Nov 28, 2014


Announcer: These are the conversations happening inside health care that are going to transform health care. The health care insider is on The Scope.

Wendy: My name is Wendy Chapman and I am the chair of biomedical informatics at the University of Utah, and we have a guest here, Dr. Ross Koppel. Dr Koppel is here to give us a talk in our seminar series on becoming a better learning health system and we're focusing on how can we learn from our actions for the data we collect, so that we improve health care delivery and also the health of communities. Dr. Koppel is faculty at the University of Pennsylvania. He's in the Department of Sociology. He's an affiliate professor of medicine and a senior fellow at the Leonard Davis Institute of Wharton. He leads many projects dealing with health information technology or health IT.
He's a very well-known health care informaticist who has hundreds of publications and dozens of projects he leads in this topic, so it's a pleasure to have him here today. We'd like to your thoughts, Dr. Koppel, on the electronic medical records system which, right now, there's a rise in the uptake of electronic medical record systems in the United States. With that, comes many opportunities to help collect data, share data and learn from data and I'm wondering your thoughts about how that has changed the practice of medicine and helped us become more of a learning institution.

Dr. Koppel: Well, thank you. It's a pleasure to be here, Wendy. HIT, or healthcare information technology, is the bedrock of a learning health care system. It has all of the information that we'd want from patients and of their treatments and medications, hopefully, their history and all of that. Without the EHR, electronic health record, as the basis, we really can't have a learning health system because we need all of that information from which to learn about what we're doing right, what are doing wrong, about the ideology of disease, about what methods seem to be efficacious, what seem to generate unwanted side effects and the like.
That said, there are the same problems with electronic health records that affect our ability to use them for learning health care systems. So, for instance, the fact that we didn't focus on interoperability, on the ability to read data and accept data from other healthcare systems, or even from across the hallway, means that a full record of the patient's problems is not available to the physician, or the group, treating the patient when she or he is in a particular hospital. We don't know what the other medications are. We don't know what previously has been done because we didn't focus from the get-go on data standards, on data format standards, enhance on the ability to share information across health care systems, or even, as I said, sometimes, across the hallway.

It's a, sort of, Tower of Babel, or really a city of Towers of Babel, that comprise the healthcare systems' IT. And from that, we are blind and deaf to a lot of what is going on with a patient, since patients see more than one doctor, more than one physician group, more than one hospitalization. And of course, America is particularly challenged in this regard because we are the only health care system, the only nation, really, that doesn't have a unique patient ID and without a unique patient ID we really don't know who was sitting before us.
My own health care system which is preeminent and superb, University of Pennsylvania, we counted 35 ways of recording a patient's name. So, there's "Koppel, R.", "Koppel, Ross", "Ross Koppel", "Ross J. Koppel", "Koppel, R.J.", and it goes on and on like that, and then, there's also spaces and the like. So, the lack of ability to combine data, just the data itself, assuming we knew the patient, is exacerbated by the lack of ability to know which patient we are seeing before them.

We could mine the vast oceans of data about people in EHRs now and so instead of doing a study on 30 people taking a particular drug with a previous history, we could have the, more or less, 310 million people in America and mine that data collectively. But we can't, because the data is not shared across. I would venture to say, that we could accelerate the progress of medical research by 10, 20, or even 30 years, if we could mine the data that has all of the information, about all of the patients, all of their treatments and all of their drugs, but we can't because they are in separate silos, in separate Towers of Babel.

Interviewer: What should we be doing about this?

Dr. Koppel: I think we have to ask the people in charge of regulation to focus not so much on how many places have implemented HIT, but rather on the quality the HIT, on the usability and usefulness, a word we haven't touched yet, usefulness of the HIT, and on how well it works with the workflow that clinicians are engaged in. There's no reason why the software can't accommodate what makes sense for patient safety and for clinician efficiency.
The other thing we should do, is we should have clear testing standards and certification processes that make sense. Currently, the certification process is sort of pathetic. You know, I used to joke when these systems came out that every physician who had a brother-in-law who knew VisiCalc and owned a garage can build an EHR. The market is still open for all, and there are these legacy systems out there which are not terribly usable, often. Sometimes they are wonderful, but they are being used by one person. They talk to nobody. They go nowhere. The learning health care system that you talked about in the beginning is totally thwarted by these systems.

Announcer: is University of Utah's Health Sciences Radio. If you like what you heard, be sure to get our latest content by following us of Facebook. Just click on the Facebook icon at

For Patients