Do the secrets to a healthier future lie in untapped healthcare data? Microsoft and UCLA Health want to find out and have built a platform that will make it easier to deploy artificial intelligence across heaps of clinical and research data.
“Part of what we are doing here is to enable precision health,” said Dr. Mohammed Mahbouba, chief data officer at UCLA Health Sciences.
Mahbouba thinks that machine learning is the natural next step for hospitals following the widespread implementation of electronic health records over the past few decades. The new platform will take in data from several sources so that it can be analyzed in a holistic way and, eventually, allow the hospital to tailor care down to the individual.
One ongoing effort at UCLA Health is an algorithm that identifies patients whose condition is likely to deteriorate. Mahbouba thinks that the platform can help caregivers preemptively take action for patients who are at risk for things like heart disease and diabetes. These actions might, in turn, reduce costs and prevent avoidable hospitalizations or emergency room visits.
The effort is a response to the growing appetite for predictive analytics based on structured data, such as lab results or personal information. But there’s also the potential to pull insights from unstructured data — more unwieldy sources of information that can include everything from genetic tests to radiological scans and ultrasound readings. “We’re being asked to put it all together and have data scientists develop algorithms,” Mahbouba said.
Making sense of medical records is a popular goal. Google Cloud is working on a service that pulls together disparate sources of medical data. Amazon is also in the race with a service called Amazon Comprehend Medical that mines patient data records, including unstructured data. And electronic health record provider Epic builds its own algorithms to help inform clinical decisions.
“It’s only really been in the last couple of years that Microsoft Azure, as well as our competitors’ clouds, have reached the level where they can safely handle and manage at scale the massive aggregation of health data that UCLA and Dr. Mahbouba are envisioning here,” said Peter Lee, corporate vice president at Microsoft Healthcare.
Microsoft has steadily made inroads into hospitals in other ways as well. Earlier this year, it rolled out an Azure API for health record sharing as well as Microsoft 365 for hospital teams.
Healthcare providers have been slower than other industries to embrace cloud computing in part due to concerns over data security. Microsoft, Amazon, Google and others have responded by making their cloud services compliant with federal regulations such as HIPAA.
But Mahbouba said that the healthcare industry has started to come around to cloud computing. “I would argue in many ways, especially for smaller organizations, it’s much safer to have the backup of an organization like Microsoft to make sure the security standards are upheld,” he said.