Episode 37: Standardizing Healthcare: Carol Macumber Talks HL7

Join Charlie Harp and Carol Macumber, EVP of Client Services at Clinical Architecture and Chair-Elect of the HL7 Board of Directors, as they unpack Carol's informatics work and her participation in HL7 through the years. From Carol's early days in informatics to leading initiatives like Gender Harmony, this episode takes a deep dive into the organization’s efforts to standardize healthcare data exchange.

They break down the evolution of FHIR, the hurdles of achieving true semantic interoperability, and how AI is shaking up healthcare standards.

Read more about her chair-elect announcement here.

Contact Clinical Architecture

• Tweet us at @ClinicalArch 
• Follow us on LinkedIn and Facebook
• Email us at informonster@clinicalarchitecture.com

Thanks for listening! 

Transcript

Charlie Harp (00:10):

I am Charlie Harp, and this is the Informonster Podcast. And today on the Informonster Podcast, I have Clinica Architecture's own Carol Macumber. Good morning, Carol.

Carol Macumber, MS, PMP, FAMIA (00:20):

Good morning, Charlie.

Charlie Harp (00:21):

And today on the Informonster Podcast, we are going to talk about something near and dear to Carol's heart, and that is HL7. Now I'm going to ask you some questions that people probably don't normally ask you about HL7. This is kind of like Gotcha, gotcha. Journalism. The first question that I'm sure you've never been asked before about HL7, this is how exactly do you spell that?

Carol Macumber, MS, PMP, FAMIA (00:42):

You want me to spell it out?

Charlie Harp (00:43):

No, I'm just kidding.

Carol Macumber, MS, PMP, FAMIA (00:44):

Because full descriptions are important.

Charlie Harp (00:46):

So I think one of the things that would be great for the listeners is for you to give kind of your history and a little bit about your history and involvement with HL7.

Carol Macumber, MS, PMP, FAMIA (00:58):

Sure. Straight out of graduate school, I started with a company called Apelon. They were focused on terminology, terminology, standards, and deployment via their open source tooling. I had the opportunity to attend my very first HL7 working group meeting, I think it was 2002, long, long ago apparently. And it was extremely eyeopening, I guess I should say, as a young informatics person just starting out, seeing how the sausage of standards was made and having some of, I think some at that time, some of the pioneers and informatics on the boots running work groups. I think Stan Huff, who I'm sure lots of our listeners know, he's actually been on our podcast, I think was chair at that time. And many of the folks who I consider mentors were working on the initial standards that were being developed from there. I mean, I spent many years kind of just attending and learning and really soaking things up so that I could apply some of the standards in the work that I was doing at Apelon.

(02:07):

And it eventually kind of evolved into me not ducking fast enough to be honest, and having somebody nominate me to become a terminology infrastructure or what was vocabulary work group then to become a co-chair. And they did me a favor, honestly, I had really kind of was reluctant about taking on a role like that, still feeling like quite a new kid and wet behind the ears as far as my understanding and my ability to contribute to the standards. But it turned into something that I think will be one of my fondest memories and or achievements depending on how things go here of my contribution. So that was maybe a decade ago. Since then, I've also taken on roles in the HL7 terminology authority, which was kind of rolled up into what is now the terminology services management group.

Charlie Harp (03:04):

-And you're about to take on a new role at HL7?

Carol Macumber, MS, PMP, FAMIA (03:07):

I am, yes. So as of January, I'll be the chair elect for the HL7 Board of Directors, and we'll spend a year hopefully learning as much as I can about the board and their role in guiding HL7 as an organization and engaging the community to continue to evolve standards. And then I'll take the helm a year after that for two years and then get to be past chair and

(03:39):

help mold and mentor hopefully the incoming chair thereafter. So four years in total as a member of the HL7 board.

Charlie Harp (03:49):

So I mean, we will get into the weeds a little bit on HL7 and what HL7 does, but when you think about HL7 as an entity, what would you say you think its strategic objective is or its mission statement? What is the mission of HL7?

Carol Macumber, MS, PMP, FAMIA (04:05):

I think, I mean, for me, again, I guess slight disclaimer and that I'm not speaking for HL7 as an organization, but really kind of talking about it from my perspective and its role in delivering standards to help improve healthcare and healthcare outcomes via the standardized exchange of healthcare information. They provide the standards and the framework by which they can do that. You can do that in a hopefully transparent and intuitive way.

