Interoperability is a 'unicorn,' says Franciscan Health's Chuck Christian

Chuck Christian has been working on healthcare data exchange and interoperability for decades.

Mr. Christian, the vice president of technology and chief technology officer of Mishawaka, Ind.-based Franciscan Health, has served as an executive with the Indiana Health Information Exchange. He's been on the Healthcare Information and Management Systems Society's interoperability committee. He teaches a class on interoperability at the College of Healthcare Information Management Executives.

Becker's did a Q&A with Mr. Christian for the final installment of its series exploring what it will take to get to interoperable healthcare data exchange in the United States. (Read parts one, two, three, four and five of the series here.)

Note: His responses have been lightly edited for clarity.

Question: Where would you say the healthcare system is now in terms of interoperability?

Chuck Christian: Interoperability means different things to different people. And a lot of people co-mix data exchange and interoperability. Then you have the purists that believe interoperability is one system, just sharing data with another one and it gets automatically incorporated into the workflow of whoever needs the information.

We in healthcare are in the business of patient care, and the whole idea of interoperability is providing the right information to the right person at the right time. Healthcare is very complex; it's not like going to the ATM and taking out 20 bucks. You need your card, your account number and your PIN, and you can get 20 bucks.

But if you're a neonatologist, you'll need information about infants that may be in crisis. If you're a general practitioner or family doctor, you're taking care of the patient from beginning to end, and so you're doing routine visits, routine lab work and then doing referrals.

It is absolutely a complex process, and we've tried to legislate what needs to be done, and then we as an industry depend upon the EMR vendors to solve the problem. All of this was well-intentioned, but in my humble opinion focused in the wrong direction.

When ONC was founded, we started looking at the standards. The good news is we have standards. The bad news is we have lots of standards. And in many cases, we have competing standards.

One of my pet peeves is the fact that in 1996 when HIPAA was propagated as a federal law, it had four identifiers: physicians, hospitals, insurances and patients. Through the power of the appropriations bill, the ability for the federal government to spend money on creating national patient identifiers was nixed.

It's just as bad to give somebody an allergy they don't have as to miss one they do have. If we really want to be serious about interoperability and getting the right data at the right place, we've got to do something about patient identification. There are those that argue that by using demographic information, you can pretty well get to who the person is anywhere between 86 to 92 percent of the time. I'll ask you, would you like to have your paycheck deposited in the right bank account 86 to 92 percent of the time?

It's not the usual John Smiths and Betty Whites who have the problem. It's my name. I've got some good friends who get mixed up all the time; one gets mixed up with her sister-in-law.

The trick is to tell physicians what they need to know. With the technology we have in hand, we should be able to know, based upon either chief complaint or letting the machine understand who the patient is, what comorbidities they have, what disease processes they have, what information the physician treating them needs.

Have we made progress? Absolutely. But there are still competing propositions on how data needs to move. I work in a facility that is an Epic customer. We move data pretty good. Epic communicates really well with other Epic customers; not so much with non-Epic. But large medical centers in systems like mine are typically always going to have Epic, and we can move the data around pretty seamlessly. The way that gets done is controlled by Epic, and with some other integration engines we have we can continue to do that.

We also happen to live in a state that has probably one of the oldest and best HIEs in the country now. The Indiana Health Information Exchange has been around 30-something years.

It was started by a bunch of physicians who wanted to be able to see what was going on because we had a lot of ERs in Indianapolis: Methodist, Eskenazi, IU. And patients would frequent all three of those emergency rooms. So the physicians needed to see what happened to that patient previously, to see if they needed to be treating them differently based upon what medications or what labs they'd been given, so as not to increase the total cost of care.

Q: What was the opposition to a national patient identifier?

CC: This goes back to '98. U.S. Rep. Ron Paul wrote the language. It's just a couple of sentences in the appropriations bill. They basically said no federal money or appropriations are allowed to be used for the creation of national patient identifiers. That language has just carried forth through the years. Through education, and a lot of years, CHIME, which is one of my professional associations, has been able to have a conversation with the folks in the House and have it removed from their part of the bill.

For whatever reason, the Senate still refuses to remove that language. Once the language is removed, the next thing would be to get funding appropriated for that.

If you look at the credit agencies, they take data from a lot of different places and do a pretty good job of telling us who you are, but there are still the stories of people who get entries in their credit history that are incorrect.

One of the things Ron Paul was afraid of is there would be a way of spying on people and using information against them. Today in this day and age, I think most everybody puts their second and third thought of the day on Facebook by the time they get up and have their first cup of coffee. So it's ridiculous to think you have any privacy in this country, to be totally honest. Every time my wife wants to sign up for a customer thing, she gives them her contact information. And I'm going, "They're just tracking what you buy, and then they're selling that data. They're monetizing it. You may get something free, but it'll cost you in the end."

Q: To go back to your opening point about people confusing interoperability with data exchange, are the two terms not one and the same?

CC: One of the early tools we used for data exchange was called a CCD or CCDA, consolidated clinical document. It was a lot of data: You have the demographic header, then you have some diagnostic testing, then you have some narratives and stuff, and it can be sent as a PDF.

Well, you're exchanging data, but it's not useful, because a lot of the early CCDs that came out of EMRs could be, depending upon the course of treatment of the patient, 80, 85, 90 pages long. And most of the physicians that I had to send it to — because we were required by the meaningful use regulations to send it — begged me to stop because I was flooding their fax machines or stuffing their mailbox at the hospital. Because they had to go wade through that basket of data to find a piece of information.

