Health Care Insider: Learning Health Systems: More Affordable Health CareFeb 14, 2014
When people hear about cutting costs in health care, the assumption is often that quality of care also suffers. Proponents of a learning health system say that it will actually increase the quality of care while decreasing costs. We asked Charles Friedman what a learning health system is, how it would improve heath care and obstacles that need to be overcome to implement it.
Interviewer: A more affordable healthcare system? Better care for few dollars, is that even possible? We're going to examine that next on The Scope.
Announcer: Medical news and research from University of Utah Physicians and specialists you can use for a happier and healthier life. You're listening to the scope.
Interviewer: It's something that's on everybody's mind: how can we get better healthcare and spend fewer dollars? Well, Dr. Charles Friedman says it's with a learning healthcare system. I'm intrigued by that title. A learning healthcare system, if I understand correctly, is one that's constantly taking in data from physicians, patients, research, demographic data, evaluating it, which gives us a better picture of healthcare needs, and will provide information to patients and providers which would mean better care at decreased cost. Is that a fair summary of what a learning healthcare system does?
Dr. Friedman: Yes, that's exactly right. I actually can't say it better myself. I want to emphasize, though, that it's all about making people healthier and doing so, as you said, at lower cost. We can do this now, and maybe it's worth talking about what's different and why we're having this conversation.
Interviewer: Yeah, I think that's good. So from what I understand, there is a lot of data out there, but we're not using it very well.
Dr. Friedman: Correct. Heretofore, the data we had about healthcare and what actually happened when patients interact with care providers was trapped in paper records.
Dr. Friedman: Think about this for a moment, what it takes to do a study that looks across the experience of different people so you can learn from those experiences and improve. What you had to do is have someone first find all those paper records, open them up, find the data of interest, which is often hard to find, and then code it and eventually put it into computers so it can be analyzed. The big difference now is that the nation is headed toward a system where care is routinely documented in digital form, which means we are in a much better position to study what actually happens and learn from that.
Interviewer: Yeah, and it kind of blew me away as I started learning more and more about this. A lot of other companies are very successfully using data from all different areas, but medicine really is kind of behind in that. The information my doctor has isn't compatible with genetic information the state might have or cancer records that might be someplace else. Why is that?
Dr. Friedman: Well, there are a lot of reasons for that, and it is the case that probably health is the last sector of the economy to go digital. There are a lot of reasons for that, and among them is the fact that healthcare is very complex. And the information that is collected has a lot of what we call nuance to it. It's very hard to get a precise definition of what a clinician means when he or she says something about the state of the patient. We're developing techniques to do that now that we didn't have before. So, a lot of things are happening that are that are good trends.
We're overcoming some of these barriers; the healthcare system is going digital. Most experts believe that by 2019 or 2020, we will have an 80% digital system, which means that 80% of the care that is delivered in this country will be documented digitally, which means from the point of view all the learning healthcare system, those data will be amenable to analysis without going through all the steps that were required when the data were captured on paper.
Interviewer: Which makes it much faster. So then in real practice would that mean that, well, you've got physicians all the United States treating people for a specific condition and they might discover, "Well, we're prematurely treating this condition," that data would actually help show that right?
Dr. Friedman: Right. Those kinds of results would be very, very easy, and very, very rapidly unearthed if we had this kind of learning system where, in an ideal situation-but I think it's an attainable situation-we are continuously looking at the data that emanate from the actual care that people were receiving. Now this has to be done in a way that respects patients privacy, that usually, I should say almost always, looks at data that are de-identified, because to do these kinds of studies, and this is very important emphasized you don't have to know who the people are. You're just putting together information about people in a way that makes this information comparable. But you don't really need to know the identity of these individuals. So, inherently, this is a process that preserves privacy.
Interviewer: And with this information, you can actually get a more accurate picture of the state of healthcare, the state of our health, what works and what doesn't. Right now, we're just kind of guessing, right?
Dr. Friedman: Correct. An example I give of what wasn't possible a decade ago and what is possible now or is increasingly possible now is the story about Vioxx. And many people may remember that Vioxx came on the market in 1999 and almost immediately caused an increase in the rate of heart attacks in the country. The problem was this was hard to detect quickly. It actually took many years before the association between Vioxx and heart attacks was detected and, in fact, the healthcare delivery organization, Kaiser Permanente, which detected it, because of advances they've made, has many of the features of the learning healthcare system.