Charlie Harp (04:35):

And what does Health Level Seven, why is it called that?

Carol Macumber, MS, PMP, FAMIA (04:41):

Oh, you're going to really go

Charlie Harp (04:47):

Well, I remember back in the day when I was a young man, I developed laboratory interfaces for Smith Cline Beach and Clinical Labs and Smith Cline Beach and Clinical Labs used ASTM standard interfaces, which were OBR, OBX and those types of messages. And so when I left and I went to do other things, and I ultimately came back and I was told that we're using this thing called HL7, and I saw, well, when I used it, it was ASTM. And once again, I'm talking about the 2.X version of HL7, not the CCDA and FHIR, but I thought at the time, I just don't recall. I remember looking it up at the time, but I don't recall why it's health level seven.

Carol Macumber, MS, PMP, FAMIA (05:38):

So layer seven,

Charlie Harp (05:40):

Layer seven,

Carol Macumber, MS, PMP, FAMIA (05:40):

Layer seven is the open systems interconnection model layer for applications. So to achieve that, you need to be able to exchange at an application level.

Charlie Harp (05:50):

Okay. Because one of those things where we throw on HL7 all the time, and I was just thinking as we were sitting here talking, why did they choose the name HL7? So now we know know,

Carol Macumber, MS, PMP, FAMIA (06:01):

At least from my recollection,

Charlie Harp (06:03):

Get to level seven.

Carol Macumber, MS, PMP, FAMIA (06:04):

It's get us to level seven.

Charlie Harp (06:07):

So one of the things that when you and I talk about HL7, it seems like it does a lot of different things. So I think a lot of people may think of HL7, they think about 2.X and the classic messaging format stuff around 2.X, but HL7 is a lot more than that. So if you were to throw out the, let's call 'em the big categories of what HL7 does, how would you categorize those things and just as an exemplar category, the 2.X classic messaging format, rules and guidelines?

Carol Macumber, MS, PMP, FAMIA (06:43):

Yeah, I mean, I think there are some fairly large buckets, but at the heart of it is really, it's a community. It's a community of individuals and organizations and countries that are working together to create standards for healthcare. I mean, I think at the root of it, that's how I would kind of characterize it. And they produced various things which include base standards like v2, X and v3, FHIR, CCDA, but also provide education and training that's critical to the future of healthcare and standards development in terms of educating folks about what the current standards are and engaging those folks to help develop standards of the future. So they provide education and training. They also provide the framework to develop implementations, guides to actually utilize the standards. So there's tooling all around that as far as increasing the ability for people to uptake standards and actually implement them. So it's not just the development of the core standards, but also supporting the customization of those things for jurisdictional-based use cases. I guess those are my buckets.

Charlie Harp (07:57):

So as you know, but maybe our listeners don't know, I've been working on this PIQI framework and we've kind of made the decision that we're going to turn it into an open standard for scoring patient information. And so I guess my question is, we've talked about taking that and taking it to HL7 as a cross paradigm.

Carol Macumber, MS, PMP, FAMIA (08:21):

Yeah, it's kind of like how a schoolhouse thing, how Bill becomes a law.

Charlie Harp (08:27):

I'm just a bill. Yes, I'm only a bill.

Carol Macumber, MS, PMP, FAMIA (08:29):

How bill becomes a law.

(08:31):

It's like how an idea of something like this, and it is at least I think a really good one becomes a standard and there's a path for that. I think a lot of times people may shy away from doing what I think we've agreed to do because it can be a very long laborious process. However, there are intermittent steps that get you there and bolster a community of people to help get you there. We shouldn't be developing standards in silos, shouldn't be putting it on a single individual or single entity to try to do that. And HL7 has a path that you can start to engage the community either first being producing what we refer to as an informative white paper to say, Hey, here's what the problem is I'm trying to solve, and here's how I think I'm going to go solve that problem utilizing existing standards or establishing a model for a new one.

