Why a learning health system is important for patient care

A report published this past December by JASON--an independent group of scientists that advises the U.S. government on science and technology--states that the ultimate goal in healthcare is to achieve an "agile, national-scale 'Learning Health System' for identifying and sharing effective practices of care."

The organization's previous reports called for building a robust infrastructure, saying that Meaningful Use Stage 3 should embrace interoperability at its core as a basis for that learning system.

A range of stakeholders aiming to galvanize a national grassroots movement to create a learning health system, now known as the Learning Health Community, grew out of a May 2012 summit focused on reports from the Institute of Medicine and sponsored by the Joseph H. Kanter Family Foundation. 

That community envisions a public-private partnership that involves harnessing big data and feeding knowledge gleaned from collective experience back into the system to inform better care.

After serving nearly 20 years as executive director of the North Carolina Healthcare Information & Communications Alliance, and as a board member of the National e-Health Collaborative, Holt Anderson is now heading efforts to develop the governance and policy framework required for such a learning health system.

Anderson spoke with FierceHealthIT about that work.

FierceHealthIT: Creating a learning health system is not just a matter of having health records on computers. How would this system work?

Anderson: What if, when you went to the doctor, they were able to find out of hundreds of millions of records, people with a very similar health history and DNA with that same diagnosis? And those other 100 people or 1,000 people treated for that diagnosis had all types of outcomes, ranging from poor to good to excellent. Wouldn't you like to know which of those treatments had the excellent outcomes so that your physician could use a decision support system to determine the best treatment for you?

Physicians and nurses have pointed out to me that this is not cookbook medicine. You can't just take data from one patient and apply it to another patient. You've got to consider social, economic and environmental factors for that individual. You have to factor in whether they're able to comply with any orders. I think it would be a good research topic to figure out how to have structured data that characterizes social and economic factors for an individual, too, so that they're easily comparable.

FierceHealthIT: What is your role in this?

Anderson: The responsibility I've been asked to undertake is to take a deeper dive in some of the specifics of governance and policy.

I'm happy to say the recent [interoperability] roadmap that ONC published a couple of weeks ago is a 10-year vision, and the way I read it, the goal is to reach a learning health system. A significant portion of it is governance and policy. So after working in this space for 20-plus years, I believe we have to develop a learning health system. There is no option. We can do so much better in providing the information that informs a physician or a nurse or public health person to do their job. There's so much capability that can be given to those individuals to boost their confidence that they're making good decisions because they have the information they need.

I've talked about this recently to physicians and nurses, and they say, "We're going to get there in 2024? But that's way too long."

FierceHealthIT: What is the path to a learning health system, from a technology perspective?

Anderson: The first piece has to be standards, that this piece of data means the same thing as the other systems you're exchanging records with. That the data is valid. Then standards among systems that get us to interoperability. Even a hospital might be on a different version of the same technology used by a neighboring hospital. There are all kinds of technology challenges even within the same system. [Some lab companies] report differently for the same test. So it's a standards issue.

There are a number of challenges, including control of data, intellectual property, workflows, security and privacy, and economic incentives.

If you use a patient portal, then you go to a different hospital, there would be a different portal there. At this point, there's not a combined record. People who use the [financial software] Quicken know they can put information from multiple banks, multiple brokers in one place. We need that capability for individuals [in healthcare].

Another factor is competition. Everyone will use these patient portals as a competitive advantage, a marketing tool. It's like your loyalty cards at the grocery store. It's a benefit that people wouldn't want to move because my portal is better than that one over there. Competition is one of these challenges we're going to have to get through.

Patients and their families are getting impatient, and that's a good thing. That will ultimately force change in the system. It's up to the will of the general public and their representatives, health professionals as well as payers and others to accelerate this change. The technology is there; the innovations are there. So it's the ability to coalesce around a set of standards and policy that can overcome inherent business protective tendencies and economic factors. There has to be economic return on investment.

Anyone in this learning health system, if you look at it from a selfish perspective, is going to have to give up something. Their information about their patients, their procedures, their successes in coming out with a better treatment, which could be a competitive edge. Is the benefit of being able to see the treatments and outcomes for 100 patients worth my giving up my one patient. This learning health system is all about Six Sigma, it's all about learning.

FierceHealthIT: Do you see this interoperability roadmap as being an effective way to get there?

Anderson: We're doing analysis of it now. We have nine work groups just on policy and governance. Others are analyzing it from other perspectives. It's an open comment period. Comments are due April 3. The public needs to weigh in on this. But it gives us a place to collaborate, to dialog and to debate this. If we can engender the public understanding, the will and passion to improve, this roadmap can be a good start.