Behind-the-Scenes Look at Providence-Humana FHIR-Based Connection

In an October news item we described an effort involving Humana and the Providence health system that leverages HL7 FHIR standards, Da Vinci Project Implementation Guides, and APIs to improve interoperability between the payer and provider organizations. Last week, we took a closer look at this project by interviewing Michael Westover, Providence’s vice president of population health informatics, and Chris Walker, Humana’s associate vice president of interoperability. 

Healthcare Innovation: Could you guys talk about the history of fragmented systems and manual processes that have hindered payer-provider coordination in the past that this effort seeks to overcome?

Westover: I can talk to Providence's experience. We have armies of people whose job it is to take information out of our EHR and manually copy it over to the payer. We have two individuals whose full-time job is to take financial information off of spreadsheets and then put it in our databases. We have 30 people whose jobs involve just member rosters and claims data from all of these different formats. I think we have over 80 member roster formats — that's just things like first name, last name, which contract you're a part of. We have 80 ways that that comes into the system. 

HCI: How about from your perspective, Chris?

Walker: David, I feel like you're asking a question of why interoperability? And I  think Michael articulated that really well. Data exchange offers a really great opportunity to improve all the inefficiency that Michael just described……As we move to more standards, there's an incredible amount of efficiency that's gained, and it takes the waste out of the healthcare system.

HCI: Could one of you explain what the CMS-0057 rule is, and whether that's lending urgency to these types of efforts?

Westover: The 0057 rule is an interoperability rule that goes into effect in 2027. It covers data exchange between payer and providers. There's a provider access API, for example, and it also involves the process that covers whether a prior authorization is required. I think that a lot of organizations are looking at doing the minimum necessary to comply with 0057 — like, what's the least amount we can do? Humana and Providence are saying if we take the compliance out of it, what do organizations need to do to thrive and to take care of our patients, and what data needs to be exchanged? This is foundational to value-based care and the businesses of our organizations. So I feel like we're going to do the compliance part with 0057 and we're talking about a bunch of other data sources that aren't a part of that, because they're important to us. 

Walker: We’re very much committed to 0057. Those are important topics to get right. But that doesn't mean that an organization can't go above and beyond what those requirements are in the spirit of really pushing the industry toward better data exchange through improved and more common standards.

Westover: A practical point on this topic: we were just on the phone with Humana representatives yesterday talking about the provider access API, which is the clinical and claims information from Humana, and we're talking about whether it has everything we need. Should we add in the supplemental financial data so we're doing more than the minimum necessary? Because it matters to organizations in managing those populations together.

HCI: Did Providence and Humana come together to do this work through the HL7 Da Vinci Project? Or was it separate from that?

Westover: We're both part of Da Vinci. I'm on the steering committee. But because we have value-based care contracts in place, Humana and Providence have a long history and a relationship. We have 30,000 lives in risk populations that we're jointly managing, so we have strong incentives to work together to nail this.

HCI: But was it important to have Da Vinci project implementation guides in place? Does that help the work that you guys are doing together?

Walker: The short answer is absolutely. It's a question of standards, and where Humana has been historically on this topic is that we are very committed to ensuring that the work that we do scales for the industry. So the work that we're doing with Michael is intended to not just be a Humana-specific solution, but it is a solution that's scalable. We need to ensure that when these types of data are exchanged, they can be exchanged not just one time, but it can scale for the system as a whole.

Westover: I think the standards and the modern APIs allow this to be so much bigger than just Providence and Humana. In the previous world we would have to have a different connection with every single payer, and then we had to struggle on data formats. We have these data stand-offs where they say, use our standard, and we say, use our standard. They're really not standards, right? So now we're saying, why don't we all just use the national standard? They're Da Vinci standards, to your point, for member rosters or claims data or gap lists. I feel like that's a much easier conversation that we can come together on, and that will work for other payers that we work with in our in our markets and the vendors we want to share data with.

HCI: Does this involve making changes to how data is pulled from the EHR and other systems prior to exchanging with Humana? 