(09:27):

And so I think our first step will likely be that produce an informative white paper about the picky framework, and then from there you're able to create a standard for trial use and get it out into the larger community, go through the HL7 and Nancy certified process for balloting, get community feedback. And through that feedback, people engage and they engage in sometimes disagreement. But that's the beauty of at least I think the beauty of the process where you can get a diverse set of feedback to ultimately improve the standard that you're proposing, which is the goal. I mean, I think ballot resolution for those of us who have gone through it, I mean it is a slog. It can be a slog, but what comes out of it, I think is a standard that's better for the greater good.

Charlie Harp (10:20):

So when you think back on your history with HL7, is there anything, I know you've done a lot over the years with HL7, but is there anything you look back on or let's say things that you look back on and say, I'm really proud of that

Carol Macumber, MS, PMP, FAMIA (10:39):

Thing? Yeah, I mean, off the top of my head, two things come to mind. And for those people who have been on this journey with me, it won't be a surprise that the gender harmony work is at the forefront of probably what I'll always remember and that drive to help better represent the community and solve a problem that it was like a can that got kicked down the road a lot to disambiguate sex and gender and be a little bit more specific about how we should be representing that in standards. And that three year, four year long journey was absolutely worth it from my perspective in terms of what we've been able to accomplish, see it being mentioned in things like USCDI and various others

Charlie Harp (11:26):

And talked about on the Informonster podcast. I forget which episode but yeah.

Carol Macumber, MS, PMP, FAMIA (11:28):

And talked about on the Informonster podcast

(11:31):

Yeah, it was very, please refer to that previous episode and give it a listen. But yeah, so certainly the gender harmony work to be able to see that come to fruition, have aspects of that be taken up in other countries. I mean, it's something that I don't think in 2002 I would ever have said I would achieve. In many ways, I still feel like a new kid on the block because there are people who are so passionate about HL7 that they have been doing this for 30, 40 years and will continue, I think sometimes into their retirement and still contributing, which is amazing to me that people are that passionate about doing something that honestly they're doing as volunteers.

(12:11):

I mean, there are absolutely instances where there's corporate support here where it's part of what we feel is important and therefore we're putting our assets into that continued development, but it's mainly volunteer based. So that being the first, the second, which I think is something that on the board I hopefully will help continue to focus on, is mentoring the next set of HL7 contributors and members and co-chairs. And that our ability to embrace new people into the community and help educate them, not berate them for not understanding what we did 10 years ago or 17 meetings ago, I think is critical to the future success of HL7.

Charlie Harp (12:57):

I think that's one of the things that could be challenging in a community with a long history is

(13:03):

Not to become insular where you become so clique-ish and closed that new people can't feel like they belong and contribute. And my experience on the periphery of HL7 is it's a very welcoming group. It's a very open-minded group, and you're right when it comes to this kind of volunteer work, it's not a job for the people that are involved in it. It's really more, it's something typically people are very passionate about and they build these lifelong relationships in this community and they have a sense of purpose about it. And it is really cool when you're a part of something like that. So when you think about, so we had two x, which is still out there. We have CCDAs, which are still out there. We have FHIR, which is a lot of initiatives wanting to push FHIR for a lot of good reasons. It's a more robust standard than we've had historically, and it came around during a time where we had the technology and the bandwidth to support something like FHIR. When you think about the future of HL7, what are some of the things you think are going to be impactful or significant?

Carol Macumber, MS, PMP, FAMIA (14:18):

I mean, I think everything else. I mean, we have to be able to evolve with technology and advancements. And obviously if I don't say the word AI in this podcast, it may get screened by people because it's everywhere you are, everywhere you go, every conference we attend, initial conversation we have with folks was what is your approach to harnessing and utilizing artificial intelligence and large language models and so on.

Charlie Harp (14:46):

But how do you see AI playing a role in some of these standards?

Carol Macumber, MS, PMP, FAMIA (14:49):

I mean, I don't necessarily think that AI will play a role directly in the standards development, but certainly the way that we look at things and how the standards would be used should take into consideration any large technological advance that includes AI to say, Hey, we know that systems are going to be using it. Is there anywhere amidst the standards that we're creating that we need to take that into consideration in terms of what is going to be utilized from within the standards or to help ensure that AI doesn't do something we don't want it to do? I mean, I don't really know how we would do that, and I'm certainly not an expert on AI, but I think there is definitely a recognition within the community since we're all in our corporate lives being asked those questions about how it can be used.

