Cars and trucks aren't the only thing whizzing along the I-44 corridor in the center of the country. Thanks to Mercy, a large four-state, 35-hospital, and approximately 700 clinic health system based in Saint Louis, terabytes of data are moving along the highway.
Mercy's main Epic (Verona, Wisc.) clinical repository contains 25 terabytes of data, housing information on patients across the four-state region. And as Paul Boal, director of data engineering and analytics at Mercy says, "That's just the starting point." The actual scope of the data goes well beyond that massive number, including another 25 terabytes of data on patients’ privacy information that’s stored in another database.
The challenge for the health system is taking all of that data, information on roughly eight million patients, and drilling it down to actionable level for individuals. Aggregating data in terms of averages is easy within a traditional database, says Boal. Getting to the point where you can track and analyze outliers is much harder and costlier.
“When you start to drill down and ask, ‘Who are the patients that don’t match these criteria?’ you find the analysis gets more complicated and everything starts bogging down,” Boal says.
That's why Mercy turned to something called Hadoop, an open-sourced framework for storing and processing large datasets. Using a platform from Hortonworks (Santa Clara, Calif.), the health system replicated and optimized its data from Epic into Hadoop giving providers real-time functionality to analyze and act on it. The platform is interfaced within Epic, meaning providers can pull out information directly within the EHR.
"One of the neat things that Epic allows you in that integration model is you can...create hyperlinks in that report that make the data actionable. I can click that hyperlink and it will pull up a specific page on that patient in Epic hyperspace," says Boal.
The main purpose for using the open-sourced framework, for now, is to improve medical documentation. Boal says that Mercy is aiming to get documentation up to speed by the time the patient is discharged. This not only helps the providers with actionable data but coders, who can get the right information in the chart and create a more efficient reimbursement process. “[The coders] have better information up front about the patients that are in the hospital right now and which patients they should be focusing on when they do their chart reviews to review with physicians during huddle times,” he says.
They also have a few other Hadoop-related projects ongoing. The labs are consolidating their technology footprint, says Boal, and through Hadoop, Mercy is developing a way to “Google” their way through lab notes. The health system is also planning to move the nearly 25 terabyte privacy database into Hadoop as well.
In the near future, Mercy is planning on using the framework to integrate vitals information from the health system’s large electronic intensive care unit (eICU) into the EHR for preventive care. The platform can allow for more detailed readings of vitals, and in real time, that allow for better analysis.
“What we’re building out is a real-time clinical applications platform, so we’re looking for other opportunities to turn that into decision support,” Boal says. “We’re looking for folks that are interested in device data integration.”
Tech and Business
Moving to Hadoop was a whole new learning curve for Boal and his team of technical experts. He said one of the challenging aspects of the implementation was bringing everyone up to speed on a new technology, since many were either traditional programmers or database developers that wrote SQL for a living.
Despite this, Boal considers moving to this open-sourced framework a “leap forward” in terms of healthcare analytics. Those who are interested in moving in this direction simply have to understand that a two-sided approach to implementation is necessary, he notes.
“There is a technology play at work here, in that we’ve got a great new technology that the world has made available and we have to figure out how to take advantage of that. And we need to implement it. We need to be doing more with this technology. It’s powerful,” says Boal. “At the same time, we need to make sure we understand what are the particular business needs [for this technology]. You can’t just sit and wait for the business part of the organization to tell you to implement big data. The business problems aren’t big data problems or little data problems, they’re business problems.”