How Is Patient Data Consolidated and Presented to Doctors in the US?
In the United States, individuals' healthcare information is dispersed among various healthcare providers. But many companies have been working on creating consolidated patient views, Reveleer being one of them.
Data fragmentation often occurs because people tend to switch healthcare providers when they change jobs and, as a result, their health insurance plans. Since insurance companies have specific networks of affiliated healthcare providers, a change in insurance necessitates a change in providers. Because providers use different information technology systems, individual healthcare data becomes compartmentalized and difficult to consolidate.
In this episode, Jay Ackerman, CEO and president of Reveleer, a healthcare technology workflow, data, and analytics company, supporting payers and risk-bearing providers in their value-based care programs, explained:
how Reveleer consolidates patient data to give clinicians a single overview of the patient,
what are the biggest pain points in healthcare data management in the US,
how is generative AI affecting Reveleer’s product development?
This is the full raw transcript:
How would you describe what your company does?
Reveleer is a data and analytics company that supports payers and risk-bearing providers in their value-based care programs. Essentially, we aim to help them be properly compensated for taking care of members, whether they're healthy or really sick, and we want to reward them when they provide great care to their members and patients. A lot of this is achieved by capturing clinical data and the medical record, and understanding how service has been provided to them and what has happened during a member's health journey.
You aggregate data from over 45 EHRs, if I'm not mistaken. Can you discuss that a bit?
Yes, 45 EHRs and probably 70 HIEs. Capturing and aggregating data on members and patients is really hard. Members and patients will see care providers all over the place. You might be on vacation and pop into a minute clinic, go out of network in your neighborhood, or enter an emergency room setting or urgent care. That data sits in those siloed EMRs and doesn't get consolidated. So, we've worked directly with EMRs to connect to them for larger health systems and with other providers to access data through HIEs, where there's a trading and sharing protocol around how information is shared to make it more publicly available. We work across a lot of different channels to pull all that together and have an accurate view of a member when they come in to see their doctor.
What's the most difficult thing for you in this process? How much do you have to update the connections and integrations that you already have for data consolidation?
A lot. Actually, I'd say that the bigger challenge is, when you are able to capture so much data on a member or patient, how do you then consolidate it? Distill it down to the most important issues that might be going on? Because when the doctor sees that patient, they have 10, 15 minutes. So if we capture 1000 pages of medical record data, we need to surface the three or four things that are most important. Using machine learning and AI is what we do to take that large data set and consolidate it into something actionable.
What has changed for you with the emergence of generative AI? What's your story with that technology?
We've been in the AI journey and the AI world for going on five-plus years. We've been using what's called more traditional machine learning AI tools, and we're now experimenting with generative AI and large language models. But we have work to do in how we leverage those tools. One of the things we have done with generative AI is enhance our clinical dictionaries. We've used those tools to pull a lot more data that we'll be able to run through our algorithm.
Can you talk a bit about how data harmonization works when you pull data from different EHRs? How do you create a consolidated view based on the different types of data?
What we generally do is sweep all the available data sources within, let's call it, typically a 20-mile radius of where that member lives. We take that data, pull it all in so we have a single record of that member, then run it through our AI. That's where we're looking for previously existing conditions on a member, where we might connect a couple of conditions and then identify a suspected diagnosis that we think is going on. Then we present that in a single patient compendium to the provider. So they have one page to look at when that member comes in, and they can click through that and go deeper if they want to see the supporting evidence on it, but we're giving them one single page to look at.
Is your system integrated into all the EHRs, or is it just something that another screen that the doctors need to open to have the picture of the patient?
Oh, now you're talking to the pain point of a doctor. We're able to do it in a variety of means. There are docs who still actually would prefer a PDF printout, and we'll serve it up in that fairly traditional and inefficient manner. But in the most advanced and most efficient, we'll show up in the EMR, where it creates a nudge for them when that patient comes in, they'll know that there's the Reveleer Compendium for them to review. And that's the preferred native in their EMR. Not an additional location, not an additional website or portal for them to have to access.
What's your biggest pain point for you and the whole development of healthcare data management in the U.S.?
I think our biggest pain point is continuing to optimize that workflow. So we are able to show up in the EMR. That takes a lot of work to be able to do that in such a wide array of EMRs. We're able to push through to that pretty easily, but then writing back to the EMR has its challenges, but we are writing back and we're looking to use our AI tools to write back the doc patient's notes AI-assisted on behalf of the doc.