Charlie Harp (15:44):

Well, what's interesting is I talked to people in the industry and some folks come up to me knowing that I exist in the terminology world

(15:56):

And the structured data world and say things like, oh, Charlie, we're not going to be using terminology. It's going to, five years from now, I'll be having a coffee and say reminiscing about how back in the day we used terminologies. I remind them that we still use fax machines in 2024, but I also think that I don't see us getting away from structured data. I don't see us getting away from terminologies. I think they anchor us into reality. I mean, if we do get away from terminologies, and one could argue that we probably wouldn't need standards for transporting terminologies from point A to point B. And I think that when someone takes a step back and looks at the broader ecosystem, it's a little naive to think that the entire healthcare industry will shift in even a 10 year period to something that is fundamentally different than what we do today.

Carol Macumber, MS, PMP, FAMIA (16:49):

And at least I understand the terminology community, the view on that, I mean is really terminology and the terminology models that are in place in our standard terminologies like SNOMED in particular could be used to help train AI to do a better job of parsing text. And you certainly can't replace it, but rather utilize it to make it better.

Charlie Harp (17:16):

I think that AI is a technology that fundamentally, at least the current generation of things, like large language models, is they can be used to augment human activity or accelerate human activity. But the thing I keep reminding people that are like, oh my God, Charlie AI, it's so great, is it is a rear facing bias based technology, and the only way to unbias it is to re-bias it. And so we're always, the metaphor that I used when I spoke about it at AMIA is it's like a Ouija board where history's hands are on the planchette, moving it across the board,

(17:56):

And we just have to be aware that even though it's a clever simulation of an intelligence of a human being, it isn't. And if you look at some of the responses you get or look at some of the videos that it creates or whatever, it quickly becomes apparent that that is not something that came from a human being. But I am biased. I'm a technology zealot. I love new technology. I've played around with Chat GPT a lot, but I still think that it is a technology that maybe isn't appropriate for use in some of the ways people are using it. Oh my God, how did this become a podcast about AI?

Carol Macumber, MS, PMP, FAMIA (18:36):

Because I said the word, and that's what happens.

Charlie Harp (18:39):

I know ai,

Carol Macumber, MS, PMP, FAMIA (18:40):

It has already trained us.

Charlie Harp (18:41):

I know. So back to HL7, other than AI, which I would argue should not impact HL7 other than if HL7 starts to tell us when things came from AI versus came from a human, maybe that's something we add to HL7.

Carol Macumber, MS, PMP, FAMIA (18:58):

Sure. I mean, obviously there's been an increased push towards security and privacy, and I mean that's a great use case to say, okay, well, are we able to throughout the ecosystem of utilization of HL7 standards from v2 and CCDA and FHIR all mixed into the same healthcare ecosystem, are we able to say what came from where? Which data sources did it originate from? What amount of augmentation has occurred and where it's certainly that's something that standards can help determine, make transparent in the information that you're receiving.

Charlie Harp (19:37):

You want to get super nerdy?

Carol Macumber, MS, PMP, FAMIA (19:39):

I don't know, maybe. I have a cup of coffee in my hand.

Charlie Harp (19:41):

You've had some coffee. Let's get super nerdy.

Carol Macumber, MS, PMP, FAMIA (19:43):

One hand, although this is hotel coffee, so we'll see.

Charlie Harp (19:45):

I'm going to put forth a statement and then we can debate this statement or agree. When I look at HL7 historically, and I look at the standards, because remember I was a developer. I built lab interfaces back in the day. That's what I did. I really feel like the standards we have in healthcare, and the reason why some of them are pretty flexible is because they evolve from being a representational technology where I'm electronically sending you data so that I don't have to call an auto dial printer or fax you a report because a lot of the stuff from HL7 came out of these transactions like a lab report. So I think when you look back historically, we had these things that were representational, which they could be a little loosey goosey, and even CCDAs are really representational. They are something that allow you to construct a human readable piece of information.

(20:48):

One of the things that I think is interesting about FHIR is I think that, and one of the reasons why I like that FHIR is a little more prescriptive than the previous standards is I think that people talk about how we've exchanged data for the last 20 years, but I really think we exchanged the representation of data for the last 20 years. So semantic normalization wasn't important because I'm just printing it out and you can read it with your human eyeballs. I think with FHIR and with where we are today, there's really a need for us to not have a representational sharing of data, but a computable sharing of data. And I think that's one of the reasons why we're starting to see data quality become more of an issue, semantic, normalization, interoperability, because we're not just sharing something you can print for a human to read. We really need something that software can understand without having a human to spoonfeed it. So that's my statement.