So within their own population of nine million patients, they were able to see this. So that's a really good example of what we're going to be able to do with a learning healthcare system, putting our data immediately to work, to learn how to do things better. Now, a Vioxx kind of problem is relatively rare, but when things like this do happen, I'd like to think that with a learning healthcare system we will detect them much more quickly than we did with the Vioxx problem.
Interviewer: For me, it makes sense how this would mean better care, but how does it mean lower costs?
Dr. Friedman: It's all about health in the end, and it's all about making people healthier. But if you study the processes in the same way that you study the actual care that is delivered-for example, the medicines that people receive-you can see which processes work better and in that way, you can take cost out of the system without compromising the quality of care.
Interviewer: So, for example, you can see a bunch of physicians prescribed drug B, which costs less, but gave just as good of results as more expensive drug A. You'd be able to easily see that?
Dr. Friedman: You'd be able to see that, and you'd be able to see, for example, around process of care that perhaps in a particular case patients who were sent home after a certain treatment did just as well or maybe even better than patients who stayed in the hospital.
Interviewer: All right. What can a patient do? What can I do to help facilitate this?
Dr. Friedman: The learning health system absolutely requires patients-everybody, because everybody sooner or later is going to be a patient-to be part of it. In fact, many people envision ways in which patients themselves can use the learning health system to help them make decisions.
For example, suppose a man is diagnosed with prostate cancer and is offered many treatment alternatives. Right now, it's very, very hard to understand what treatment is going to work best for someone very much like you. You can read the literature, but the literature and the studies that are done are about very, very large populations of people, and the results are about what happened to all of the people in that population. You can't, from the studies, drill into that and say, "Okay. There were a few people in there who are just like me. What happened to them?" So many people believe that individuals can use the learning healthcare system directly to help them make decisions. They can press what would be akin to a patient's "me" button. Find patients like me. From that, tell me what choices they made. And, as a consequence of the choices they made, what actually happened. This is real data. This is real experience, not what somebody happens to remember, which is a lot of what patients have to go on now. Or what happened to their friends and other people they now.
Interviewer: So as part of me being a participant of this when I go into my doctor's office, they ask me to sign a form that said, "Can we use your medical information?" Is that part of what a regular person could do if they think this is a good idea and they'd like to see it move forward?
Dr. Friedman: It's not exactly clear yet what form of consent will be required. It may be the case they'll have a situation that's called an opt-out. We would expect, as the benefits of this become known and as the threats to individual privacy are known to be very, very small-not zero, but small-most people will see more benefit in being part of the system. They will see that greatly outweighing any risks associated. So we hope and fully expect that very few people will opt out as the system develops. I want to go back to a question you asked earlier. What is actually happening? Well, there is a growing national movement toward a learning health system, and it's seen in a lot of places. A number of books and reports and plans all refer to the learning health system now and call it out as an imperative. Many organizations within their own boundaries-I've mentioned Kaiser Permanente-are developing the capability to put their data to work within their own boundaries to study their processes and their treatments and based on the study and the results of the studies, actually improve what they're doing. Equally important, as we think about this eventually becoming a national system, a number of learning networks are developing that have members that come from a range of organizations. So we're seeing learning develop within the organization and we're seeing learning develop across organizations. And these organizations are developing ways of sharing their data and learning from all of the data exists in all of the organizations. And this is very, very important. I really do believe that these trends are combining to begin a march. It will be a long march, but it will eventually get there, toward a learning health system for the nation.
Interviewer: We have a couple seconds left. Do you have any final thoughts or is there anything I left out?
Dr. Friedman: It's very important to emphasize that the most important word in the learning health system is health. This is all about making people healthier, making the healthcare system safer, assisting public health. For example, some uses of the learning health system that we did not discuss up to this point will meet the needs of the public health system.
For example, when an epidemic like a flu epidemic happens, the learning health system will enable the whole nation to track the spread of that epidemic much better, predict where it's going and actually alert clinicians who are in an area not yet reached by the epidemic that the epidemic is headed their way. That's, of course, very helpful to any clinician who has to properly diagnose patients who come in with symptoms as to having the flu, which might be the disease of the epidemic or not. So there are many ways the learning health system will be useful in the future.
Announcer: We're your daily dose of science, conversation and medicine. This is The Scope, University of Utah Health Sciences Radio.