How would you describe the collaboration between healthcare IT providers in the U.S.? We mentioned like 45 EHRs, that means Forty-five integrations, 45 discussions to get those APIs working.
It's a land of co-opetition, and look, some smaller players are obviously more willing to collaborate. Bigger players, like some that might sit in Minnesota and have a large EMR footprint, they're less likely to work and collaborate. They want to run you through their app exchanges, which aren't that friendly. But it's not super, it's getting better and we're finding ways to work within it and also ways to work around it and provide better support to our payers and providers.
And what are some of the key things that you hope to achieve still, say, in the upcoming few years, especially with future development of generative AI?
Look, I wish we could capture data across the entire country on a single member, single patient. And we can't do that efficiently. So that's hard. We identify, like I said, like a 20-mile radius where we're sweeping. And that we've been pretty successful with that. But look, ideally, wherever you've seen a doc, that should all be rolling into a single place where you have a consolidated picture of your health journey. And I've seen that play out. Like I have a mother who passed away this year. Medicare Advantage member. She suffered an accident outside the U.S. She had medical data there that was not transported back to the U.S. She then saw a bunch of specialists and that information wasn't flowing easily and smoothly. And so when you go through that and you see the limitations of it, it does, it drives you to do a better job at it.
Patients often change healthcare providers due to job changes and healthcare plan alterations. How do you ensure a consolidated picture of the patient when they are in different systems? Do you use patient identifiers to know it's the same patient?
Within our system, we have a unique patient identifier. So, if a patient moves across from one health plan to another, we're able to maintain a single view of their health. When we're sweeping and capturing data, it's irrespective of what plan they were a part of. But that is a critical challenge. It's like having, you know, your social security number. We're not using that, but a unique identifier allows us to keep a long, longitudinal view of your health.
You also store data. Do you have any plans, or can you even do any secondary use of data because you are building a huge database? What's your use case? How big is your database?
This year, we will ingest over 150 million pages of clinical data alone, just this year. And probably two or three times that number next year. So, we have a growing clinical data repository. Today it's used solely for the payer or the provider that's contracting with us. Could it be used elsewhere? Perhaps, but that's not our focus at the moment.
How do you see the growing number of data ponds that are present in the U.S., like Komodo Health and Epic Cosmos, especially considering you have a large data pool? How do you perceive that in the healthcare space and the impact it might have or already has?
I think there's an explosion of data, and there's a wide array of uses for that data. But there are a lot of concerns about how people might package that up and use it. Right now, we're staying focused on a pretty important part of the value-based care arena. And we're 100 percent committed to that right now.
One thing I'm wondering about data consolidation is that, as I mentioned, many companies are doing it and all the companies probably have different approaches to how they're going to normalize and harmonize that data. So how do you see that? Is there a danger that there would be discrepancies or inaccuracies because everybody has a different process in data cleaning?
There's certainly going to be differences in how people standardize, normalize, and how they uniquely identify a member or patient. And I think the risk is, I'm Jay Ackerman, and there's a Jason Ackerman, and somehow our data gets combined into a single view. Clearly, we don't want that to happen. I think there are risks of that and risks that people are going to take that data, package it up, and use it in ways that they shouldn't. So for us right now, in terms of our mission, we're focused on making sure that the information on Jay is only Jay's, and the information that we have on Jay is used solely for the purpose of making sure that when they walk into a provider, we accurately understand what's going on and we're able to take the best care of them and be able to report that accurately on behalf of the payer and the provider.
Who decides that you can get patient data? Do patients have any say in that, or is it just the provider who starts to work with you and they say, we're going to do this?
The provider, who provides care, has the right to that data, and the payer, who's providing insurance coverage on their behalf of them, has the right to the data. And that's who we're working in support of their needs. This is why we don't take that data and then turn around and package it up for some other purposes.
What are you still missing in the data management interoperability space in the US?
We're lacking common standards. What I just described and what we're doing to capture data, like, I equate it to dragging for fish in the ocean. You put a net in there, you're gonna get some really big fish. Maybe if you layer a second or third net over that, you can catch a lot of smaller fish. And to get all that data, you have to go through a lot of different vehicles to do it. And so it's pretty inefficient. And still, at the end of the day, there are data points that slip through your hands.
Is there anything specific that you would like to see in terms of the progress made in the interoperability and data standards space?
Yeah, look, I'd like the big players to be more open and sharing. And so just because you have the lion's share of the market doesn't mean you can put up big walls around you.