Carol Macumber, MS, PMP, FAMIA (21:45):

What do you think? I don't think you'll be surprised that. Well, maybe you would be that I agree. I mean, I don't think there's any,

Charlie Harp (21:49):

You're a reasonable intelligent human being.

(21:52):

Of course you would agree with me.

Carol Macumber, MS, PMP, FAMIA (21:53):

Yes. I mean, so v2 flexibility was its strength, but also its achilles, right? I mean increased implementation dramatically. It's probably the most utilized. I don't have any statistics to back that up. People out there,

Charlie Harp (22:10):

I heard that I would not be surprised

Carol Macumber, MS, PMP, FAMIA (22:12):

And it allowed people to send information in a structured way, but along with CCDA where the structure was even more prescribed, the content was not as much, right? I mean, you could place things. There's a lot of flexibility there as far as the content and people like to save. You've seen one. You've seen one because based upon the institution from which it's being generated could technically be conformant, but the content itself is a varying quality. And this push towards saying, Hey, the data that's actually in that envelope has to be of higher quality, and we have to have a way of assessing that and having a feedback loop in place to say, Hey, these are areas that are really important that will ensure that there's semantic interoperability. I can't just receive the thing and understand the size and shape of the envelope, but understand the contents that are inside. And having people be able to in systems, be able to take that feedback in a standardized way and make improvements. I mean, I don't see any other way to really achieve what we've always kind of been working towards in terms of semantics and semantic interoperability and equivalence. 

Charlie Harp (23:32):

I think that when you look at HL7 v2, it was a very flexible thing. I think that one of the challenge we have with CCDAs is that CCDAs sometimes I feel like when people develop a standard, because a standard is kind of a guideline, there's nobody that says that your v2 is not compliant, so therefore you're not getting paid. I mean, there might be, but I don't run into them. It's not unlike X12, 837 where you have these formats where if you don't send the data in a format that we recognize, you are not going to get reimbursed for this. There's no real impact on that. And so we end up burying the cost as the receiver to make sure that we can have interfaces that can handle all these different variations. And it's the same thing with CCDA. Ideally we'll have that less so with FHIR because FHIR is more prescriptive. And theoretically you could see a time where if we say that the exchange of data is not just referential, it's critical and to be a good part of the ecosystem for public health and for everything else, you have to send compliant FHIR.

(24:50):

So it's not an option. There's no flexibility. And I'm one of those people where people tend to flexibility, it gives them freedom. But I'm of the opinion, if we really care about exchanging data, we almost have to be draconian about formats that we have to slow the role so we're not creating a new version of a format every two years because we're then forcing the entire industry if we're saying you have to be compliant with it to rewire everything to be compliant. So I think it does a couple of things. It makes the standard more draconian. It makes it so you almost have a minimal viable set of information that you have to exchange, which is kind of where things like USCDI come in, and then you have to basically tell the industry, we're not going to change this every two years. Right. Does that make sense?

Carol Macumber, MS, PMP, FAMIA (25:39):

It makes sense. I mean, I've been waiting for this moment in the five years that I've been working here to turn the podcast around on you and ask you a question, Charlie Harp.

Charlie Harp (25:47):

Oh my God, I'm ready.

Carol Macumber, MS, PMP, FAMIA (25:49):

One of the things that people love about FHIR is that it's the concept of extensions is baked into the approach. With too many extensions, what does that do to our draconian standard? 

Charlie Harp (26:08):

I mean, it's one of those things where it goes back to the siloed nature of healthcare. I'm a firm believer that healthcare evolves from the edges, which is why I don't think we can have one terminology to rule us all or one model to rule us all. I think that we don't want to stifle innovation by saying, this is the way we do it. So things like extensions and those things are good. I think that if what we're saying is this is a way that we exchange data, we can't expect that if we create an extension, unless it's for a particular business purpose or a particular industry purpose, if everybody creates an extension, then we don't have a standard. A standard is something that we all agree on. So I think that if nothing else, an extension is an innovative mechanism to evolve the standard in a particular direction.

