The Sequoia Project: Advancing Data Usability in Healthcare

Stephanie Broderick (00:06):
Thanks everybody for coming to the Clinical Architecture Data Quality Theater. My name is Stephanie Broderick. I’m EVP of Strategic Initiatives for Clinical Architecture, and I am super excited about this particular session. Clinical Architecture has been involved in The Sequoia Project for about the last year and a half, and there is some really important initiatives that the group has undertaken around data usability and Clinical Architecture is very involved and fortunate to have the chairs of the workgroup that are leading this initiative. And so I’m going to go ahead and have them introduce themselves.

Didi Davis (00:47):
Hello everyone. And again, welcome. My name’s Didi Davis. I’m the Vice President of Informatics Conformance and Interoperability, meaning I wear lots of hats for The Sequoia Project and I’m very grateful for the opportunity to share a little bit more about data quality. But what The Sequoia Project does is really focuses on interoperability issues that are plaguing us here in the US and trying to advance ways to help remediate those for the public good.

Stephanie Broderick (01:14):
Wonderful. Dr. Gregg.

Dr. William Gregg, MD, MS, MPH (01:15):
Great. Thank you. Bill Gregg. I’m Vice President of Interoperability at HCA Healthcare and also co-chair of our Data Usability Workgroup at The Sequoia Project. And so from our perspective as HCA as being one of the largest healthcare providers is the critical nature of data usability, both for the efficiency of how things operate, but also the experience of our care providers and our patients. And so this is such an important issue to us and we see it at the tip of the spear of exactly where these things are happening, and so we’re happy to support it.

Stephanie Broderick (01:54):
Fantastic. Charlie?

Charlie Harp (01:56):
I’m Charlie Harp. I’m the CEO of Clinical Architecture. I am not a chair of usability, but I am one of those slats in between the chair legs and a big supporter of The Sequoia Project and what The Sequoia Project is trying to achieve.

Stephanie Broderick (02:14):
All right, so Didi, you can tell us a little bit more.

Didi Davis (02:17):
Yeah. So for those who may not know what Sequoia has been doing for data usability, I promise we’re only going to spend a few minutes for a few slides so that everybody’s on the same page before we start our discussions. So basically we published a version one implementation guide December 2022.

(02:38):
So we actually published guidance. We started as a workgroup. This is kind of the history and purpose of our workgroup from the charter. Our board of directors prioritize data quality and usability as something we really need to help make some advancement on, to move the needle. We have lots of data moving. I don’t know if you realize, but there are over 9 billion clinical documents moving across networks today in the United States, and some are really good quality and some are hundreds of pages and not quite as useful. So that’s what we’re really trying to attack. The board prioritized that, and they actually prioritized for us to work on it at least into the next decade. This is not something that’s going to be one and done. We’re going to have to continually make incremental process to actually get this moving forward. So our goal was to publish the implementation guide, and again,

(03:33):
Dr. Gregg’s here. We are missing Dr. Adam Davis, who’s a pediatrician from Sutter Health. So I have two wonderful clinical advisors helping make sure we stay grounded, but basically this was kind of our goal. We don’t want to create new standards, we want to build upon the standards that exist today, but help elevate those things that can actually make a difference in the real world with the systems that are out there today, the publication of the guide focused on six topic categories. So on the right side of this slide you’ll see those six. There’s things like provenance, where did the data come from so that it could be trusted, making sure you know where it came from, what daytime it came from. Things like being able to reduce the impact of duplicates. When people query for data, a lot of times they get the same information from multiple sources. How do we reduce some of that? How can we do things to help alleviate? Those are just some examples, but the implementation guide, this is kind of like the outline. We have the executive summary and the entry level pieces. There are the six topic categories, and then there’s some appendices that support that. It’s only 42 pages, and if you take and look at just the requirements part of it, it’s probably less than 10. So it’s not a huge lift, but hopefully something that’s easy to look at.

(04:50):
Each chapter focuses on what is the use case, what is the actual standards that we’re going to use and what’s those additional constraints we’re going to add to the standards to help make this real and actually achieve improvement? And we also kind of alluded to some of the future efforts in version one of what we didn’t get to, but kind of put in a parking lot that we knew we needed to address in the future. Now that was version one. We didn’t stop there. We’ve already started working on version two. So in February, one month, we took a whole month off and we started reconvening the group to start working on those parking lot items. In February of ’23, we spent all last year going back through the parking lot items, reprioritizing, rescoping, what we wanted to try to include in version two, and we started those efforts.

