Dr. Krumholz sat down with Austin Jaspers of ISPS Health on a recent morning to discuss big data and improving patient outcomes, as well as the nexus of scholarship and medical practice.
Harlan Krumholz, M.D., is a cardiologist and Director of the Yale-New Haven Hospital Center for Outcomes Research and Evaluation (CORE). At Yale School of Medicine, Dr. Krumholz is also Harold H. Hines Jr. Professor of Medicine, Professor of Investigative Medicine and of Public Health, and Co-Director of the Clinical Scholars Program.
Austin Jaspers: Your field is cardiovascular medicine: Today, how is data collected for cardiovascular patients and how is it used to improve care for the next patient?
Harlan Krumholz: For the most part, the data is collected haphazardly and the only systematic collection in the United States is the billing data. There are some registries which collect information about selective groups of patient which can funnel into quality improvement programs. We are still in need of a unified system that provides reliable, accurate information about the care of patients. That can be used to assess our performance today and improve for tomorrow. We’re still trying to find ways to learn from our practice and translate that learning into improvement.
AJ: Using cardiology as an example, how would access to more data impact doctors’ work?
HK: Currently, we’re using billing data to characterize the performance of hospitals and try to identify where there are gaps in outcomes. But there is a need to understand what the underlying causes are and to develop strategies to elevate practice That is best done with data that is available in real time, that includes clinical information about patients that is constant across sites, and that is available to people who want to investigate ways to improve care.
I would say that the cardiovascular field has been ahead of most in its use of data. Groups of physicians do collect data. But it’s done by brute force: The data is collected often by nurses and other health care professionals on charts, but there is no good way to get good data from electronic health records, yet. With billing data, we’ve been able to use clever approaches to elucidate valid information, but I think we can do better.
AJ: Would patients always benefit from increased access to data, about their own health and the health of others with similar conditions and circumstances? How would this affect decision making?
HK: There is always the concern of the harm of data breaches, where people’s personal information becomes available in ways they do not permit. If you can ensure data security, there’s no question that people would benefit from more wide-availability of data about health care. The main thing we waste in medicine is clinical experience. We usually fail to learn from the experiences that patients, doctors and health care professionals have every day – because it is not collected in systematic ways. If we were able to solve the problem of collecting the information in ways you could trust the data and the security of the system, there would be benefits for everyone.
AJ: What areas in medicine and health care have been most successful in the integration of data feedback loops?
HK: All groups, including cardiology, have been slow to include patient-reported outcomes in their databases. This information goes beyond survival data and tells doctors how people feel. The medical community is starting to collect more information. The surgical community – particularly for different types of surgery – has been collecting data. The Society for Thoracic Surgeons was one of the first to begin collecting data.
AJ: The Yale Open Data Access Project (YODA) has made tremendous inroads in the area of leveraging big data to improve health care. Who else – in government, academia or the private sector – could help usher in the use of big data in medicine?
HK: People are increasingly appreciating the importance of making data available for the public good. First and foremost, we don’t have enough mechanisms to bring in the kind of data we need. There are two ongoing parallel efforts: 1) Improve the quality of data that exists and 2) make data that exists more widely available. Some people are nervous about greater access to data. Some people are served by having the data be limited – people who generate and own the data and people who are concerned with what results might emerge from it. Academia, too, is slow to embrace the idea that we should be sharing data. The culture is starting to change. The EMA in Europe is leading the way, doing things that the FDA is not able to do. We’ll have to stay tuned to find out what happens.
AJ: Do you have reservations about how private sector companies could use large data sets? What policies should be in place to ensure fair, effective use of these data?
HK: YODA is focused on the release of clinical trial data. The big data that’s out there – the societal data – tracks many aspects of our lives. This is a big source of concern for many people, including me. I would like to move toward a situation where people can opt-in to opportunities for others to use their data for medical research. We want to respect their rights and their agency over the data about them. Right now in business, there are no rules. I’d like to see medicine begin to leverage the large bandwidth of information being collected and set an example for how we can protect people’s rights.
AJ: Could Yale use supercomputers to process and analyze big data sets? What’s next?
HK: We are developing a Center for Big Data at Yale. We will take advantage of the expertise in data that relies in virtually every area and discipline. We want to solve some of the fundamental problems about advancing methods and seeing greater dissemination and application. No single group is going to be able to approach this on their own, so we need to come together. There are large scale questions about how we can use inductive reasoning, instead of deductive reasoning often promoted in medical science. How do we deal with the ethics and legal issues of the situation? How do you integrate in a way that incentivizes patients, doctors and researchers to seek the information so that they can be better informed? We want to create a platform that takes advantage of one of the unique features of a university, which is the vast diversity of expertise that exists in a concentrated geographic area. We’re blessed with remarkable talent and we want to see how we can contribute to the future of big data in health care.
AJ: What role can the Yale ISPS Health Policy Center fulfill at the crossroads of scholarship and practice?
HK: I think Zack Cooper is doing a great service by working to create a network of people who are communicating and collaborating at Yale. I have great hopes that we do not ever isolate people who are working on research to improve health care. Individual contributions can be so much larger if they cross fields and connect with others.