Trinity, in addition to its hospitals, has 432 outpatient sites and 33 long-term care facilities in 10 states. As such, Trinity has long been familiar with big data, according to Kyle Johnson, its vice president and chief analytics and integration officer.
“We've had a relational database warehouse for the last 13 years here at Trinity,” Johnson said. “We're steeped in data, clinical, revenue and supply chain data. You have to think, why would Trinity buy Explorys? It has to do with velocity and volume.”
Velocity, according to Johnson, means more than just data processing speed—although that's a part of it. It's also the “speed at which you can acquire new data sources,” she said.
“It takes four to six months to acquire a new data source, and we've been doing it for years,” Johnson said. “With Explorys, we may be talking four to six weeks, and we think that we'll get better as we get familiar with their tools.”
Meanwhile, data-processing horsepower at Trinity is under strain. “Our revenue system takes about six or seven hours to process and it can't start until midnight, but we have to have it ready by 6 a.m.,” Johnson said. Trinity plans to offload some of that data processing work to its new contractor, he said.
Explorys uses a Hadoop-based data-processing framework. Derived from work initiated by Google, with later contributions by Yahoo, and now available under the Apache open-source license, Hadoop software is capable of handling massive data flows by breaking them down into smaller chunks and assigning them across multiple servers for faster processing. It's also designed to easily handle both discrete data—a lab test result, for example—and unstructured data, such as a clinician's typed or dictated note.
“Today, our relational database environment can't handle unstructured data,” Johnson said. “It might be clinical notes, genomic data, medical device data, but you want to pair it with structured data (and) you can't pair that in a structured database. We're talking about petabytes and exabytes. These Hadoop and big-data environments can handle the volume. We think (this is) where healthcare transformation is going to push us; these are areas we'll have to get into, maybe not this year, but in the future.”
Johnson said Explorys operates a private cloud, with its own data centers in Cleveland and Texas, which helps assuage concerns about privacy and security that a pure cloud might have generated. Trinity also plans to retain its data warehouse, so any data processed by Explorys will be copied back to Trinity and stored there, she said.
Both big data and cloud computing are getting a lot of buzz in health IT circles, but, beyond the hype, the potential benefits are real, said Dr. Kevin Fickenscher, president and CEO of the American Medical Informatics Association, a professional association for research and clinical informaticists.
In healthcare, cloud computing remains “at the formative stage,” Fickenscher said. “Up until about a year ago, there were individuals that were concerned about whether we could create a secure environment in the cloud. I think we've proven that that is possible.”
What will drive adoption of cloud computing in healthcare will be its economies of scale, which will be needed to cost-effectively handle the explosive growth of data volumes in genomics and microgenomics, he said.
“Every single healthcare system is looking at their costs,” Fickenscher said. “Trying to create enough capacity at a single institution or data center is extremely costly, so moving to a cloud is going to create a number of opportunities.”
Follow Joseph Conn on Twitter: @MHJConn