Westover: The EHRs have these big reporting databases — ours with Epic has 60,000 tables in it. Someone has to go write a query, and that takes a long time; then the data is wrong, and you go back and forth, and then you automate that query, and then two years later, the query breaks; then the person who wrote it doesn't work here anymore, and you have to go figure it out. That’s not the way to manage a business, but that's how data exchange is usually done right now. But because certified EHRs need to expose the data in FHIR, we can pull the data natively in FHIR and filter it to the population and the data that is needed for Humana and then send it to them or other payers. So I think it's more accurate. We've seen a huge increase in the quality and the amount of data when we're pulling it using these standards. It's much better than an analyst writing a query with 27 tables trying to pull the right content.

Walker: I would say it's probably a fairly similar story on our side, but just really in the other direction. We exchange the same types of data, but through different standards. And as a result of that, it produces an incredible amount of work for us to manage. But as we start to move toward an industry norm, everyone's working from the same rubric that we need to on that particular type of data, so it's just really a productive way to work within the organization.

HCI: I read that the first phase was automated member attribution for Humana Medicare Advantage members, and that went live in October. Could you talk about why you picked that capability first? Was that low-hanging fruit or the obvious thing to tackle first?

Westover: When we’re talking about value-based care, where you start is the member population, right? Everything else is based on that. And as I mentioned, it's the one where we have 80 different formats. Once you have the member roster, then we can bump that up against our EHR, and pull the clinical data necessary to send. And then Humana, on their side, can use it to get the right claims data for the correct population. So we have our list of data domains that we're going after, and I think member rosters is the logical starting point that I think everyone should go after. It's wild to me that member rosters are not a part of some of these national data exchanges, too. This is another good reason why having direct collaboration between payer and provider allows us to work through some of these practical challenges and fix them on the ground, and then we can scale up to sharing them with our other payers and then with their other ACO-aligned networks.

HCI: Will future steps be things like automating clinical data exchange, or helping to close quality gaps?

Westover: We have a list of five data domains we've talked about, but we expect this to be an ongoing relationship. We’ve talked about member rosters, and then we've talked about claims data — so the details of what is happening outside of Providence's four walls for the populations we're managing. We're working hard on that. We're also talking about gap lists. Humana is the arbiter of how we're performing on clinical quality measures, so getting that in a standard format, as opposed to multi-tab ugly Excel spreadsheets, right? We've also talked about getting clinical data from Providence to Humana in cleaner ways. We employ Epic Payer Platform with Humana, but we're also experimenting with sending the clinical data through other channels to Humana that has the information they need to close clinical quality measure gaps. We've even talked about the financial performance information. I think it's exciting that we have a large payer and a large provider working through the kinks on some of this stuff and solving some of the problems for everybody else in the industry.

HCI: You guys have talked about scaling this up so that you can use it with other partners. So Michael, are you talking to other payers? Are those organizations close to being ready to work with you in the same way?

Westover: We want everyone to come along for the ride, and we think that it makes sense that they will. Why would a payer want to accept 300 formats from providers? We actually have strong relationships with Premera and Regence in Washington, and they're committed to exchanging these types of data with us. And actually, since we made the announcement with Humana, we've had other large payers reach out to us and say, we want to do this, too. I think what makes Humana special is that they are the thought leader, they're the innovator, and they're national, so that other folks will follow them. If we're just a small regional payer in Northwest Washington, that would be cool, and we could  learn a lot, but that might not move everyone in this direction.

Walker: I think that's right. It is about scale. And to Michael's point, the hope is that there would be interest in from others in growing and using that platform, because that's where standards-based data exchange starts to yield its benefit is when it does scale. And the same is true for providers, as much as payers.

HCI: Anything else you want to add about this work?

Westover: I might say that if folks don't know where to get started on this type of thing, Da Vinci has what it calls Trebuchet for FHIR pilots. It’s a really good way to connect to people who are doing this, who know it inside and out. Or they can reach out to one of us individually. I’d be happy to chat with folks, too. Anything we can do to get people moving in that direction is what I'm interested in.