Big Data’s True Promise: Better and More Individualized Care
At a recent HIMSS conference in Washington on big data and analytics, I found myself vividly remembering two young college students, covered in black and blue blotches and desperately struggling to live. It was years ago on a beautiful Saturday morning in June, in a college town drawn from a Saturday Evening Post cover and I was the emergency physician on duty at the local emergency department. As the experts at the conference heralded the new era of big data capture and analysis, images of these students were playing back in my mind.
What is big data and how will it impact care at the bedside? I’m not sure the conference gave a definitive answer. As Adi Ignatius, Editor in Chief of the Harvard Business Review, described in the October 2012 issue, there is a “huge interest in big data, but also a lack of clarity about it.”
The first challenge of big data is freeing yourself to appreciate the vast dimensions of data that is available. It has been estimated that Walmart collects more than 2.5 petabytes (PB) every hour. Clinically, a single patient when genomic and imaging data is included may exceed a terabyte (TB), and much of the additional patient data is in unstructured form. There are data that can be measured directly and recorded from the patient, including his past history and therapies.
Beyond clinical, physiologic metrics, we ought also to capture the data of all that affects the patient. Much of this expanded data will be unstructured such as is present in social network data set or quantified but predicted such as weather reports and pollen counts. Some may reasonably be skeptical of the clinical impact of capturing this hyper data set.
We have come of age in medicine capturing and analyzing physiologic parameters such as blood gases, making predictions on the information generated, and instituting therapies off these predictions. In managing the resources of healthcare we become facile at matching resources to predicted needs. We readily map projected service line volumes with cost and revenue data to create sustainable decisions. This is the simple use of data, consistent with the Peter Druckerism “you can’t manage what you don’t measure”.
Redefining Patient Data
Big data takes us further, expanding the data set, continually monitoring parameters in real time, and processing with user-friendly decision support systems to enhance the accuracy, efficiency and effectiveness of therapies and interventions for the care provider. As we move from the traditional fee-for-service model to being compensated for maintaining the health of our communities, heat waves and pollen counts become important.
But there is yet more to be harvested from truly BIG data. This brings me to the desperately ill university students I was treating that Saturday morning. Their lesions, bleeding into the skin caused by vascular injury from the infection, were suggestive of the most lethal infections, meningococcemia, a bloodstream infection by Neisseria Meningitidis.
Beyond providing care for these critically ill patients, there was also a need to prevent further transmission of the disease. To do so we had to identify the close contacts of these two patients and begin antibiotic chemoprophylaxis, preferably within 24 hours. We mobilized the human resources of the University, the health department and township police to painstakingly interview and reconstruct all the contacts of the two patients. This was a process of days not 24 hours.
Flash forward to 2013. Given the same presentation of two college students with meningococcemia, how much improved would our care be if we had access to their Twitter and Facebook data as we sought to identify those who had close contact with the students?
I’ll never know for sure, but I’m encouraged at the innovation evident at the conference and all around us. Public health officials are already using social media to track infectious disease outbreaks. At the bedside, IBM is analyzing real-time data streams from patient monitors to identify symptoms of traumatic brain injury.
Big data is everywhere and a new generation, some who are perhaps as young as the students I treated that day, will harness it for better, faster and more individualized care.
Frank X. Speidel, MD, MBA, FACEP is Chief Medical Officer for Healthcare IT Leaders and a former hospital CEO and CMO.