The health care industry is facing a challenging transition as it tries to use data more effectively. Gathering, sharing and analyzing patient information is an incredibly challenging process that is contributing to a technological revolution across the health care sector.
One of the main challenges in our health care system today is the difficulty in working with patient data, because it exists in a variety of formats, both structured and unstructured, and it often exists in large volumes, making it difficult to handle in many data centers. This is where supercomputing comes into play, and the Oak Ridge National Laboratory (ORNL) is leading a project that is taking key steps forward in driving innovation.
According to a recent article in Communications of the ACM, ORNL has combined three advanced computing resources to tackle many of the data issues facing the health care industry. The results are a better ability to process and analyze large data, more powerful data modeling and better capability for data-driven innovation.
Looking at the ORNL’s health data initiative
The article explains that ORNL is using Titan (a Cray® XK7™ supercomputer), a cloud computing platform and its Apollo in-memory graph computing solution based on Cray’s Urika™ platform. All told, the amount of compute and storage power being devoted to the project is considerable, and necessary.
A major limitation of today’s information technology is that information tends to end up in disconnected “data silos,” making it extremely difficult to bring information from different sources together in order to make better decisions and innovate effectively. ORNL is working around this issue through its Urika solution, a graph analytics platform built around the idea of scalable graph computing. Representing structured and unstructured information uniformly as a graph makes it much easier to identify the underlying patterns in the information. This enables researchers to quickly understand their data and advance their programs efficiently.
Using supercomputing to solve key health care problems
The U.S. health care industry is facing many problems, and high performance computing systems could hold the key to unlocking the sector’s full potential. Sreenivas Rangan Sukumar, a research scientist in the Computational Sciences and Engineering Division of ORNL’s Computing and Computational Sciences Directorate, told the news source that ORNL began to recognize its unique ability to contribute to innovation in the U.S. health care industry a few years ago. Since then, the organization has been working to tackle key data problems by finding ways to drive innovation in health care.
“We believe that the transformation of health care delivery toward improved quality and health outcomes has a better chance of success if patients, providers and payers all benefit,” Sukumar told Communications of the ACM. “We are using our computing resources to identify such opportunities.”
Georgia Tourassi, director of ORNL’s Health Data Sciences Institute, echoed this sentiment by pointing out that data problems are only going to get worse in health care, and that ORNL is laying the groundwork for handling future demands through mathematical innovation and the development of future algorithms.
Building a healthier future
Figuring out how to use patient data effectively could prove critical in establishing a more efficient, effective health care system. Analyzing enough patient data can enable health care organizations to use predictive analytics and similar techniques to anticipate care demands and establish preventive care methodologies instead of reacting to patient conditions. The cost savings, health improvements and other advantages of such methodologies are substantial, but they are only going to be obtainable as solutions like graph-based computing are used to make sense of all of the unstructured patient data that is available to research organizations.