(05:41):
This is the workgroup makeup. Today we have almost 400 organizations and almost 500 participants that have actually signed up to be part of the workgroup and part of the roster. And this gives you kind of a look at where they come from, what stakeholder hats are represented across the board. So we tried to have a broad representation. We do this in phases. So we finished phase one, February through June. We started phase two in July of last year. So we’ve spent one year actually putting pen to paper and starting writing information down. For version two, our goal is to put it out for public comment in July, hopefully for about 30 days, and then publish the final version two document by December of this year. This is how if you want to follow along, this workgroup is open to anyone. You don’t have to be a member of Sequoia.

(06:31):
This is the link. These slides have been made available and hopefully you can download them yourself, but you can get all of the meeting logistics, download the calendar invites, listen to any prior meeting recordings or catch up as you would wish. Now, we didn’t want version one to sit on a shelf and collect dust. We wanted to actually have people implement it. So we partnered with the American Health and Information Management Association, AHIMA. They joined us to try to actually move version two into a deployment and build what are called a community of practice. So what we’ve done is taking the development work. So if you look at this part of the slide, this is the work. We’re developing the guidance and we’re actually now trying to deploy it. And that community of practice, we have monthly round tables. They meet the last Wednesday of every month and we actually are developing technical assistance resources to actually help those implementers.

(07:28):
We just announced last Friday in a press release that we just launched a new testing platform to be able to test so people will know what they don’t know. How does my documentation score when it comes to data usability? And then the in-person convenings, we try to do one in-person summit per year. The last one we had was in September of last year. And the goal is to have people sign up as supporters, implementers or sponsors because we are a nonprofit and we do appreciate any help that we can get from the folks that help us along. We launched the pledge for people to formally sign up. We’ve had a lot of verbal agreements from almost 60 organizations and we’re now actually asking them to pledge so that we can use their logo saying that they are a supporter and as implementers, they’re going to agree to report metrics on how many of their customers have deployed this over time.

(08:21):
So there’s little things like that. It’s called a letter of understanding. It’s not a big legal document, but just enough so everyone understands their part and what role they play. This is what you can learn more about data usability, the workgroup itself and the Taking Root Movement itself. So that’s the overview and we’ll go into the panel discussions.

Stephanie Broderick (08:41):
Fantastic. Didi, before we move on, can you talk a little bit also about the Tiger Lab?

Didi Davis (08:46):
Oh yes, good, good point. So in version one, we found that lab interoperability was one of the hardest problems, that no one’s really made a lot of headway addressing so far. Now we didn’t have enough standards to appoint to that existed yet to do some of this guidance that we needed to help improve this. So in version one, we created an appendix saying, here are the top 200 ish labs we are looking to focus on.

(09:17):
So start looking to see what you’re doing with those. And because last go around doing version one, every time we talked about lab, it seemed like it sucked all the air out of the room. So what we decided to do is create a Tiger team. So a good point. Thank you for bringing that up. We do actually meet as the workgroup the first Thursday of every month. And the lab interoperability Tiger team meets the second Thursday of every month. So it’s a one-month monthly meeting. And we have representation from large labs like LabCorp Quest. We have folks from subject matter experts across the industry from the standards organizations as well as those folks who are actually doing some of the landscape assessments and helping us figure out what should we focus on, what are the use cases that are going to give us the biggest bang for the buck.

Dr. William Gregg, MD, MS, MPH (10:05):
And I think that’s well said. I want to emphasize one of the things that Didi said and that’s about creating that community that’s really key. We’re not a standards organization and not intending on doing that, but there’s a lot of standards out there that aren’t adopted because there’s no teeth to them or anything like that. The ones that get adopted are the ones that are required. So what we do is we point to things that might be relevant to really have the biggest bang for the buck. What can we improve in an 18-month cycle? And the Tiger team is really good about doing that. The whole life cycle of a lab, there are so many places where things can break down. And so how can we bring all those stakeholders together, not stay in our silos, but force everybody to go outside their silos and make meaningful change to the end user and the end consumer of that data. And I think that’s different than what others have done, but we’re complimentary to that. We’re not trying to take anything over but really say, alright, here’s that last mile of improving the usability. And we think that now is the time is right, the standards are more mature, the technology is better, and the industry is ready for this, we believe.

