In November, CURE and Yale, in collaboration with Boehringer Ingelheim, presented “Patients and Big Data in Healthcare: Deriving Value and Accelerating Innovation.” This panel discussion, moderated by Richard Foster, PhD, Emeritus Director at McKinsey and Co and Lecturer at the Yale School of Management, brought together forward thinking minds from across healthcare industry and academia for a discussion on the challenges, successes, and potential that big data in medicine has to transform the way healthcare is delivered, and financed.

“Big data” represents data that exceeds the conventional computing capacity of traditional database systems as a result of volume, velocity, or variability (The Three V’s). Next generation analytical tools and models are necessary to make sense of this data and uncover previously hidden patterns, relationships, and information. There was a general consensus among panelists that healthcare lags other industries in harnessing and acting on its reservoirs of data.

Dr. Krumholz began with a look at the reductionist approach to much of the research in medicine, “where we strip the complexity of the data and miss the richness; we don’t take it as a systems biology approach where there are multiple interactions among the thousands and millions.” Risk scores that present patients with estimates for conditions ranging from cardiovascular disease to musculoskeletal disorders typically take into consideration a handful of variables, whereas the reality presents a much more diverse picture of complex interactions. Even with the data that we do have, we still see marked disparities in health outcomes among institutions, clinics, and individual providers.

“In the everyday operation of healthcare, we generate data that we are failing to use effectively to promote the best delivery of healthcare, personalize it in the ways we can, and be accountable for the results that we are achieving.”

Harlan Krumholz MD, Professor of Cardiology and Director of Yale-New Haven Hospital for Outcomes Research and Evaluation

Mr. Matteo of Optum identified the shifting landscape of healthcare financing and delivery as holding promise for more effective data utilization. As fee-for-service models make way for value-based compensation in which payers and providers share risk, there will be greater inclination to share data across stakeholders. The integration of clinical, claims, administrative, and socioeconomic data can serve as a boundless reservoir of powerful data that are invaluable for researchers and clinicians.

“It’s the biggest opportunity I have seen in healthcare in 20 years. While every other industry has harnessed big data in a way that focuses on the end consumer, we have been slow to adopt these practices. How can we remove these blockers and accelerate the use of big data to achieve a better patient condition?”

Mike Matteo, Chief Growth Officer at Optum

Other panelists, including Rishi Bhalerao and Fay Lindsay, gave their input as well, based on their professional experiences:

“At PatientsLikeMe, we are aspiring to make the human condition computable. We are exploring ways to treat healthcare as a mathematical model and create mathematical models of disease?”

Rishi Bhalerao MBA, Director of PatientsLikeMe

“Big data is only going to get bigger, and as it keeps growing we seem to not know what to do with it. Some industries are better at mining big data, but healthcare is leaving a lot on the table in that regard.”

Faye Lindsay MBA, Lead IT Business Consultant, Managed Markets Team at Boehringer Ingelheim Pharmaceuticals

“In addition, it is not uncommon for patient data being outsourced to third parties, who providers have to pay to retrieve when needed,” Dr. Feuerstein explained. These practices are happening on a systematic level, and present a challenge to success. Often, large employers purchase services as efficiently as they can, resulting in many different partners for each of the services. How can we reconnect those data sources, focus them on the patient, and provide actionable steps to providers at the right time?

We already know of lots of clinical problems through the data we have, and need to be careful not to get distracted from what we can fix today and have big data be something behind the curtain that allows us to ignore what data shows to be the problem.

Seth Fuerstein MD, JD, Assistant Clinical Professor of Psychiatry, Chief Innovation Officer, Magellan Health Inc.

nov-bio

Dr. Krumholz closed out the discussion by charging the community to come up with apps that can take large-scale population data, embrace the complexity, and identify patterns of utilization or changes in habits evolving over time. Early interventions are invaluable, and can be more accessible with this type of information.

If we’re going to realize a system of truly patient-centered care, we need to move beyond what happens within the four walls of the clinic and appreciate and embrace the limitless data continually being generated by individuals.


By Shervin Etemad (edited by Amogh Sivarapatna)