Google wants to continue edging its way into the healthcare market by easing one of the major pain points for healthcare...
IT pros: bringing together disparate and isolated data sets.
At the recent Radiological Society of North America (RSNA) annual meeting, the company announced new Google Cloud for healthcare features including two new codelabs or tutorials on how to build AI applications on top of medical imaging data, as well as making public data sets easily accessible for healthcare researchers and support to redact sensitive data for its Cloud Healthcare API.
For the last two years, Greg Moore, an engineer and practicing neuroradiologist, has been leading the Google Cloud healthcare and life sciences charge. Moore, vice president of healthcare for Google Cloud, received his Ph.D. in radiological sciences from MIT and served as the chief emerging technology and informatics officer for Geisinger Health System before joining Google.
At RSNA 2018, Moore talked strategy for the company's Google Cloud for healthcare offerings, which he said could break down data silos for the industry. He also weighed in on the role AI is playing in healthcare and his hope to bring about "positive digital transformation."
Editor's note: This interview has been edited for brevity and clarity.
What's the big Google Cloud message at RSNA 2018?
Greg Moore: One of our core offerings is Google Cloud Healthcare API. It can ingest healthcare data, including imaging data, in its native form, so it's fluent in [Digital Imaging and Communications in Medicine], but it's also fluent in other healthcare data types like [Health Level-7, version 2, Fast Healthcare Interoperability Resources] and others. That's really the core, our beating heart offering.
The ability to take those silos of information that are in healthcare, that's our core competency. Imaging at RSNA is one silo of data -- typically in [picture archiving and communication systems] that you see throughout the floor. Medical imaging, while important and a critical piece of healthcare, is just a piece. It's an important piece, but the clinical records are equally important, to get that electronic healthcare record phenotype together with the imaging phenotype … all of those traditionally are in different silos. What we do at Google -- our core competency is to manage that data. We will break those silos down and bring it to a common platform so we can bring value to patients.
AI is a major theme here at RSNA 2018. How is Google Cloud using AI? And what is driving the trend of AI in healthcare?
Moore: We've made massive investments in machine learning and AI. We are an AI-first company. … We have offerings like AutoML -- auto machine learning tools -- where, for instance, a healthcare system or an education institution wouldn't need to have a whole team of data scientists and hundreds of engineers to develop [machine learning] products.
We are at an interesting point of time -- the intersection of data availability that our partners and customers are creating with higher resolution machines and imaging systems, and we have now the availability of very powerful compute. What we're bringing to them are these [machine learning] algorithms and the compute space so they can gain meaningful insights from that data for their customers and their partners. We're providing them that platform and infrastructure.
What will be big for Google Cloud next year?
Moore: We are seeing this explosion of data, and that creates tremendous possibilities for companies in this space that partner with us to gain great insights from that data for patients.
It's a conversion of data plus powerful compute, and, in 2019, I see those two large drivers coming together. I see this becoming incorporated into workflows to create accessibility to healthcare, to create safety, to create security going forward. I see that becoming operationalized -- and much more quickly than anyone realized. That sort of happened overnight with your mail and photos and your texting.
I think what you'll see is these great assists, likely not in diagnostics since it's got to go through a routine regulatory pathway, but in the routine workflows of doctors every day, whether that's dictating notes or documenting a patient interaction or scheduling patients or predicting no shows and making sure people get to their appointments, even the back office and billing. I think that's where you'll see first deep integrations in the workflow. Some of these things that are part of running the pieces of healthcare; I think that's what you're going to see widely adopted in the coming year.
What challenges can Google address in healthcare? And what about opportunities for the company?
Moore: My thought is that we can disrupt without being disruptive. We don't want to be disruptive to the patient experience. We don't want to be disruptive to anyone's business. But, certainly, there is vast opportunity for improvement going forward. We know something like 80%, 85% of healthcare transactions are still by fax. Here we are in 2018, almost 2019, and we're having to teach medical students how to use a fax machine. We can do better. Healthcare can do better than that.
I see technology and cloud in general providing an impetus for positive digital transformation in the healthcare space -- to let healthcare focus on its core competencies by providing care for the patients and not have to worry about the IT. Healthcare systems have a lot of IT from everything from call centers and schedulers to back office to their clinical healthcare records, and much of that is stuck in 1970s era. My vision, my hope for positive transformation, is that we give back to healthcare providers, healthcare systems -- the ability to focus on their core competency and let us help them with the technology piece.