovation, Associate Professor of Surgery, UCSF

“There are thousands of concepts that a radiologist needs to know, detect and determine in their practice. The best approach is to articulate these concepts as use cases determining which are of highest clinical value and most amenable to AI, given the machine learning algorithms available today. From there, you should deploy resulting the AI models on a platform to test it in various environments and at scale to ensure it evolves.”

– Dr. Keith Dreyer, Vice Chairman of Radiology, Massachusetts General Hospital

How will AI change the role of clinicians?

“We will need to continue to evolve with the set of capabilities that become real and available to us. The term AI today stands for Artificial Intelligence, but as it relates to healthcare it should really be Augmented Intelligence. AI, done right, will make us better clinicians and accentuate the most human aspects of why we got into care in the first place – to focus on patients, not to be inundated with mundane tasks around documentation and data crunching.”

– Dr. Rasu Shrestha, Chief Innovation Officer, UPMC

“AI won’t replace radiologists, but radiologists who use AI may replace the ones who don’t.”

– Dr. Mark Michalski, ‎Executive Director, MGH & BWH Center for Clinical Data Science

“It’s easy to imagine that in the near future we will develop algorithms that are numerically as good or better than doctors at making a diagnosis, or recommending a treatment.  However, there will always be the need for experienced clinicians in the complex, emotional undertaking of providing healthcare.”

– Dr. Michael Blum, Associate Vice Chancellor for Informatics, UCSF

*UPMC, MGH, BWH and UCSF are all collaboration partners of GE Healthcare and as a result, have a financial interest in the development and commercialization of certain GEHC next generation imaging products.