Precision Health AI Launches Platform to Deliver Data-Driven Insights in Oncology
Precision Health AI has launched Eureka Health, a platform that leverages electronic medical record (EMR) data to provide artificial intelligence (AI) through more than 60 pretrained and learning-enhanced modules designed specifically for oncology.
“If we’re trying to build meaningful artificial intelligence, the thing you’ve got to have is deep, longitudinal data about patients,” said Brigham Hyde, PhD, chief executive officer, Precision Health AI, in an interview with The American Journal of Managed Care®. Through a partnership with the American Society of Clinical Oncology's CancerLinQ, Precision Health AI collects data from provider sites, including major hospitals and community centers. According to Hyde, this includes data on radiology, genomics, patient and physician notes, and genetic testing. The data is then aggregated, de-identified, and integrated into the AI platform.
The AI platform then looks for specific questions important in oncology and infers information about precision medicine, such as what treatment or test a patient should get and what the outcome would be. When a researcher or clinician logs in to the system, they have access to more than 60 modules that are pretrained on the collected data.
“Somebody from our team already went in and said hey, ‘Let’s try this algorithm and that algorithm with this data and that data,’ and picked out the best 1, the 1 that performs the best and is most accurate,” said Hyde. “Then they productize that, and that module can run on top of our data or on top of data locally. It can feed intelligence to researchers or feed intelligence into a platform like an EMR or clinical decision support tool.”
One of the algorithms the platform has looks at adverse events from drugs. According to Hyde, a big problem in oncology is chemotherapy-induced neutropenia, which affects 1 in 5 patients. With many doctors not knowing which patient will develop the condition, they start blind and uninformed of the risk factors. The Eureka Health platform is able to predict which patient is going to get neutropenia as a result of the drug.
In addition to helping guide clinical decisions, the platform can also assist in workflow. While physicians don’t need help determining the stage of disease a patient is in, they do need help with the documentation of it in the record, said Hyde. The platform takes the busy work away from the doctors by being able to accurately predict the stage.
“The way we see the world is that we’re building a system of intelligence. We’re taking expert knowledge and understanding of the setting of care and the issues, and we’re encoding it into this AI platform with these individual modules,” said Hyde. “The system of intelligence works really well when partnered with the system of engagement. We do not want to make another stream for doctors to look at, and because of that, we think this should live within the EMR or another clinical decision support tool.”
Hyde compared the platform to the production facility for Netflix, citing Netflix’s need for someone to make new shows to keep people engaged in its system. EMRs require the same kind of team, and Precision Health AI provides the content.
In the long term, Hyde hopes that the platform can provide a strong real-world evidence base to help drive discussion between different stakeholders, including insurance companies, providers, and drug companies. Right now, said Hyde, insurance companies don’t feel like they’re given the transparency and evidence on a lot of new drugs, such as immunotherapies, where many patients who don’t actually need them or are not good candidates for them are given the type of drug.
“My dream is that we end up talking around a common, real-world source of evidence to all make decisions, and then the negotiations about value are more straightforward,” said Hyde. “That may be a bit of a utopia, but I think we can get there in oncology.”