Apixio announced today the release of its new cognitive computing platform, Iris, which it says will bring advanced data insights to healthcare by extracting and analyzing medical data previously trapped in electronic health records.
The U.S. annually produces 1.2 billion clinical care documents, but about 80 percent of the data is unstructured and difficult to access.
[See also: Natural language processing in high demand]
"Making sense of unstructured healthcare data is extremely challenging and requires sophisticated technology like cognitive computing to make the information useful," said Bob Rogers, chief data scientist at Intel, and one of Apixio's co-founders, in a press statement.
Iris is meant to give healthcare institutions access to patient data to create a more accurate care profile, thus improving the quality and efficiency of healthcare.
"We want to bring real world data to individuals to enable individualized care," Apixio CEO Darren Schulte, MD, tells Healthcare IT News, "by assembling patients with similar care profiles to find out what works and what doesn't."
[See also: Apixio serves up data]
Schulte says there are three fundamental components to the platform: data acquisition and assembly, learning and adaptive analysis capabilities and a user-friendly nature.
The platform uses Intel Xeon processor-based servers and an Apache Hadoop solution stack for scalability to process massive data volumes. The architecture is built on four layers implemented with open source and commercial products, in addition to its own patented innovations.
"By using Intel Xeon processor-based servers, Apixio's cognitive computing platform has the performance needed to do computationally intensive workloads that ultimately unlock untold value in healthcare data," said Rogers, in a statement.
The Iris platform uses proprietary data extraction tools and machine-learning algorithms: a self-learning system with real-time healthcare data, Apixio officials say, helping create a fuller picture of the patient.
Its analytics engine uses natural language processing and machine-learning technologies to create health profiles, evaluate risk and enable better quality care decisions and performance.
Apixio's platform groups similar patients and cases to achieve evidence-based patient data. Data changes over time, and by looking at similar patients the data can tell healthcare providers what works.
Apixio's isn't the first cognitive computing platform to be applied to healthcare. IBM's Watson is perhaps the best known, of course.
"We're not trying to compete with or denigrate Watson, but rather to take a different approach," Schulte says.
Apixio also announced the release today of its HCC Profiler, built upon Iris. The profiler mines clinical medical charts, Medicare-reported chronic condition data and medical billing data to compute patient risk scores. It creates targeted care delivery and helps Medicare set payer and provider payment for patients enrolled in Medicare Advantage.
San Ramon, Calif.-based Hill Physicians was one of the first providers to to try the HCC profiler pilot. Jennifer Pereur, the physician network's director of government programs, tells Healthcare IT News that the platform offers a way to "automate a manual process."
"Everyone in industry tries to get records, which is sort of onerous process," says Pereur. The profiler "removes that burden factor. It made it easier as an administrator because we can access all of the records. (It's) using technology to redesign a process we've been doing for years and years."
As healthcare undergoes massive transformation, Schulte says it's imperative to "improve care with computers – to assist and coach, not replace healthcare providers – and to use the information as evidence for more personalized care."
[See also: Big data holds keys to population health]