Big Data Use In Healthcare Needs Governance, Education
report, from the Institute for Health Technology Transformation (IHT2), a New York-based research and consulting firm.
Citing a 2011 McKinsey & Co. study, IHT2 said the U.S. healthcare industry could potentially save $300 billion a year with the help of advanced analytics, but healthcare organizations continue to struggle with managing and leveraging the vast stores of data they are building up.
- How Healthcare Payers are using Customer Communications to Improve Productivity and Effectiveness
- Learn how Kettering Health Network maximized clinician patient time by virtualizing clinician access to data
- The Value of Analytics in Healthcare
- Redefining Value in Healthcare: Innovation to expand access, improve quality and reduce costs of care
By 2011, U.S. healthcare organizations had generated 150 exabytes -- that's 150 billion gigabytes -- of data, IHT2 said. Kaiser Permanente alone might have as much as 44 petabytes of patient data just from its electronic health record (EHR) system, or 4,400 times the amount of information held at the Library of Congress.
[ How keen are healthcare pros to use big data? Read Healthcare Organizations Go Big For Analytics. ]
"It's true that big data promises to ease the transition to authentic data-driven health care, allowing healthcare professionals to improve the standard of care based on millions of cases, to define needs for subpopulations, to make more personalized decisions for individual patients, and to identify and intervene for population groups at risk for poor outcomes," the report states. "But while big data has transformed much of American industry, it's also true that massive information sharing and analysis has yet to generate significant benefits within health care."
"There are so many executives out there that just don't know where to start," IHT2 CEO Waco Hoover told InformationWeek Healthcare. Having a framework is "extremely important" to sorting it all out, he said
"A carefully structured framework for enterprise-wide data governance is arguably the first and most critical priority to ensure the success of any effort to leverage big data for health care delivery," said the report, which includes a summary of the Data Governance Institute's sample framework. The Data Governance Institute defines this framework as a "logical structure for classifying, organizing, and communicating complex activities involved in making decisions about and taking action on enterprise data."
But a management framework is not enough, says IHT2. The report recommends that healthcare organizations include providers in analytics planning because doctors often are resistant to efforts to change their practice styles. They are the ones "in the driver's seat" of care provision, so they are the most reluctant to change, Hoover said.
Healthcare organizations also need to "close the quality loop," said the report, by having data analytics specialists work closely with quality improvement teams to assure their strategies are aligned. "Achieving health care transformation requires dramatic and sustainable changes to the structure and processes of health care," the report said.
Hoover said he heard at a recent executive forum that some healthcare networks are hiring mathematicians with PhDs who "really understand how to extract data."
Longer-term, medical schools need to "bake analytics into training," according to IHT2. "It becomes a part of how medicine is practiced," Hoover explained.
"As we move forward in this new era of care, that has to be an integral part of delivering care," he continued. However, Hoover he predicted it would take 10-15 years before medical schools add this to their curriculum and start producing physicians trained in analytics.
In the meantime, more than a few provider organizations are counting on big data to help them manage population health and reduce variations in care.
"Electronic health records and automation tools already exist to identify and stratify individual patients who need special attention or care; identify care gaps; measure outcomes; and encourage patients to assume more responsibility for their health. However, they cannot store, manage, and distribute comprehensive, timely, and relevant information to the degree needed for public health management (PHM)," the report notes.
"In addition, many clinical analytics tools currently in use are quite primitive, reporting just a few basic facts and figures about a patient panel. The next generation of BI tools will have to be predictive and prescriptive to make PHM a reality."
A lot of the direction of an analytics strategy depends on how organizations define "big data," though. "What we're really talking about is leveraging your data," Hoover said, whether the database is big or small. "It's about getting value from data."
Sharp Community Medical Group, part of the Sharp HealthCare organization in San Diego, is starting to look at natural-language processing, according to medical informatics officer Dr. Neil Treister, one of the authors of the IHT2 report, because that seems to be helpful to physicians.
"[Sharp's analytics strategy] is driven by our doctors' priorities," Triester said. "Data is extremely important, but getting through their day is more important."
Triester said analytics are great for population health, but it is perhaps harder to derive benefits at the individual level. He told InformationWeek Healthcare said there has been an internal debate at the practice over how much granularity clinicians need. Many Sharp physicians are inputting a lot of data as free, unstructured text. "It doesn't make business sense to have discrete data" in some cases, according to Triester.
As large healthcare providers test the limits, many smaller groups question the value. Also in the new, all-digital Big Data Analytics issue of InformationWeek Healthcare: Ask these six questions about natural language processing before you buy. (Free with registration.)