The CIO of Arizona's Health Current describes the health information exchange's efforts to serve its participants during the COVID-19 crisis – and discusses its ongoing efforts to boost data quality and consistency across the state.
Phoenix-based Health Current, Arizona's statewide health information exchange, serves more than 500 organizations: ACOs, behavioral health and community providers, emergency medical services, rural health clinics, hospitals and health systems, payers, labs, LTPAC organizations, and state and local health agencies.
For some time now, the HIE has been engaged with a major initiative focused on data governance and data-quality improvement, working to integrate more clinical and behavioral data across its network. Having seen 700% growth over the past three years, with its participants becoming ever more diverse, Health Current has been grappling with big challenges related to the sprawl and diversity of coding terminologies and data content across its stakeholders.
Since 2017, it has worked across its community of data suppliers and end users to work toward more commonality and uniformity in how data is relayed. That approach has been helpful in recent weeks as Health Current has had to pivot to focus on marshalling healthcare-data resources across Arizona in response to the COVID-19 crisis.
The HIE is working to ensure its data centers and IT infrastructure are solid and backed by multiple redundancies, even as its team members are observing social distancing by working remotely. The aim is to ensure all participants have easy access to critical data without any drop in service levels.
Health Current is enabling widespread access to longitudinal patient medical histories and data from patient encounters during the pandemic via its secure online portal, and with tens of thousands of alerts sent to clinicians and other healthcare staff. It's also offering expanded services for members to help monitor vulnerable populations, and prioritizing recruitment of new participants to expand its range of customers.
Healthcare IT News spoke recently with Keith Parker, Chief Information Officer at Health Current, to see how the HIE is serving as a critical public health infrastructure.
Q. How has Health Current pivoted in recent weeks, in light of the COVID-19 pandemic, to better serve the healthcare organizations across Arizona that rely on the HIE?
A. We're reaching out to our stakeholders and communicating with them about how best to use our data and the infrastructure we've put in place here in Arizona: everything from data-mining our database for our state or other authorized individuals to use for hot-spotting, to using our alert mechanisms, so when different labs or different high-risk patients come across, we can make sure that their care teams and their providers are aware of it as well.
Q. What is the size and scope of Health Current?
A. We don't really frame it in the context of the number of providers, but instead look at it through market-segment penetration. So we have about 98% of our acute care facilities connected and sending information bidirectionally – sending and receiving information.
Q. So the HIE is serving as a pretty critical information backbone to kind of help coordinate response to this crisis statewide.
A. We do have an infrastructure that is a pretty critical backbone for the state, as far as alerting and getting data to distribute across our healthcare community. And we send out, from a broad message perspective, well north of 20 million messages a month. But from a raw perspective, one encounter could equal several different message types that are going back and forth.
Q. That's just on a typical day?
A. That's typical.
Q. So you expect this to ramp up as the crisis continues to unfold.
A. Yes, and we've got the infrastructure in place. We've been doing north of 10 million alerts for the past almost year now.
Q. Talk about the importance of data quality.
A. We're on a journey to improving the data quality here in Arizona from a systems approach: How do we actually put in place mechanisms that engage our stakeholders, as well as the infrastructure that we have currently, to ensure that we have quality across the data continuum.
And what I mean by that is the data has to be input directly from the data source to us: The connection has to be built correctly – and to be built correctly as far as the types of data – and from a validation perspective: Is it actually syntactically correct, coming across? Is it complete and is correct?
We started down this path a little over a year ago, and I've got to tell you, it still feels like we're just at the beginning of that journey. There's just so much work to do to be able to clean up the data, make sure it's accurate, make sure that it's coming across on a regular basis.
Q. How do you approach that work?
A. We take an approach that's more collaborative. We have a data governance council. A lot of our ideas go to them. We put in place standards that go through our data governance council that we then take to our board, which is a representation of Arizona's market: payers, hospitals, providers, behavioral health, physical health. And then we move forward with it.
So from a data-governance perspective, that collaborative approach that promotes getting in the weeds, we actually have started running data-quality reports back on a subset of our facilities to say, "In these segments, this is what we're seeing." We've identified that we're missing specific maps, because not everything meets national terminology. And some of the areas that we've had to do this in his round of facility naming conventions, because what they name their facilities is not going to be in any national terminology book.
A. That's where having the foundation in place is valuable. We already know what those codes are, so we're just going in and validating that we're receiving those codes that we can data-mine against those codes. So when you take a look at it, it's the CPT codes, it's the LOINC codes for the labs, it's the CDC guidelines for symptoms or other areas that they're looking at.