Stephanie Broderick (11:13):
That’s fantastic. So Charlie Clinical Architecture participates in the Data Usability Workgroup. We’re also an implementer of the Taking Roots initiative. We’re also participating in Tiger Lab. From your perspective, why was it important to you and to Clinical Architecture for us to be involved in this?

Charlie Harp (11:31):
I just wanted to hang out with Didi, really. No, I think when you think about Clinical Architecture, we work with people across all the verticals of healthcare. Providers, public health, government, life sciences, payers, and data quality is kind of our thing. And the things that are happening inside The Sequoia Project, basically an orchestrated improvement of data quality, data usability, data sharing. And so it’s kind of one of those things where it’d be crazy for us not to be involved in something like The Sequoia Project. It’s kind like it is, therefore we are.

Stephanie Broderick (12:09):
Great. So this is a group topic, so you guys answer however you want. Contrast data usability versus data quality. I hear this question a lot.

Didi Davis (12:21):
Do you want me to start then you guys chime in? Alright, this is a good question. So they are not separate. Data quality is the foundation for data usability. So data quality is making sure that you have the right data formatted in the right sections with the right codes. It’s using a RxNorm code if it’s a medication or a SNOMED code or a LOINC code if it’s a lab, making sure that those adhere to the specifications and the requirements. Now that’s great and that’s what everyone is doing today because they have certified systems, but they don’t necessarily have that data being presented in the workflow and usable to that clinician or public health official or even consumer as a patient the way it could be in the maximum effort. So if you think about usability, it’s trying to make that data actually usable to that end person who’s taking care of the patient, helping them help that patient have the best journey possible.

Dr. William Gregg, MD, MS, MPH (13:22):
Yeah, I think that that’s well said. One thing I would add is kind of as an example is well, data quality is necessary, but it’s not sufficient for full data usability. As Didi said, it’s a component of it. It’s a very big component of it, but we can have high-quality data, but if a sending system actually doesn’t even send that part of the data, it doesn’t matter, or if they send it, but they put that high-quality data in the wrong section, it won’t be consumed properly or the receiving system doesn’t render it effectively, it doesn’t make any difference. So it’s all of these pieces, it’s like a puzzle and data quality is a huge piece of that puzzle, but then how do we make sure the other things happen so that high-quality data doesn’t just sit on a shelf.

Charlie Harp (14:13):
And I’ll add another layer to the cake. I agree with everything these guys have said, but I also think that one of the things about usability and quality is it’s got to be fit for purpose. And so usability changes depending upon what you’re using it for. And the other thing too is you can have the most beautifully, well-formed conformant data, but if it’s not right, so there’s the quality in terms of measuring what’s in each bucket. But if I’m getting data for Charlie Harp and it’s really Bill Gregg’s data, it might be quality data, but it’s not useful because not in the right place. And the other side of it is not having data is also something that affects data quality. So data quality isn’t about measuring what’s there. It’s also about knowing what’s not there.

Stephanie Broderick (15:05):
Right. So Charlie, there’s a big focus on AI, but as you always say, garbage in is garbage out. So from your perspective, what is the state of data usability and data quality today?

Charlie Harp (15:19):
I mean, I think that when you look at the data that we have today, and we did our survey last year, we’re in the middle of the survey this year. I mean, most people say that their data quality is not good. And I think that part of that is an artifact of our industry and how we have dealt with data and things like longitudinal data. When the industry decided that we were going to have longitudinal data, we didn’t redesign systems to support longitudinal data, we just stopped deleting things. And so what we really have, it’s kind of like a landfill. We have a landfill of historical data from all these episodes that happened from time immemorial, and when somebody says, give me a summary, we take that and we back the truck up and dump it on top of them. And some of that stuff is old and some of it’s wrong and some of it was copied forward when it shouldn’t have been copied forward. And so I think the problem we have today is that we have this tsunami of data. And so even identifying what the quality data is, you really have to deal with the noise-to-signal ratio to be able to do that.