I don't know how long your office visits are when you see your physician, but the average is about eight to 10 minutes for a typical visit. How long would it take them to flip through 80 pages of information to find there's really nothing there?

It goes back to "tell me what I need to know, what's changed since the last time I saw this patient." Interoperability is basically taking data, then moving it from one system in a discrete data format into another system, and that receiving system can take it and make useful data on it.

We do lab tests on patients all the time. If I do them at Franciscan, then the data goes into the system in a codified way: It's discrete data, and it's very easy to pull together. If I have similar lab work done at another system, it could be coded differently; it could have a different unit of measure. And so if you send that raw data into my system, it's not going to know what to do with it. It doesn't know where to put it because nothing matches.

One of the things we're required to do is provide quality data back to the federal government. And if you have someone who has heart failure, what is the patient's injection fracture, or how well is the heart working? Well, in some systems, depending upon how that's reported, that will come across as a numeric value. In other systems, it is part of a narrative or a sentence. And then you have to understand what the context is if you're able to pick that value out.

Is some of it useful? Yeah. It meets the requirements. But is it totally useful in taking care of that patient effectively? I don't know.

Q: Have regulatory issues held back interoperability?

CC: It's a mixed bag. There are regulatory requirements — HIPAA as an example. HIPAA is the floor of privacy: rules and security rules that we have to use in managing data. Every state in the union has the ability to create their own privacy and security rules around that data, but it's additive.

I worked at a hospital that was on the border of Indiana and Illinois, and we serviced patients from Kentucky. And so we had three states with laws that we had to deal with about how we handle the data, how they handled privacy.

And patient consent is another challenge. If, say, you're in Pennsylvania, and you go over to New York to get care, New York is an opt-in state and Pennsylvania is an opt-out state. What that means is if you get services in Pennsylvania, unless you tell them not to, they're going to share your data appropriately. But if you're in New York, then they're not going to share your data at all until you tell them to.

That makes it very difficult because if the patient doesn't know that, if they don't know the advantages of being able to share their information with other caregivers, then their data is just going to be stagnant. It's not going anywhere.

The Cares Act has a big provision called information blocking where you cannot not share information. A good double negative there. If you have the data, unless you have a really good reason — and there are a few reasons that you can't or shouldn't share the data — then you have to share the data with who requested it, which is the patient or another provider.

So there are a lot of regulations that have been promulgated around this area. And so I don't think it's for a lack of regulation. It's a lack of clear regulations that don't stomp on each other. In the very beginning, some of the information-blocking rules that were put forth actually conflicted with the HIPAA rules. I think they've reconciled those now. But every time that we get those kinds of rules it adds to the cost of healthcare. We've had to add staff in order to make sure that we were meeting those rules.

It's not as simple as just opening up a pipeline, because we have to be careful about how we share the data. Because healthcare data is probably one of the most expensive things on the dark web that people want to buy. Credit cards are a dime a dozen. Give me a medical record, and I can steal your identity. And I can also do medical identity theft. I can go get healthcare with your insurance.

Q: What is your definition of interoperability?

CC: Interoperability is kind of like a unicorn. Everybody can describe it, but nobody's really seen it.

With the federal funding that was made available through the HITECH Act back in 2009, we spent billions of dollars and didn't get a whole lot of value out of it because interoperability means a lot of different things to a lot of folks.

What kind of problem are we trying to solve? If you're working at a reference lab, your problem is getting that data to the physician who ordered the test. Getting it to them is just part of the solution. You need to get it to them so it drops into their workflows and their EMR and they look at that result and it can be compared to other similar results. If that patient had, say, a hemoglobin A1C test, if it comes out to 7.8 [grams per deciliter] that's abnormal, but is it better than the last one? Was it 8.2? Is it going down? Is it going up? How's it trending? Looking at that single result without the patient's history may or may not be clinically relevant to that physician and what decisions they have to make.

It's not an easy thing to make a determination of what good interoperability is, and I wish I could give you a definitive answer. When I was at the Indiana Health Information Exchange, one of the things that developers did was work with IU Health and ER physicians on FHIR [Fast Healthcare Interoperability Resources] protocols. The Indiana Health Information Exchange is a curator of what we call the Indiana Network for Patient Care, which is basically a massive clinical repository of all the data that comes into the exchange — it gets normalized and it gets stored, and it's from every hospital in Indiana.

And so we tried to come up with a way of telling the physician what they needed to know. So when a patient comes into the hospital with a chief complaint of chest pain, what is it the physician needs to know about that patient, whether it's five or six things that they'd always want to know?

And so what the developers did inside of Cerner, they wrote an FHIR app that would, when a patient came in with a chief complaint of chest pain, search the INPC for those five pieces of information and bring it back and put links on the page for the physician. So all they had to do was click on it and there it was. They didn't have to go hunt for it. Well, one evening, one of the ER physicians who was working with us had a patient come in with chest pain, and he noticed that all five pieces of information were there, but he also noticed that they all came from IU Health. So it actually was in his Cerner system.

So he decided he would do a little test to see if he clicked on the links, how quickly he could get to the data if he went in Cerner and looked it up. He got to the five pieces of data inside the FHIR app in 12 seconds. It took him four minutes when he looked it up himself. That was a big plus, and so they now have 11 different diagnoses that they're working with.

So that's a form, in my brain, of interoperability, because it provided the information when he needed it.

Copyright © 2024 Becker's Healthcare. All Rights Reserved. Privacy Policy. Cookie Policy. Linking and Reprinting Policy.

 

Featured Whitepapers

Featured Webinars