So we've pretty quickly developed what that compendium looks like, and then we could go in and validate. But part of what we've done is, when we bring in our large data sources right now, those 98% of our hospitals, our acute care facilities that are sending us information includes labs. We've already got the foundation in place to be able to move that information back and forth. So it's just a matter of us going in and validating: This is the appropriate coding that we're looking for.
And then we can connect with the data sources and say, yes, that's how they're settling in as well. When it comes down to the national coding elements, it actually makes it a little easier because from a LOINC perspective, they follow those guidelines.
And if they don't, we can identify it. Then we can talk to them as far as how they actually code it. And then we have the appropriate data maps in place to be able to bring the data across as usable data. So the foundation really lends itself to us being able to stand up pretty quickly, to be able to say, as the testing is being distributed, we can then already have mechanisms that we're wrapping up or putting in place to be able to alert on those tests.
Having the right pipes in place is key – having those pipes support national standards, and having the appropriate mapping in place so we can make that data usable, and then doing spot checks (Is it actually working?), and validating against it. So from that perspective, it makes it a lot easier when something like this does arise to be able to work with our stakeholder and say, "How can we support you better?"
Q. Do you get the sense that the stakeholders, generally speaking, are appreciative of the role you're playing here and are using you to the fullest potential?
A. I would say here in Arizona, yes. So, we've got a pretty collaborative state when it comes to working together and sharing the information. We try to hit 90% or better on the data quality coming across. But the different data types is going to lead to natural variation that's going to be in the data coming across.
And our data sources work hand in hand with us to make sure the data quality is there. And then the end users: how is the data actually going to be used? A lot of our data goes out to those care teams, those care navigators that are supporting the rest of the care teams, the providers, the hospitals, the outpatient facilities on how to better manage those patients. The 10 million alerts that we send out are because most of our complex patients are on multiple provider panels.
Q. In recent years, as interoperability and data exchange imperatives have evolved, many HIEs have similarly innovated the services they provide. Talk a bit about how Health Current has changed its own business model in recent years.
A. Like evidence-based medicine, we pursue an evidence-based strategy. When we started sending out alerts we were at maybe 10,000, we thought we're doing pretty good. But when we reached out to our stakeholders and actually had deep conversations as part of what they really needed with the individuals who are using the alerts, we went from less than 100,000 to over 10 million a month.
We saw individuals starting to update their panels, stay more in touch, so it's really listening to our stakeholders. We do that. We take that same approach when it comes to data use from, say, pop health or analytic platforms. So we've spent a lot of time and efforts on developing different queries that we hit against our system to be able to support specific measures.
So if Organization A is using a given analytics platform, instead of just giving them CCDs, we give them the specific national codesets that they're looking for for those. Along with specified demographics, it's a lot cleaner. It goes into their system a lot easier. We do the normalization on our end, so it just feeds right into their environment.
We also work on it probably hitting their system in a few cases where we can support them with the alerts, to where it can support a care pathway. So they're using IP or solutions out there to be able to say, "I've got a diabetic; he presents in this way; this is the recommended pathway for the care team," they need that information to hit the system.
And we provide a mechanism when it's done to be able to hit their system like that. And we try to do it from a real-time perspective, so that information comes across. We immediately set it off to their system. The coding is then in place for them to be able to run whatever algorithm is downstream with their partners, with their care teams, as well.
So it's really listening to our stakeholders and understanding the direction they're going. It's also looking at it from an integrative perspective. We've the approach to where we're not shying away from 42 CFR information, if we bring in substance abuse information and make it available in accordance with SAMHSA, with national direction on that as well.
So it's really listening. As our environment's going to integrated facilities, integrated care, it's making sure that we have a data infrastructure or a direction that we're moving in that can support that as well. So it really goes from HIE to data management. So from the exchange to how do we actually manage data across market segments, and aligning both state and federal guidelines.
Q. We've seen often how in times of natural disaster, whether it's wildfires or hurricanes, HIEs really serve as key infrastructure: a "public utility … as critical as having roads, as having fire hydrants, as having an electricity backbone," as another HIE director once told me. Is that how Health Current sees itself as the COVID-19 crisis unfolds?
A. We see ourselves as a partner here in Arizona to make sure that we're providing the best care possible. So we see ourselves definitely as a piece of that puzzle. But only a piece of the puzzle. In hurricanes, when records were lost, well, you've got at least the start of a longitudinal record inside the HIE that you can begin using. And as the data elements become more complete and the quality improves, that only gets better over time. But yes, we definitely see ourselves as an integral part of the healthcare system and providing better care.