Innovation Pulse: An early glimpse into a learning health system
When NYU Langone Medical Center established a Department of Enterprise Data Quality, the academic medical center homed in on exactly how information is collected, categorized and distributed -- as well as how to validate quality across the entire clinical enterprise.
"It has a lot to do with how we leverage data inside our system," said Paul Testa, MD, Langone's chief medical information officer. "We pride ourselves on clean data, which enables a learning system."
NYU Langone is among the new breed of academic medical institutions -- Indiana University, University of Michigan, University of California at San Diego, among others -- that are increasingly focusing on data quality for education, research and care and, in so doing, are becoming learning health networks that offer a glimpse into the future of a learning health system.
Grand slam?
Indiana University has, since 2011, been participating in the collaborative Hospital Medicine Reengineering Network, sportively dubbed HOMERuN, to transform itself into a learning organization.
HOMERuN involves research into ontologies and classifying data so it's accessible between institutions in meaningful ways within an applied environment based on agreement to data sharing standards that include ways to use that information for measuring and benchmarking efficiency, quality and outcomes.
A foundational challenge to any of the emerging networks is granting data access while protecting individual patients across distributed research networks, said Lucila Ohno-Machado, MD, chair of the Biomedical Informatics department at UC San Diego.
[Also: Read more Innovation Pulse columns]
"If we're going to build an isomorphic virtuous circle where we learn how interventions lead to population health outcomes, we have to really conduct management and analytics to become data-driven," said Ted Hanss, Chief Information Officer at the University of Michigan.
UC San Diego, for instance, uses analytics and algorithms to reach data and get the results it wants for research without actually moving any patient information, Ohno-Machado explained.
"A corollary challenge is getting systems and services defined for healthcare then repurposed for research, such as EHRs and the data therein," said William Barnett, the first and current Chief Research Information Officer for Indiana's Clinical and Translational Sciences Institute and the Regenstrief Institute.
Yet doing so can payoff rather nicely. At NYU Langone, for instance, a big success story details how the hospital system is leveraging its EHR to implement what Testa described as "electronic pathways" to better manage the process of caring for specific disease states in a standardized way and derive value from that data. Langone is doing this for patients who had colon surgery, those with congestive heart failure, and pneumonia.
"The paradigm shift is value-based medicine, quality divided by cost times patient experience -- you have to address all three," Testa said. "We move patients out of ICU faster, out of hospital faster, and we can detect why people are moving off of a pathway."
What surprised Testa the most is just how quickly the clinicians and care teams got in the groove of electronic pathways.
"It allows us to see changes in the system very quickly," Testa said.
Indeed, by leveraging the data warehouse, EHR, and a single booking system that connects more than 100 NYU Langone sites with its labs, the health system was able to pinpoint between 30 and 40 instances every month in which having lab technicians simply call a doctor back to double check whether they really wanted to run a specific test, or not, eliminated waste and redundancies.
"I look at our core missions: academic, care delivery and research -- getting those three to leverage the same data, that's a whole different game," Testa explained. "It's then a fait accompli to bring in a research machine to do the same analysis.
Big guns sparking change
The U.S. is amid "a shift in the informatics space of groups actively collaborating together to provide information and analytics toolsets," University of Michigan's Hanss said, on a national level that was not happening even as recently as five years ago.
Federal government heavyweights including the National Institutes of Health and National Science Foundation are funding and fueling data networks, namely the Patient-Centered Clinical Research Network otherwise known as PCORnet, and the National Center for Advancing Translational Sciences' Clinical and Translational Science Awards that encompasses 62 participating institutions.
Those networks enable UC San Diego, for its part, to access the analysis of information pertaining to 13 million patients in California via a data warehouse in a distributed fashion that protects the data's integrity and patients' privacy.
Reality when?
With its 10-year interoperability roadmap, the Office of the National Coordinator for Health IT is driving toward such a learning health system.
It's a grand vision, indeed. But the interoperability roadmap is still being finalized. The persistent question, then, is how far off is this living organism from reality?
"It's not so far-fetched that we won't see it in our lifetimes. We've been seeing incrementally more interoperable systems that can be used for analytics," Ohno-Machado said. "It will not come with a bang but will come slowly. I think within five years we'll see it."
Under the leadership of NIH, PCORI and NSF within that time "we'll be able to show outcomes we've never seen before," Hanss said. "It will be qualitatively better than what we can do now."
Gazing ahead 10 or even 20 years, Barnett envisions much deeper understandings regarding common roots of disease and compliance to physician directives that trigger substantive behavior changes.
"We know what it looks like," Barnett said. "We just have to scale it."
Previous Innovation Pulse columns:
ICD-10: Pick your own adventure
The bold art of build your own IT
The high road to information interoperability