(26:58):

But we shouldn't have a thousand extensions happening at the same time. I mean, I think you can, but I think what you're really doing then is you're going back to this world where you're like, well, maybe we don't have a standard. Maybe we just all kind of do our own thing and we throw something at you and it's on you to figure out how to unravel it, which puts us right back to where we are today. But that's why I'm not the interoperability czar because I'd lock everything down and say, Hey, put that down. Stop messing with that. But I think it's one of those things where it's a challenge. I mean, for example, I was looking at USCDI v4 because this is the other thing that happens sometimes with standards. It happens with standards, it happens with quality measures where somebody sits down in a room and designs this thing.

(27:45):

That is an ideal thing, and then they say, this is what we're going to do. Here's an ideal thing. And not to get in trouble with anybody. But when I think about USCDI v4, I think it looks good. I think there are things that aren't in USCDI v3 that should be there. Every lab result should have a date and a time that's not there. It's kind of implied, but it should be there. When I think about USCDI v4, one of the things that added to the medication is adherence. And they say there's going to be a SNOMED code that says whether somebody is adhering to their, well, here's my question. Who knows that? That's going to send a message, whether it's FHIR HL7, CCDA, who truly knows whether the patient is complying with their medication regimen?

Carol Macumber, MS, PMP, FAMIA (28:33):

It's a good question. I mean, I think there is some precedence for that in the quality measurement space. They retroactively look at adherence for quality.

Charlie Harp (28:45):

Now, that's what I mean. 

Carol Macumber, MS, PMP, FAMIA (28:47):

That's what live interaction…

Charlie Harp (28:47):

I wish we had this information. If we had this information, we could answer questions.

(28:52):

But it's kind like some of the things you see in social determinants of health. You can put a question out there, but if the person doesn't trust the person that's asking the question or they think you're asking a question about food security because you want to take their kids away, there's so much uncertainty in that that we have to solve some of these problems and we need to figure out how we're answering that question. And with the medication adherence, unless we're asking the patient, unless the patient's sending us data saying, or they have a device that measures every time they take the pill, which could happen in the future, who knows? My Apple Watch might be watching me right now and saying…

Charlie Harp (29:28):

Why didn't you take your unload?

Carol Macumber, MS, PMP, FAMIA (29:30):

Know you're getting that refill, but you're just stockpiling it.

Charlie Harp (29:32):

That's right. So those are the kind of things where I think that, I guess I'm trying to think of what my actual point is with this, but it's kind of one of those things where I think with standards, we need to make sure we know what's realistic.

(29:47):

I think we need to avoid, back when I worked at First Data Bank, I noticed this thing about subject matter experts. I'm not qualified really to do anything, not even this podcast, but I noticed that subject matter experts are really proud of their knowledge. And so when you talk about a clinical scenario or you're designing a system to support something, they really kind of gravitate towards the edge case because it demonstrates their expertise. And well, that's true 80% of the time, but 20% of the time this esoteric thing happens. And I think sometimes that bias to want to make sure that people realize that you are a subject matter expert tends to drive design and things into this edge case. And you see some of that in the standards, some of the things in FHIR have a lot of complexity in their design, and that complexity is going to be engaged 3% of the time. And I could certainly accept an argument where somebody says, Hey, if it's you and you're in that 3%, don't you want to be represented? That's true. I do, but I also feel that perfection is the enemy of good. And so how do we make sure that we know what the minimal viable or the least common denominator or the most frequently used kind of pattern is, and then tell people this is what should go in the most reasonable pattern, but there could be these other situations that might be outside of that.

Carol Macumber, MS, PMP, FAMIA (31:17):

Yeah, I mean, as far as FHIR is concerned, I mean that's why there's, at least to me, why there is a maturity model in place so that things don't become part of that core, a normative standard until it has been proven to be applicable to the larger masses. So there's a lot in FHIR, but only a portion of it is actually normative. And so the things that are in there is trial use are for trial use and for people to help refine. And when they get to the point where we have evidence and meet the criteria within the FHIR maturity model to promote them, then that's how they get promoted.

Charlie Harp (32:02):

So if you were, the interoperability zar, all hail, Carol Macumber, Interoperability zar.

