Intermountain Healthcare Exec Discusses Fine-Tuning Data Supply Chain Strategy
The term supply chain usually refers to production and distribution of a commodity. But some health systems are beginning to think of their own internal data assets in terms of supply chain.
“If we think about the job of an effective supply chain, it's to make sure that we get the right supply in the hands of the right person at the right time, place and cost,” said David Dirks, vice president of strategy for Utah-based Intermountain Healthcare. “We are thinking about our data assets in those terms.”
Just as it would with a manufacturing supply chain, Intermountain is seeking to strike the balance between something that is high quality, but isn't overbuilt. “We don't want expensive or unnecessary supplies that aren't justified and just complicate the process,” Dirks said. “In terms of the right person, we want to make sure that people are practicing at the top of their license — that highly skilled employees are focused on those tasks that only they can do.”
Dirks said that many health systems are focusing on data aggregation — getting as much data as possible, to spin up data lakes. They are thinking about massive amounts of data. “We tried that for a while, and that's great if you're an analyst or you really care about data architecture. But we found that for front-line workers, there's no way that they can absorb all that or utilize all of that data,” he said. “We've been focused on the data that we actually need to care for our patients. So rather than an approach where we onboard as much as we can, we're very focused on making sure that the right person has the data they need to carry out the job. We want to make sure that we're customizing the data that we surface so that it's meaningful, and that the person can act on it. We're aligning those data elements and allowing people to practice at the top of their license.”
For instance, value-based care and population health often relies on a claims data feed. “That data is very dated, and it's looking in the rearview mirror,” Dirks explained. Physicians want more real-time data that allows them to make interventions as they practice.
Another hot topic at most health systems is embedding data and solutions in the clinical workflow. “We actually think that can be problematic because we want a physician to be able to look a patient in the eye, to listen, respond to their needs, so putting a bunch of data or tasks within a physician workflow — to some extent, we've backed away from that,” Dirks said. “Instead, we look for what data we can tee up at the beginning of a provider's day and simplify what we're asking them to know about a patient — and do that in the right place within the structure of their day rather than just throwing a bunch of data and tasks into their workflow.”
Intermountain tries to offload some tasks to a different set of resources. It has a function called Care Traffic Control. Its employees are nonclinical but highly trained on all of the data that the organization has to be able to resolve issues, clean the data, and make sure that the clinician and the downstream caregivers have the information they need, so that the amount that they have to do during the visit is as little as possible. “Part of that is making sure that you have that full data supply chain running so that everyone is looking at the same data, and they're all communicating the same way.”
I asked Dirks whether a strong data governance process was key to this effort.
“Data governance is job number one,” he responded. “We have a clinical integration committee made up of not only Intermountain physicians, but also affiliated physicians, who make the big decisions around data governance. In the area of value-based care, the idea of attribution drives so much of the data that's being shown to those physicians. You've got to have a group that is responsible for helping to make some of those key decisions.”
The supply chain metaphor works because the whole downstream process breaks down if the health system is not introducing good quality raw materials into the process, Dirks said. “The second that a physician doesn't trust the data, doesn't trust the information, everything else we're trying to do on their behalf doesn't make sense,” he adds.
As Intermountain works with other health systems and with cloud storage providers and analytics companies, the data supply chain gets complicated, he admitted.
“When you take on a population health approach, and you're bringing on a bunch of data sources outside of your enterprise, you have to be very careful about who can see the data, how and where you can store it and surface it,” Dirks said. “You no longer can just say we're going to put all of this information in one single data warehouse. I've got to have certain teams that can only see this data, and other teams that can only see that data. And so the level of complexity increases pretty significantly.”
Besides being vice president of strategy at Intermountain, Dirks is also serving temporarily as interim CEO of Intermountain subsidiary Castell, which has a focus on value-based payment analytics. Castell is the first company to grow out of a large integrated delivery system that is working in a meaningful way with affiliated physicians and other partners that aren't part of the delivery system and in a multi-payer environment, Dirks noted.
Castell also has developed a close working relationship with Arcadia, a healthcare data and software company focused on value-based care. “One of the things we identified as Castell sought to enable population health is that there really was no single comprehensive population health vendor,” Dirks said. “We knew that we'd have to do a lot of development on our own and bring other tools into the environment to round out our capabilities. The main reason we selected Arcadia was its track record, at scale, of being able to take in multiple data sources, from payers, clinical, social determinants, and integrate them, normalize them, and clean them. And they are agile and nimble in helping us to build out the toolset that we needed to support our work.”