Google Cloud says HIMSS20 attendees should eye AI, interoperability and security
There will be many technologies, strategies and trends discussed at HIMSS20 in March. But which are most important? Which deserve the time and attention of healthcare CIOs and other leaders?
Healthcare IT News spoke with Dr. Joe Corkery, director of product, healthcare and life sciences, at Google Cloud (Booth 3729), to get his expert view of the health IT terrain. He identified what he says are three very important trends and technologies for HIMSS20: AI in healthcare, data interoperability, and data security and privacy.
AI in healthcare
The evolving role of AI in healthcare is sure to be an important topic on people’s minds at HIMSS20, Corkery stated.
“The past few years have seen some incredible demonstrations of the power of and potential for AI in healthcare – for example, breast cancer detection and diabetic retinopathy detection,” he said. “As well as a lot of discussion around how AI will be used, integrated and regulated as part of clinical practice.”
As an “AI-first” organization, Google has significantly transformed the way it works through AI and brings this expertise to healthcare, he said. As part of that effort, Google is focusing heavily on addressing some of the key requirements necessary for success with AI, he added.
"Data silos continue to be a significant challenge in the healthcare industry and, as a result, data interoperability continues to be an important theme as we head into HIMSS 2020."
Dr. Joe Corkery, Google Cloud
Corkery identifies three requirements for AI success. The first is supporting a data ecosystem.
“To build AI models, healthcare providers need large, well-labeled and relevant datasets on which to train,” he said. “To help address this, we’re working with a number of generators and providers of public data – for example, the National Institutes of Health – to make those datasets available on Google Cloud for healthcare institutions to use.”
In addition, Google is investing heavily in addressing the challenges of document processing and data interoperability to make it easier for Google users to create and join multiple distinct datasets together in the same format to provide a consistent base on which to train, he explained.
“Finally, we’re providing tools to make it easier to de-identify healthcare data so that our users can remove personally identifying information before working with those datasets,” he added.
Tools and frameworks
The second key requirement for success with AI that Corkery identified is tools and frameworks.
“Google Cloud provides solutions for AI developers across the spectrum of skills and experience,” he said. “Our AI Building Blocks make it easy for developers to add sight, language, conversation and structured data capabilities into applications. As part of that, tools like AutoML are making it easier for less experienced developers to build and deploy their own new custom models for those capabilities with less training data.”
The Google Cloud AI Platform provides a code-based data science development environment that empowers machine learning developers, data scientists and engineers to go from ideation to deployment quickly and cost-effectively, he contended. And Google’s Healthcare API is designed to enable Google Cloud to “speak the language of healthcare” by providing the necessary connections between standardized health data and Google’s Cloud AI capabilities, he added.
And the third requirement for AI success Corkery notes is computation scale.
“Training new AI models is computationally intensive,” he explained. “Google Cloud provides the necessary scale and elasticity to ensure that AI developers have the resources that they need to build their models. In addition, hardware like Cloud TPUs are making it even faster and more cost-effective to build AI models using Google Cloud.”
In addition to AI’s continuing evolution in healthcare, Corkery says another major trend at HIMSS20 will be data Interoperability.
“Data silos continue to be a significant challenge in the healthcare industry and, as a result, data interoperability continues to be an important theme as we head into HIMSS 2020,” he said. “In 2019, we saw the announcement of the CMS Interoperability and Patient Access Proposed Rule, and its focus on FHIR as the API standard for patient data access.”
Interest in and adoption of FHIR as a standard is rapidly increasing, with this proposed rule as just one prominent example. Google likes to see this momentum around FHIR and the opportunities that it will enable, especially regarding the development and deployment of new clinical applications to improve patient care.
“Google is a strong supporter of open standards broadly and, in particular, FHIR for healthcare,” Corkery said. “The Google Cloud Healthcare API provides an engine for data interoperability as a cloud-based managed service that supports multiple industry standards including FHIR, HL7v2 and DICOM. It bridges the gap between existing systems and the advanced data analytics and machine learning capabilities offered by Google Cloud.”
Data security and privacy
The third trend or issue Corkery and Google say will be key around HIMSS20 will be data security and privacy. Not a new issue, but one of increasing importance.
“Healthcare institutions are increasingly worried about the security of their systems and, in particular, the patient data stored in those systems,” Corkery said. “This is partly due to an increase in breaches and cyberattacks, and also to a growing realization by providers that their legacy IT systems have become too sprawling, complex and vulnerable.”
Simply put, provider organizations do not want to be in the business of IT security – they want to be in the business of providing better health outcomes to their patients. So, as healthcare institutions turn to newer, cloud-based solutions, many of them come to Google to leverage both the breadth and depth of Google’s security expertise in cloud infrastructure, productivity tools like GSuite, and cloud-native hardware, for example, Chrome devices, he added.
“Individuals also increasingly are worried about their data and how it is used by their providers and their business associates, such as EHR vendors, lab companies, etc.,” Corkery said. “Of course, HIPAA provides stringent rules for both data security and privacy, but the legislation often is not well understood by most patients, which frequently leads to confusion.”
Providing greater transparency as to how data is used and protected, for the provision of care, improving the tools used to provide care, and for research is top of mind for providers going into HIMSS 2020, he contended.
“Google Cloud is explicit in our role as a data processor, our compliance with regulatory requirements like HIPAA, and our certification with frameworks such as HITRUST,” he stated. “Google Cloud also is making investments in providing healthcare institutions with the tools they need to facilitate their own HIPAA compliance when operating in a cloud environment, including tools to assist in the de-identification of healthcare data.”
But, he concluded, Google also realizes that more education is likely needed to ensure providers leverage these tools – and patients understand how their data is being used.