Dr. William Gregg, MD, MS, MPH (16:27):
Yeah, I want to echo that comment that’s so important. If we talk to any of our providers of care in our hospitals and that includes our physicians, our nurses and other stakeholders, they will say exactly that, it’s a signal-to-noise problem. They may not use those words, but that’s what there is, that there’s a lot of noise out there. Now that doesn’t mean it’s not useful data, but for what their purpose is at a moment, that’s not the signal. It’s important. And so that’s why in this data quality and data usability work, it’s not just to make it consumable by the human because I think our industry for too long has put it on the backs of the humans to say, well, they’ll sort through it. That’s not possible anymore. Now that’s where we start to say it’s got to be machine consumable to be able to identify that signal and that’s where the AI comes in. But as Charlie said, it is garbage in, garbage out. It’s not going to work unless we do this. We’ve got to do the hard work, the grunt work that no one wants to do to get to that next world of where we can really make our caregivers and our patients work seamlessly with the data,- aspirationally. But it’s going to take a lot of things to get there.

Didi Davis (17:38):
And one thing that we did do in version one guidance was add a recommendation requirement that these electronic systems that send data not only give that summary of care document, which is supposed to be a snapshot, I’m referring the patient, it’s supposed to be everything, the allergies, meds, that kind of information. But we need encounter level data as well. We don’t want just everything in a hundred-page document where somebody has to filter through 15 different encounters or 30 different encounters. We want to have them start making those encounter documents like discharge summaries, referrals, progress notes, histories and physicals. Those types of documents need to be also available to be sent. And just doing that has made a huge difference. Vendors thinking about, oh yeah, meaningful use kind of on the path. I just put everything in one document, but there are other documents out there they can leverage. And I think that will help with the AI part because it’ll help segregate those things and help be able to slice and dice the data as you need.

Stephanie Broderick (18:41):
I’m curious, and I’m going to go a little bit off-topic. When you think about the six areas that The Sequoia Project is focused on that the group prioritized, is there an area that you think has the opportunity to advance the needle more than the others?

Dr. William Gregg, MD, MS, MPH (19:04):
My gut feeling is I really think that if we can make progress in the lab space, it will be very meaningful. So much of healthcare is at a junction point that involves laboratory testing. And right now clinicians see one piece of that, but they’re not able to assimilate all the things together and really make meaningful assessments of it without a lot of time. And so if we can solve that all the way from those who are creating the data to those, passing it along, there’s so many steps along the way and it fails at so many points, I think that would be a great example of something that we can solve. And I think it’s solvable because it’s all discreet data and it’s one of those pieces and it won’t solve everything, but I think it would be such a meaningful change for our users that I’m really hopeful about it, that we can make progress.

Charlie Harp (19:58):
Well, I was going to say that the lab is important also because if you think about the data that we get in healthcare, there are certain aspects of the data that are subjective or theoretical, lab data is not. It is hard data. Even medications you can say, hey, this person, I prescribed these meds, but you don’t know for sure if they’re taking them. When you talk about their diagnoses, there’s so many things that are really, we are kind of saying we think this is what’s going on with the patient. We believe they have this problem. Lab data is one of those things like vital signs and other types of telemetry that is hard data and it gives us a lot of really good information. So if we can solve for that fills in a lot of the picture.

Didi Davis (20:42):
We’re really trying to, with version two, pick those pieces apart. So we’ve had some good discussions already in the lab Tiger teams that have started thinking about, okay, guess what? There’s six different codes for albumin, which is one lab value when all of these vendors put these systems in their customer sites with the meaningful use dollars seven years ago, a lot of these sites have never updated those codes. There may be a better code to use for albumin based on how they tested it, how was that test resulted? So thinking and making folks really consider, are there areas that I can make this IT system harness its power and give us that data a little bit better?

Dr. William Gregg, MD, MS, MPH (21:27):
Yeah. Well, and one thing I would add to that is clinicians and clinical systems, people are often either lumpers or splitters, I want to group all these things together or no, we’ve got to maintain the fidelity and we can do both. And it’s how we utilize a system and everybody playing from the same game plan to say, here are things that I can safely lump together and not reinvent it at every site and share that information as a community, that’s going to be so critical. And that’ll involve the providers, it’ll involve the vendors of all the different systems that are part of it. It’ll involve the lab teams. And so if we can do that, it could be a big win.

