Compared to other industries, healthcare as a whole has been a late adopter of big data predictive analytics. This may be due in part because of concerns about patient confidentiality and fears of security breaches on open source Hadoop-based systems. Demonstrating a return on investment (ROI) for big data solutions to hospitals and healthcare providers can also be a challenge.
Ironically, most healthcare organizations have an abundance of patient data at their disposal that they could use to benefit patient care and medical research while driving down costs...if they could only implement a big data solution.
This paper shows how one healthcare institution—Penn Medicine—is using big data solutions to derive new insights and improve patient care.
Penn Medicine is a $4.3 billion organization with more than 2,000 physicians providing services to the Hospital of the University of Pennsylvania, Penn Presbyterian Medical Center, Pennsylvania Hospital, Chester County Hospital, Lancaster General Health, and a health network that serves the city of Philadelphia, the surrounding five-county area, and parts of southern New Jersey.
Considered a leader in healthcare predictive analytics, the Penn Medicine data science team is dedicated to improving patient outcomes through analytics. Specifically, they want to harness the full power of clinical data to help clinicians identify patients at risk of critical illnesses that may have been missed by current diagnostic techniques. In the process, they are also developing solutions that will remove the barriers to developing analytic models and accelerate the deployment of analytic applications based on those models, all of which they intend to share with other health organizations via Open Source.
To accomplish this, Penn Medicine needed a platform for rapid development and deployment of predictive analytics applications that could be applied to detect patients at risk of critical illnesses. The platform they developed is called Penn Signals. Penn Signals is a collaborative data science platform developed by the Penn Medicine data science team that combines clinical data at scale with big data to allow researchers to explore solutions, allow developers to develop predictive applications, and provide a platform for deployment. The first applications of Penn Signals focused on sepsis and heart failure.
To gauge the success of Penn Signals, Penn Medicine decided to run clinical pilot case studies involving two very different acute events—sepsis and heart failure. They hoped these pilots would provide tangible improvement to patient care that might convince other healthcare institutions of the viability of such a big data solution.
According to the Centers for Disease Control (CDC), sepsis (blood infection) affects more than a million Americans annually and is the ninth leading cause of disease-related deaths and the #1 cause of deaths in intensive care units. Worldwide incidents exceed 20 million cases a year, and mortality due to septic shock may approach 50 percent, even in industrialized countries. The mortality rate is approximately 40 percent in adults and 25 percent in children.
Treatment guidelines call for the administration of broad-spectrum antibiotics within the first hour following recognition of septic shock. Prompt antimicrobial therapy is critically important, as the risk of dying increases by approximately 10 percent for every hour of delay in receiving antibiotics.
Traditional methods of sepsis identification generally only detect about half of the cases, and even then detection typically occurs just two hours before the patient succumbs to septic shock. A diagnostic process that could help staff identify sepsis earlier could profoundly improve treatment success and lower the mortality rate.
Heart failure (the inability of the heart to pump enough blood to meet the needs of the body and lungs) is amazingly common, affecting 5.