Is the healthcare industry ready for Hadoop?
This big question looms whenever practitioners all join together to discuss advances, changes and needs in the industry.
In some ways, the answer is “yes.” There is a need, a desire and a significant opportunity here. However, in many ways healthcare organizations are just not there yet because they’re still addressing core data management and analytics issues around accountable care, readmission control, cost containment and resource staffing.
This was evident in the case studies presented during the Healthcare Analytics Symposium this week in Chicago. Throughout the 2 day conference a variety of healthcare organizations spoke about how they are leveraging data and analytics to solve real problems, which was fantastic. There is definitely a transition taking place where data and analytics are being used effectively. With all of that being said however, none of these organizations spoke about tapping into anything beyond structured data.
That is except for Dr. Mark Wainwright, founder’s board chair in neurology and director of the pediatric neurocritical care program at Northwestern University Feinberg School of Medicine. Dr. Wainwright did not speak about Hadoop or unstructured data, but he did dive into another big data capability: streaming analytics to unearth trends in physiological data that indicate onset of critical issues like sepsis, atrial fibrillation, brain inflammation and seizures in critical care pediatric patients.
Dr. Wainwright talked about the program under development in the pediatric neurology department at Feinberg where Excel Medical Electronics aggregates physiological data in real time from multiple devices and monitors, feeds the data to IBM Infosphere Streams which applies complex analytical algorithms developed by Dr. Wainwright and CleMetric and, in turn, delivers intuitive visualizations at the bedside or mobile device to raise alerts for the onset of life threatening conditions. This, and similar projects underway at UCLA and Emory University, is expected to transform how care is delivered and improve mortality.
Getting back to Hadoop
It is early, but we are definitely seeing providers, payers, pharmacy benefit managers and pharma companies embracing Hadoop (specifically IBM’s Infosphere BigInsights) as part of their overall information management, big data and analytics strategies.
For example, payer organizations are beginning to leverage Hadoop to capture customer sentiment in call center interactions and coupling this data with interactions captured through transactional systems to predict churn. Provider organizations are beginning to leverage Hadoop to predict hidden medical conditions and diseases, identify patients that have recalled implants, enhance the 360 degree view of the individual and optimize IT with data warehouse offloading, machine learning and log analytics. One other fascinating use of machine learning is around quantification of behavioral analytics, studying speech patterns to predict psychotic episodes like post traumatic stress disorder and other conditions.
So, the opportunity to leverage big data capabilities like Hadoop and streaming analytics to unearth trends that were not otherwise found in or connected across physiological data, speech patterns, biomarkers, basic clinical data, demographics, socio-economics and behavior is now here and organizations are beginning to explore the possibilities. And finally, the partnership between IBM and Apple will accelerate the opportunity to leverage mobile devices to understand the individual and engage more effectively to transform health and transform care.
- To learn more about the analytics progression in healthcare, read Harnessing big data for healthcare - Building a foundation for breakaway analytics capabilities
- For information on how to get started, read this Solution Brief
- Learn about the typical Big Data & Analytics progression path and download the new white paper: "The rise of the machine data. Are you prepared?"