Charlie Harp (22:08):
I mean just to give an example of why organizations like Sequoia are important. When you talk about orchestrating over a community, take COVID for example. When COVID hit and people were doing all the testing for COVID, we were involved in helping to wrangle some of the data, some of the lab result data for COVID. And the problem is with no orchestration and with no guidance, there was like 34 different ways to semantically describe COVID. There was 72 ways to say positive or negative for COVID. And so the problem is when you start out with that kind of chaos, you have to put so much work in to align that data to do anything meaningful down the road. If somebody at the beginning or if somebody can orchestrate and say, this is how we’re going to do this before you really get started, it saves so much time and effort and creates a much more agile environment down the road when you actually encounter and have to use that data.

Stephanie Broderick (23:05):
So there are a lot of organizations that are trying to solve the data quality and data usability challenges. You’ve got organizations like HL7 trying to define standards implementation guides. So how do you guys work with those organizations and coordinate so that we’re all trying to solve the same problems in the same way?

Didi Davis (23:25):
That’s a great question. Well, luckily Sequoia has been working with HL7 other standards organizations like SNOMED and so forth for a very long time. From the HL7 perspective, I’ll give you a couple of examples. So in the version one guidance that we put out, we actually worked with HL7 to help us develop some of the examples that we put in there. So one of our use cases was patient-centric. We wanted to have a way for folks to understand what’s the best practice when a patient wants to self-report that they’ve received this immunization or that they’ve had COVID. So that was something that all the vendors are saying, okay, how would I even put that in some electronic transaction? So we went to HL7 and we said, Hey, this is some guidance we’d like to have. We worked with what they called the structured documents workgroup, and we created an example with that workgroup and we actually included it in the guide.

(24:22):
So when we identified a need for something, they were very responsive in helping us pull that together. Another example, and we keep doing this, they have what are called implementation-a-thons where they kind of bring vendors and users together to work rolling up their sleeves and figuring out what are the problems and how can we address them, what can we do to improve it? They have one every April and August and the last three implementation-a-thons. We’ve had some kind of track to help raise visibility, have that discussion to then inform our work that we’re going to be putting out as well. So those are just s-me examples, but we pride ourselves in trying to be that feedback loop because the standards organizations, unless they actually get the feedback, they don’t know what they can improve. But that’s just one example.

Dr. William Gregg, MD, MS, MPH (25:09):
And I would add too, that many of the standards organizations, they’ll work with the vendors and they’ll work with, those are looking from an academic perspective, but they may not see all the things that are happening in the field with the providers. And so that’s part of what we do is we bring all those together. If you talk to most hospitals, especially your smaller hospitals, the teams there probably are not aware of what HL7 is doing day to day. And so it’s again, goes back to that idea of community is this is a shared exercise. And I also would add that it’s really important that we bring everybody along that it’s not just the bigger hospitals and the academic medical centers that do this. We’ve got to hit our safety net hospitals and those organizations because every patient is important. And if all of these things can work together and maybe it’ll work differently at those hospitals, they’ll need a regional HIE to help. That’s okay, but let’s find a solution that includes all of them.

Stephanie Broderick (26:07):
I know we’re getting to the point of getting to questions, but what are some of the areas of focus for the version two implementation guide and a little bit further beyond that? I know when we were doing the Tiger Lab at first we didn’t have participation by the labs. What else do you guys need? What other types of organizations do you need to come to the table to help with these efforts?

Didi Davis (26:32):
So I know I’ll start and please from a clinical perspective, keep me honest. But so the high-level items that we’ve really scoped in for version two that we’re looking at trying to do the first version guide was really more focused on CCDA clinical document architecture that everybody is using today because that’s what’s real, that’s being exchanged. But there’s a lot of work starting to happen with FHIR. So we’re going to be more technology agnostic and making sure it could be applied to version two messaging lab, which is more version two or FHIR messaging or even the CDA documents as we were talking about. So that’s one place. FHIR being added, the laboratory that you mentioned earlier, we really are doubling down on trying to have something from that. And with the Tiger team efforts, we’re hopeful that we can at least help raise visibility and educate the users to understand what they don’t know.

(27:29):
I think that’s part of it. They don’t realize there’s things they could improve. So we’re really going to try to include the guidance for the lab piece. And then overall, in my mind, it’s trying to think about the receiving systems too. The first version focused on how do you send the data, how do you send it to improve that usability? We also want to look at how do you receive it so that then clinician can maybe move the sections around easier or slice and dice the data so that a cardiologist may want to see something higher up in a document than a normal primary care provider. So thinking about those type of receiving system enhancements.

