One of healthcare’s biggest challenges continues to be data exchange. Leveraged correctly, health information technology can help automate efforts, increase transparency, and reduce miscommunication between health plans, providers, and healthcare organizations. Working towards that goal of improving communications, in order to enhance patient care, is happening at health information exchanges (HIEs) nationwide.
In that regard, Morgan Honea, CEO of CORHIO, the Colorado Regional Health Information Organization, a Denver-based HIE, will be one of several healthcare IT leaders participating in a roundtable discussion entitled, “How Colorado Is Redefining Interoperability,” and which will take place on Monday, July 15, during the Rocky Mountain Healthcare Innovation Summit, sponsored by Healthcare Innovation, and which will be held at the Grand Hyatt Denver in downtown Denver.
And on Tuesday, July 16, Honea will co-present with Dr. Timothy Dudley, chief medical officer for Colorado Care Partners, a HealthONE accountable care organization (ACO) and clinically integrated network (CIN) that encompasses more than 320 clinicians, to share with the audience how their organizations are collaborating in order to gather meaningful data for their front-line care management initiatives, through the use of real-time notifications for hospital stays and emergency department visits in the ACO’s patient roster.
Recently, Honea spoke with Healthcare Innovation Editor-in-Chief Mark Hagland to discuss the range of activities that CORHIO is currently involved in, in advance of the Rocky Mountain Healthcare Innovation Summit. Below are excerpts from that interview.
Tell us a bit about the collaboration between CORHIO and Colorado Care Partners? It sounds exciting.
Dr. Dudley is absolutely phenomenal. And he is leading HealthOne’s efforts to gather the types of data that they need to operate as an IDN [integrated delivery network]. And the thing that makes them unique and the way they’re leveraging with CORHIO—is that as they partner with new organizations, they’re trying to keep in place existing infrastructures rather than imposing the cost of ripping and replacing IT infrastructure. And the way we’re partnering to do that is by leveraging CORHIO’s infrastructure to help them manage populations and integrate physician groups. It’s happening in a couple of different pockets in the state, but Dr. Dudley’s is one of the more progressive and assertive in leveraging growth opportunities.
What have you been working on with them?
Bringing their practices into the CORHIO network. We’ve been very successful at doing that, by leveraging different federal and state funding opportunities. So we’re working under the Medicaid IAPD Program out of CMS [the federal Centers for Medicare and Medicaid Services]. It’s targeted to each state’s Medicaid agencies, dollars targeted for expanding interoperability and HIE. So we’ve been working with the state Medicaid program to help subsidize people coming into the HIE. We’ve worked to target practices important to the Medicaid agency and to these physician networks, to leverage federal dollars available. And from HCA’s perspective, they’re growing out quickly, and if they had to grow this out manually, they’d essentially be doing what we’re doing, but not everyone would have access to the network. So it’s a really great synergistic approach to leverage available funding and resources to meet the needs of HealthOne but also of the broader community.
What are some of the deliverables involved in your collaboration?
We’ve developed a three-phase approach. It involves going into each of the practices and giving them access to our provider portal, so they can see the types of information going into CORHIO from the hospital setting without having to create new interfaces; and then we extract information from the clinical setting. The goal is that you’re not ripping and replacing technology.
What interfaces have you been building in order to allow participating parties to see data going into CORHIO from the hospitals and from other medical practices?
Yes, we’ve been building interfaces, exactly. And the third leg of the stool—the thing I want to showcase in all my conversations—is that as we think about a lot of the federal policies and rules being proposed like TEFCA [the federal Trusted Exchange Framework and Common Agreement], interoperability rules, etc., all of that is based around query and receive interoperability, where a provider goes in and seeks information. One of our board members calls that “fishing” for data. So steps one and two give people the ability to go fishing; it’s aggregating the data and giving people access to the data. But the third leg of this stool is notifications. It’s about moving beyond this query functionality and into this push type of functionality. We know who the patients are whom the ACO is trying to manage, and we can set up triggers based on different events or different KPIs [key performance indicators], so that if something happens, you get the alert. So we go in and create access data, access more data, and set up triggers to let clinicians know when things are happening in their populations.
That’s the leading edge in all this, correct?
From my perspective, it is. It’s being done a lot right now. We’re among the few doing KPI-based notifications, combining our ECQM [electronic clinical quality measure] and functionality work. Electronic clinical quality measures. Combining that with our core HIE functionality. A lot of folks do ADT [admissions, discharge and transfer] alerting; but because we have such a broad scope of data, we can do not only ADT alerting, but also results alerting when an a1c comes back out of range, for example. If you’re running a clinically integrated network. So if you’re running a CIN and you’re running a KPI around diabetics being in or out of control. So we can see who your diabetics are, and we can also see the values of their a1c lab results, to notify when a change occurs, and the groups can deploy different types of chronic disease management approaches; and that’s really important, because if you’re thinking about deploying different types of approaches, you don’t want to deploy the same type of approach to someone who’s had three years of controlled a1c and someone who’s consistently had their a1c grow higher and out of control. And so having access to that kind of data is very important.
Is all that live now?
The functionality is live. I’m not sure whether it’s live with HCA. But we’ve been doing that for about two years now—triggering notifications on KPI-based data elements. We did one for diabetics, we did one for labor and delivery, around reentry into primary care, post-delivery, for the Medicaid population; and we did one for ED utilization, so that was ADT notifications. So what we’re doing at CORHIO is continuing to build out the foundational infrastructure to enable the expansion of all those services, focusing on what the ECQMs that people will be held accountable for, and how we can create standardized data assets around that. So late last year, we went through a process of LOINC-coding all of our lab values. And that’s very important, because if you’re trying to build triggers around hemoglobin a1c results, then you need to know all the results around an a1c in a lab set. We ended up taking 12,800 unique lab codes and normalized them in our system, and ended up with 1,200. We had had something like 240 codes for a hemoglobin a1c in different patient care organizations. So we could have five different lab codes for an a1c coming from a single facility. So if I’m trying to give notifications about what’s happening with an a1c back to Dr. Dudley, I need to make sure I’m capturing all those values. So we went out and normalized all those values. We normalized the 240 codes down to one, so he wouldn’t have to worry about normalizing values himself. So we’re doing that across our entire network.
How long did it take to do accomplish that normalization?
The process is iterative, and ongoing. The first time you load a data set, it’s pretty messy; the second time, it’s better. We’re on our fourth iteration of lab value normalization, and that has taken about a year.
What have been the main lessons learned so far in this work?
It’s slow, and you have to do a lot of stakeholder engagement and voice-of-customer work to make sure that your work is creating things that are of value. It’s all very interesting. There’s never a day where I sit there and look at the data and think there’s nothing interesting about it. There’s always something interesting going on with the data, but the key is to engage with your stakeholders and find out what their pain points are and how to leverage their resources. So working with our stakeholders to say, where are the areas where you’d be willing to put financial resources into things that will make a difference.
And in order to do this work, you really have to be a diverse technology shop. There’s not one vendor or platform out there that does the scope of what we’re doing. So you have to be able to use different technologies, and put them together in an architecture that makes sense, because there’s no single technology that will do it all for you.