Databases Reference
In-Depth Information
materialized into structures that can be rapidly accessed and analyzed via off-the-shelf tools or cus-
tom-built applications. Those applications then generate data back into HBase for other processing
needs. Analytics powered by Hadoop facilitate clinical-quality measure generation, measure calcula-
tions for registries, and proactive care management and other critical tasks.
For example, Explorys uses Hadoop-powered analytics for purposes similar to the following sce-
nario. To better serve its community, a county hospital might want to explore why people utilize the
emergency room (ER) rather than going to a primary physician for care. This practice is not beneficial
for either the patient or hospital, because it is expensive for the hospital and the patient doesn't get the
continuity of care that a primary physician would provide.
Additionally, it is important to analyze a bigger problem than just the ER record system. With
Hadoop, Explorys can analyze what factors are demographically different about a given population,
where patients live, and whether or not care is available in their neighborhoods. Explorys can run
these analytics daily to take immediate action, such as sending a letter to a patient the day after she
or he has visited the ER that includes information about local healthcare providers and instructions to
help prevent readmittance.
Business challenges
McHale explained,
With a clinical enterprise performance management system built on a relational data warehouse,
if healthcare practitioners want to understand something specific about a population or segment
of data, they have to go to their IT departments and then wait days or weeks to get that informa-
tion back. We wanted to provide a platform that would give them an answer as fast as if they were
searching on Google.
Rather than managing clinical, financial, and operational data in three data silos, Explorys sought
to bring all of that data together. Lougheed explained,
It's about merging the three elements and telling a story about how an organization is doing,
because ultimately what we want to do is improve healthcare and do it at a lower cost. With over
17% of the nation's GDP being spent on healthcare services, we've got to find a better way to
deliver healthcare at a cheaper cost point.
McHale added, “We had lots of experience working with traditional database platforms in our
careers in banking and telecom, and we just didn't believe they were going to scale economically or
from a performance standpoint.”
The variety of data would also present a challenge. “Electronic Health Record (EHR) platforms
are just now becoming prevalent across the medical landscape,” noted Lougheed. “There are more
and more devices that generate massive amounts of data. Patients are providing data and feedback
on how they're doing. They have devices in the home that provide data. There's a ton of data variety
coming in and it's more than the healthcare space can really handle.”
Explorys needed to find a cost-efficient technology that would help the company address these
Big Data challenges. Hadoop met both of these criteria, and Cloudera stood out to Explorys as the
most credible company delivering an enterprise-ready Hadoop solution. McHale, Lougheed, and Meil
attended the first Hadoop world conference hosted by Cloudera in 2009, and were sold both on the
value of Hadoop and on Cloudera's “ability to deliver.” They also appreciated Cloudera's contingent
