Population Health Priorities from ONC's Own Dr. McCoy

The federal government's chief health information officer is shaping ONC's interoperability roadmap. It's a foundational step to achieving goals such as precision medicine. But ONC's priorities don't always align with those of healthcare providers.

This week, to understand the future of population health, I turned to Dr. McCoy. Not the guy on the Starship Enterprise. This is the real-life Michael McCoy, MD, chief health information officer for the federal Office of the National Coordinator for Health Information Technology.

When I first asked McCoy for his thoughts on population health, he brought up a couple of HHS's big-think current initiatives: precision medicine and semantic interoperability. Here's a quick primer gleaned from our conversation:

• Precision medicine: McCoy says it takes personalized medicine to the next level. "The precision is in informing both the caregivers and the people receiving care what really works and should work for them, given a number of factors. How much of a dose do we need to give a person to treat whatever? One dose of whatever drug probably isn't the right amount. It's either too much or too little for 95% of the people that are taking it." Likewise, the duration of drug treatment might be too short or too long. By integrating genomic data into medical records, then analyzing those records across a population, precision medicine aims to cut drug costs and optimize treatment effectiveness, McCoy says.
• Semantic interoperability: "The ability to gather large amounts of good data is also predicated on having the right information about the individuals, and having it shared in a semantically interoperable way. Semantic is taking it down to the level where we mean the same thing when we say something."

As an example of creating a health system with semantic interoperability, McCoy points to the different nuances of sharing information on blood pressure readings.

"It's two numbers, systolic and diastolic. So do you do that as 130 over 70 with a slash between them, or do you make that two separate fields? Oh, and by the way, was that blood pressure taken by a blood pressure cuff? Was it taken by an interarterial pressure monitoring device? Was it sitting, standing, lying, after running? Was it on a treadmill? So you have a lot more information that you can convey that just a numbers. There's metadata that goes around that."

Another example would be recording the number of cc's of blood in reporting an incident of excessive blood loss. "A 400 cc blood loss in a vaginal delivery is not excessive, but a 600 isn't for a C-section. So when we talk about these kinds of things, you have to have a contextual understanding of more than just the data that went with it."

These are the kinds of interoperability issues ONC is wrestling with in its ten-year interoperability roadmap. Since McCoy is ONC's clinical subject matter expert for interoperability, he is in the middle of shaping the roadmap.

I mentioned criticism to McCoy that I've heard widely about the precision medicine initiative—that it puts the cart before the horse, assuming the patient's longitudinal record is in place to be married with genomic data, when clearly, due to interoperability barriers, those longitudinal records just aren't ready yet.

To illustrate, I mentioned some of the results of a pre-meeting survey that HealthLeaders Media conducted of attendees of our Population Health Exchange, which takes place later this week in Carlsbad, CA. Attendees said their top challenge was inadequate or incomplete longitudinal clinical data. How useful can precision medicine be, I asked McCoy, if that remains an unsolved problem?

"It will be very challenging to make it as widespread and as desirable as we would like for an outcome," McCoy says. "It's a linchpin for having the person-centric care, so that the person at the center of that care can direct the care to the people that they want caring for them, as clinicians [and] as hospital systems."

In our survey, we also asked our attendees what their top three data-related challenges are in the next three years. Number three, behind EHR interoperability and data cleanup, was analyzing unstructured data—which has not been on ONC's radar nearly as much as making structured data interoperate.

"Clearly, there are opportunities with natural language processing and other modalities to analyze" unstructured data, McCoy says. As an example, unstructured notes containing a history of present illness or care plan can explain whether a broken arm came from a simple fall or something that an athlete just experienced while performing. "How this is best done is a moving target, and I'm not sure that ONC at this point has any better guidance on some of that than the rest of the industry does," he says.

"We talk about the learning health system as the end goal at the 10-year mark … and that is an ambitious goal," McCoy says. "I'm not suggesting that we could tomorrow have a precision medicine ubiquitous success, absent the foundations of interoperability. The first three years is all about getting interoperability and the governance for the data standards well established. That has to go in concert though with the delivery system of reform, so that we tie the incentives to alter business practices that currently are not in alignment with sharing data.

"There's no great incentive for a large hospital system that's the dominant player in a market to want to open up its data to take care of people who may also go to competing hospitals. We have much more of a cultural and business practice challenge to overcome than a technical one. Not to say that there aren't some technical issues. I do admit that we have some work to do there as well, but I think it's less a problem than the people and process and business practice issues."

Of course, progress is unevenly distributed across healthcare. Large academic institutions who are doing well with their EHRs, such as Duke, Stanford, and the Mayo Clinic, are pushing ahead on the precision medicine front, he says.

Of course, the interoperability roadmap is itself a living document for the moment. ONC is saying a new version of the roadmap will be out later this summer. Hundreds of pages of public feedback to the initial draft are available at healthit.gov. It's a lot to comb through, and after sampling it over the past few days, I can safely say Dr. McCoy will need some real luck in getting the nation to agree on a succinct set of priorities to move federally driven interoperability initiatives forward.