Carol Macumber, MS, PMP, FAMIA (32:09):

Are, isn't that wonderful. 

Charlie Harp (32:10):

You could enforce compliance with FHIR going forward, where would that be do you think?

Carol Macumber, MS, PMP, FAMIA (32:19):

Enforce,

Charlie Harp (32:19):

Yeah. Like you said, Hey, you don't get your X, Y, and Z payments if you don't test against this standard. I'm more of a data quality result that formats out. But let's say this is a format situation and let's say it's in the future and we decide as a nation that for tef FCA for exchange, we really are trying to drive everybody to not just send us data in FHIR, but to comply and be able to prove that you are compliant with FHIR. I know this is kind of a weird question. Maybe it's a dumb question too, but there is, when you look at FHIR today as a standard, there is kind of the core FHIR and then there are extensions, and then there is where FHIR is going in the future. And so do you see a time when we could say, if we're going to try to enforce compliance with it as a format for exchange, that you could say, this is what you must comply with and this is what's coming. And these are things that are optional because they're extensions? Yeah,

Carol Macumber, MS, PMP, FAMIA (33:30):

I mean, I think it's not realistic to say there's a single version of that standard that everyone could easily comply to across all use cases. Now that being said, you set a floor and the magic, I think is setting that floor in a way that it allows for flexibility still and further constraint. So in the US at least, I mean, we're trying to do that by establishing US core, which is the US version of the underlying FHIR standard, and establishing US core that provides a floor for people to then further establish implementations of that. So for example, for public health reporting, there are various implementation guides that say, for this purpose, for this organization, this is how you must format your FHIR resources for reporting to us. And that provides a path to ensure that the information that's pertinent for that use case is included. Not everything, it's not applicable. So I think there is a path for people to say, you must comply with this implementation guide purpose..

Charlie Harp (34:47):

In a particular use case. 

Carol Macumber, MS, PMP, FAMIA (34:48):

Use case in a particular use case, right? Of course, at a national level, things like HDI one and USCDI are helping drive towards the utilization of US core as our base standard. But I think saying that isn't enough, but rather saying these implementation guides have to be implemented for specific use cases.

Charlie Harp (35:11):

I think that part of the reason why we struggle with interoperability, because we did the data quality survey, interoperability remains a big issue for a lot of people. We're trying to roll out TEFCA across the country, and most people I talk to are, they might connect to it, but they're not going to just ingest data into their system. They don't trust the data. I think that part of the challenge have is the people that are sending data out don't really, and I don't mean it this way, so if you're out there, don't be mad at me. Don't hate the play, hate the game. I think that the data we send out is less important to us than the data we receive.

Carol Macumber, MS, PMP, FAMIA (35:57):

Sure.

Charlie Harp (35:58):

I think that, and this is why we develop PIQI, is because we don't really have a way to even measure the data that we're sending out. We can kind of validate it against the format like FHIR or HL7 or CCDA. There are validators that say, oh yeah, the structure looks right, but the actual quality of the data we don't measure. And so I think that we probably do need to have something that forces people to care about what they send out. And unfortunately, the only way to do that is money at the end of the day.

Carol Macumber, MS, PMP, FAMIA (36:32):

Money or less money.

Charlie Harp (36:34):

Yeah, exactly. And when I was in the finance world for a couple of years before I moved into healthcare, and this goes back to 1989, so it was a while ago, and I remember when I moved into healthcare, I was really shocked by when we were doing decision support and things like that, they were like, oh, it's okay, because there's always this assumption. There's a human intermediary that's doing certain things. And so the software is important, but it's okay if it's not great. And that's changed obviously over the years. But the other thing that really surprised me was you'd be talking to people about things and you'd say, well, this is for patient safety. And it became quickly apparent, even though they didn't say it that well, yeah, we care about the patient, we want to take care of the patient. But money is what drives people to make decisions. And oftentimes money can and different schemes because no good deed goes unpunished. Sometimes We put incentives out there and it has a weird orthogonal effect on behavior that we, it's not really what we intend.

