Pity the clinical data warehouse.
Many organizations set them up years ago with good intentions, only to discover that the limitations of healthcare technology interoperability forced them to stand up individual warehouses for clinical applications. Warehouses multiplied, as did data center size and complexity.
Houston-based Texas Children's Hospital faced this all-too-familiar problem. Following its Epic electronic health record (EHR) implementation, Myra Davis, senior vice president and CIO, says there was a "natural belief that all the data would be in the system and readily available to whomever needed it."
That wasn't the case. Only the data generated since the EHR implementation was available. What's more, Davis says, a "laborious" reporting process gave physicians, nurses and clinicians stale data by the time reports were ready.
"No matter how fast and furious we worked to get it done, the satisfaction level was low among our audience," says Davis. Even 6 a.m. meetings to discuss the issue failed to generate a solution. "We knew what we were doing wasn't working."
Today's Clinical Data Warehouses Must Be 'Adaptive, Agile'
A while back, Texas Children's senior vice president of quality came across Health Catalyst at a conference. The Salt Lake City-based company, founded in 2008 by veterans of Intermountain Healthcare's data warehousing initiative, had - as with many problems plaguing healthcare - looked to other industries to improve data modeling.
The resulting adaptive data approach, as described by Steven Barlow, Health Catalyst co-founder and senior vice president of client operations, grabs data from source systems and applies "minimal transformation" to it, to conform names, definitions and data types within the central platform. From there, organizations can build data warehouses, connected to a "stack of linkable identifiers" standardized across sources, for various constituent groups.
It's the "minimal transformation" part that reduces the overhead and complexity typically associated with data warehousing, Barlow says. Instead of reconciling every data element that's mapped to the model, Health Catalyst's approach sticks to the core elements. In turn, this lightens the data governance load that's typically necessary for an enterprise master patient index.
"Having been exposed to a lot of data warehousing solutions, the successful ones need to be adaptive and agile," Barlow says. "The data models that are common in healthcare are incredibly dynamic and complex, so a platform needs to be able to adjust to that."
Data Warehouse a 'Risky' Investment That Paid Off
The Health Catalyst platform was new when Texas Children's first saw it - so new, Davis says, that the vendor's demos involved clients in their production environments - but the results of a three-month trial, which focused specifically on reports based on EHR data, convinced Davis to approach other executives about adopting it.
The hospital called an off-cycle, summer quality board meeting. Davis and her peers "threw ourselves out there" in sharing their vision, asking for an investment and committing to a nine-month timetable. "It was risky," Davis says, "but it worked."
Two years later, the Texas Children's enterprise data warehouse pulls from 12 systems, including supply chain, payroll and finance, and enables operational reporting. "This is very beneficial," Davis says, "because organization leaders were making executive, operational decisions with stale data, and the way I interpreted data in financial reports wasn't the same way that other leaders would." In the process, the reporting team, no longer fielding lengthy report requests, made the transition to an architecture team.
Future plans for the data warehouse include studying workforce productivity and expenses, Davis says. Meanwhile, on the clinical side, Texas Children's plans to target segments of the patient population with better care based on specific diseases and conditions.
Healthcare Innovation More Than Just Technology
Use cases like that drive Health Catalyst, Barlow says. The disruptive innovation that healthcare needs involves technology, yes, but also the quality improvement that technology can bring. As Barlow sees it, this involves an analytics system to track care standards, a deployment system for that care and an evidence-based content engine to redefine care standards. That's more than a data model, he says.
"Organizations who really want to make headway in population management and quality improvement as we move to a value-based purchasing model, an analytics solution alone won't solve the problem," Barlow says. "There also needs to be a robust data methodology and deployment system."
Brian Eastwood is a senior editor for CIO.com. He primarily covers healthcare IT. You can reach him on Twitter @Brian_Eastwood or via email. Follow everything from CIO.com on Twitter @CIOonline, Facebook, Google + and LinkedIn.
Read more about health care in CIO's Health care Drilldown.
This story, "Making the Clinical Data Warehouse Relevant Again" was originally published by CIO.