Dr. William Gregg, MD, MS, MPH (28:07):
I agree, those are all important. And one piece that we’re looking at, and this is a tougher thing to crack, is a lot of testing even whether something is conformant or not happens on QA data or test data. But where I see it break down is in the real world, what we receive. We have systems across the country that pass with flying colors, but when you actually put them in production and there’s real patient data, it breaks down because that’s much more difficult to measure. I think we’ve got to get to a point where we can safely and effectively measure the quality of actual data and the usability of that data, and that is a whole nother level, and we’ve got to crawl, walk, run. But that is something we’re going to hopefully be able to start pushing in that direction, this implementation guide, but it’ll be definitely a future thing.

Stephanie Broderick (29:03):
And I think we’ve got some ideas around that, around measurement. So I want to make sure that we leave time for questions. Do we have any questions from the audience?

Audience Speaker 1 (29:14):
I have a question for Didi.

Didi Davis (29:16):
Yes, sir.

Audience Speaker 1 (29:17):
So we talked about providers build great inventory and make sure it’s complete about provider organizations. I’m curious, when I was looking at the various guidance documents in terms of the use in research and applying this for things like the enclaves that are built for N3C or for all of us and those kinds of initiatives, it looks like there’d be some applicability. Is that in scope now or is that on the roadmap?

Didi Davis (29:41):
We’re really focused on the three use cases provider to provider exchange. So clinical for treatment purposes, typically. What we call provider to public health, so public health is not research really, and public health back and then healthcare entity to consumer. So we’re really focused right now on those three overarching use cases. Research is definitely something that we want to get the data flowing, but I definitely think there’s a place for that. And again, this is not one-and-done, so it may be something we focus on in a future version. It’s not going to be something we focus on for version two, but we definitely are open to what we can, especially with the focus from the Biden administration with the Cancer Moonshot, there’s a lot of emphasis on research capabilities even from COVID. So we want to try to see how we can support it. We got to get that data right, high quality and usable before we can start doing the secondary part of it.

Dr. William Gregg, MD, MS, MPH (30:34):
I think a lot of what we’re focusing on will benefit research just naturally. One of the ones that we’ve deliberately stayed away from is provider to payer, just because DaVinci is focusing on that and it becomes much more of a third rail sometimes among providers because all that is done in very prescriptive ways. But I think that research efforts will benefit significantly from all the quality efforts that we’re doing. I mean, you imagine we get lab figured out or at least improved, that’s such a big part of research and it’s not solving that use case directly. But I think it’ll have benefits, but I would imagine it would be one of the more natural ones for us to get to over the near term.

Audience Speaker 1 (31:20):
And so Charlie, you made a point about knowing what data is not there as being critical. And when I was looking at the NIH All of Us program, they’re trying to get as much data as I can from as many places. So the amount of missing data across domains seems like it’s going to be an exponentially increasing problem, which I hadn’t really thought about until you mentioned the knowing what’s not there. How do you come across that when you’re looking at industry and data?

Charlie Harp (31:50):
There’s this idea that if I get data from everybody, it’s going to fill in all the missing puzzle pieces. And I think that’s kind of true. But the other thing you have to realize is kind of the explosion of the reason why we have value sets. There are 350 codes that mean diabetes, and it all depends on the provider and the moment who chooses the code that they’re choosing to represent something. And so there’s a certain amount of coalescence that has to happen. But there are also situations where a provider will put something in a clinical note and they won’t put it into the structured data because the clinical note is their system of record. That’s what they really use. And when I talk about missing data, and we’ve done projects like this where we look at the data that’s there and we infer the data that’s missing. For example, you look at a patient and they’re on metformin and they’ve got a hemoglobin A1C of 7%, but there’s no mention of diabetes, but they have diabetes. So some of these things are not just about filling in the gaps. Some of these things are about looking at the data that’s there and seeing what’s not there. It’s like an interference pattern.

Audience Speaker 1 (32:54):
Thank you. Thanks.

Stephanie Broderick (32:58):
All right. Thank you. Any other questions? Okay, if there’s no more questions, I have one more question. All right. So in the HAE and interoperability, pre-form, pre-conference forum, Marianne Yeager, president of Sequoia threw out, and I heard you state the same number, which horrified me to be honest with you. So 10 years, 10 years, 10 years until we solve this? 10 years until what?