Carol Macumber, MS, PMP, FAMIA (37:47):

What you choose to measure impacts people's behavior

Charlie Harp (37:51):

And what you pay for

Carol Macumber, MS, PMP, FAMIA (37:52):

And what you pay for. And I don't recall the actual numbers, but ahead of our data quality theater at HIMSS, I did some research into that exact kind of topic in terms of what percentage of people actually trust information that they've received and what percentage of people actually use it. And clearly the first number is rather small to begin with, and the second number is even smaller. So they may receive it, put it someplace, but they don't actually use it. I've heard it be referred to as an ocean of data, but a desert of insight because the more data you get and we're producing it at a massive rate, I think the number of the doubling, the doubling of healthcare

Charlie Harp (38:40):

Every 73 days,

Carol Macumber, MS, PMP, FAMIA (38:41):

73 days, we're doubling our healthcare information. And that's overwhelming.

Charlie Harp (38:47):

Well, and a lot of it's the same stuff over and over and over again.

Carol Macumber, MS, PMP, FAMIA (38:50):

Inefficiencies in our healthcare system lead to

Charlie Harp (38:53):

One of my concerns, frankly about things, technologies like AI is one of the things that computers do is just do what we do faster. So maybe it can help us sift through some of this stuff, but it's certainly going to help us produce it. And so I think that's going to be interesting. So is there anything else you were like, Charlie, I'm not going to be good on a podcast, but you're fantastic.

Carol Macumber, MS, PMP, FAMIA (39:20):

Well, I talk a lot.

Charlie Harp (39:23):

Well, yeah, exactly. If you just sat there and didn't say anything, this would be a really boring podcast. It just be me talking. No one wants that.

Carol Macumber, MS, PMP, FAMIA (39:30):

No, I think as a last parting thing, I guess it would be remiss of me not to pay homage to everybody. That's kind of helped me get to where I'm at this point and say that I'm really excited and the opportunity to continue to contribute to HL7 and its future, the future is bright. I want to thank,

Charlie Harp (39:52):

And I think that there are people that get frustrated with HL7 and they get frustrated with standards and everybody says, well,

Carol Macumber, MS, PMP, FAMIA (39:59):

One of 'em might be in this room, two of 'em might be in this room, two of 'em might be in this room.

Charlie Harp (40:04):

I think that think standards play an important role. And I think that one of the great byproducts of organizations like HL7 and a, and there are other organizations out there, is not only does it provide things that we all get to benefit from and leverage, but it also is a collection of really smart, passionate people who are thinking about something. And I think you can underestimate the value of that. But that thought process, the designs and the consideration that drives the industry forward, and I would argue that people go to HL7, they have these conversations about standards and things, and they take it back to their organization and it may have nothing to do with FHIR, but they might say, you know what? We have this conversation about gender identity, and I really think we should change the way we do things in our system so that we're separating gender identity and birth sex.

(41:08):

And it's not necessarily to even comply with the standard just by being part of the community. It elevates the way they think about it, the way they understand things, and it drives innovation across the entire ecosystem. So I think it's one of those things where it's like, I'll use the word virus, but I don't mean it in a bad way. It's like an idea. It's a breeding ground for idea viruses that drive the incremental evolution of our industry at the edges. And I don't think that these types of organizations necessarily get the credit for that.

Carol Macumber, MS, PMP, FAMIA (41:44):

And it's one of the things that HL7 has been focusing on in the previous strategic plan. There was a great deal amount of focus on establishing an environment in which people could see how it works. So for the FHIR accelerators, the foundry that has Ference implementations, FHIR has been touted as really being able to reduce the learning curve and the barrier to new development. But even developers appreciate having someplace to go look at how does this already work in certain scenarios. And being able to see both client server end of how the standard is interacting for a given FHIR accelerator for a given purpose. And there was a great push towards that, and you see it in the foundry, and I can imagine that there will be much of that in the future.

Charlie Harp (42:40):

Alright, well Carol, thank you so much for being on the Informonster podcast with me today. My appreciation for HL7 has evolved over the years to the point where I'm really proud and delighted to have somebody who's on the clinical architecture team in such an influential role. And I appreciate all the time and energy you've dedicated to help healthcare be better. And it's obviously something that's very important to you. So thank you for being on today.

Carol Macumber, MS, PMP, FAMIA (43:07):

Thank you.

Charlie Harp (43:08):

And for all of you out there, thank you very much. This is Charlie Harp and this has been, the Informonster podcast. Thanks.