Didi Davis (33:34):
So yeah. So Marianne, for those that weren’t part of the pre-conference symposium that we had the forum, there was an HIE and interoperability forum, and she was our opening keynote, so our CEO from Sequoia. And the item was that it’s going to take 10 years to actually make a real difference. So it’s going to be incremental. I keep coining the term baby steps. I wish we could do it faster, but unfortunately, we have so many competing priorities. Vendors have regulations they have to meet. They’re looking at reporting things that they have to report for quality measures. So this data quality and usability is one piece of it, and we need to leverage legislation and regulation as it exists. Version one, we focused on the US core data for interoperability version one. Our second version guide will be focused on version three, but as the ONC continues to release the USCDI, we’re going to continue having to adopt and pull those in. We don’t want to create something that’s not already on somebody else’s radar, a vendor or an implementer itself. So yeah, it’s going to unfortunately take, we’re estimating at least 10 years to make enough difference so that research and some of these things can truly benefit because you’ve got a complete data set. That’s why I think she said that that way. But be patient, know that all of you have a part to play to help us get there.

Dr. William Gregg, MD, MS, MPH (35:00):
Yeah, I actually think that’s very optimistic. And because this is a big problem, and we’ve been digging a hole for a long time and making things more complex and less organized, and we’ve just been piling new technologies on. And I think if we end up 10 years from now, 10 years from now, we’ll either be exactly where we are now or we’ll have improved it. We might get, I think we’ll get to a point and maybe it’ll be seven or eight years where the pace will accelerate because we’ll have done enough base and foundational work that it can move more quickly.

(35:37):
But we can’t underestimate and I don’t want anyone to underestimate the challenge and think we’re going to do it in two or three years because I want everybody to participate and to be part of that hard slog to get through it. And then we will be in that better world when I need more healthcare as I’m older. And I will be very ecstatic whenever things are much better organized.

Charlie Harp (35:58):
Well, I couldn’t disagree more. I think blockchain and large language models will have it all sorted out in the next three to four months. I’m fairly convinced and I have some swamp land not far from here that’s for sale. No, I agree. I think that I’ve told people for a long time that healthcare as an industry is disrupt-proof or at least disrupt resistant and the vital nature of what we do and all of the people that it impacts and how it involves, the only way we’re going to make meaningful progress is through pragmatic, incremental improvements that we institute across the organization and with a significant amount of buy-in. Because one of the things we have to remember, I always say healthcare evolves from the edges that people in the trenches of healthcare are vital. And if we try to dictate from any standard or government or central environment, thou shalt do it this way. It almost always backfires on us one way or another. So the way to do it is doing incremental improvements where the people in the ecosystem see the benefit so that they buy into it.

Stephanie Broderick (37:08):
And I have heard Didi many times reiterate that exact same concept on the workgroup calls, and even in the HIE and interoperability showroom or showcase, just incremental.

Charlie Harp (37:20):
Yeah, that’s why she’s my hero.

Didi Davis (37:23):
We got it. We keep inspiring everybody. So all of us working together again, and I’m very grateful for being able to inspire and hopefully have the passion to try to make that difference. As Bill said, I’m getting older and one of these days I’m going to need the healthcare. I’m already a caregiver to my elderly mother-in-law and my father. So I understand the problems that we have, and I know that I would like it to be faster. But I remind another thing that Marianne mentioned earlier, the financial industry, it took them decades to get all of the banking flowing the way that it does today. You go to an ATM, anywhere in the world, you can get money. Now their data set is nowhere near as complex as ours. So another decade, it sounds like a lot. But there is a very complex problem, and those baby steps are important, but it does take the actual users helping us to get there because nobody wants to put in another project. But we want to try to inspire usability in all projects. Every time you touch a new healthcare system that you’re implementing or upgrading, ask the questions, what can I do for that clinical workflow? Because a lot of these electronic health records were put in mimicking the paper process. But guess what? If you harness IT systems the right way, you might be able to optimize some of those workflows.

Stephanie Broderick (38:42):
Wonderful. So with that, we’re going to close. I want to thank the panel Didi, Dr. Gregg, Charlie, amazing panel, and thank all of you for being here to see it. Thanks.