A lively discussion opened the Health IT Summit in San Diego, sponsored by the Institute for Health Technology Transformation (iHT2—a sister organization to Healthcare Informatics under the Vendome Group corporate umbrella), and being held this week at the Hilton Bayfront in downtown San Diego. On Monday morning, Jan. 19, Shadaab Kanwal, executive director, corporate services, and digital technologies, at the Oakland, Calif.-based Kaiser Permanente, led a panel of industry leaders around the topic, “Data & Analytic Strategies for Value-Based Care.
Kanwal’s fellow discussants were Carol Steltenkamp, M.D., CMIO at the University of Kentucky health system; Brett MacLaren, vice president of enterprise analytics at the San Diego-based Sharp healthcare; and Alex Eastman, senior director enterprise solutions, at the Premier, Inc. health alliance (Charlotte, N.C.).
Early in the discussion, Kanwal noted that “A lot of organizations are focusing on the delivery models, and there is a lot of activity going on. There is such a change from a care delivery model, so the perception of value is shifting from the provider space into other areas of care delivery.” He asked his fellow discussants, “In that context, what does value mean to each of you?”
“In terms of value, terms of how value ties to technology,” MacLaren said, “we need to think about how we can take differing incentives” and weave them together for greater alignment of all the stakeholders. “We have health plans, medical groups, and hospitals, and as we’re going through this process of transformation, not all the incentives are aligned. So we need to make sure that quality of care ties to reimbursement and everything else. I think there’s still a bit of misalignment, but we’re moving in the right direction.
(.to r.) Panelists Kanwal, Steltenkamp, MacLaren, and Eastman
Dr. Steltenkamp stated that “In terms of value, I always get back to quality over cost. Did you as a patient feel you had a quality interaction? Did you as a provider, feel that? Oftentimes, providers may be a little bit lost” when it comes to really connecting all the clinical documentation they’re required to do, with enhanced value for patients. At Premier, Inc., Eastman said, “We focus on doing data that makes it easy for people to analyze it. And I think of value as a ratio of benefits to costs, as Carol hinted at. So If I bought a car, I feel like I got a good value, with the features I wanted. The lowest cost doesn’t necessarily bring the best value.”
“How have you begun to collect, aggregate, and analyze data for value-based care in your organization?” Kanwal asked the panel. “It starts with the concepts of enterprise data management and data governance,” MacLaren emphasized. “We need to understand what we’re trying to measure, and close this loop of value. We expect our clinicians to document things, but they don’t always recognize the value. They often see documenting these things as a “necessary evil,” when in fact, it’s important to documenting value. So part of the challenge is making sure that we close that loop of value, and that we make it easy for those doing so to do so easily. Think of all the advances in natural language processing. For a long time, we tried to force all clinicians to structure, structure, and structure. Well, maybe that’s not the most efficient way to do it. I would hope that our technologies are going to help us move forward in that area. And I don’t think that our EMRs have necessarily done us a lot of credit in helping us to do that.”
“I would agree,” Steltenkamp said. “Intellectually, we know we need to do that, but sometimes, people haven’t taken responsibility for doing so. And I would say the clinicians are still in a position where they most often don’t see the benefit of what they’re documenting. They get that they need to document to be paid. And the bulk of clinicians do this because they want to help patients. And so they will appreciate the feedback or input that helps them do things. Because if you can help connect those dots for them around their helping their patients, that’s a win. Because if you show up and say, I’m here from the administration, and I’m here to help—that’s not a win.”
“So how can we physically make that happen?” Kanwal asked. “How do we ensure we have the clinical and business intelligence, and analytics, so we have focused, relevant information enabled by an agile, relevant, and able infrastructure, to help physicians to capture information? And that enables us to be equipped to deliver the care effectively across the enterprise and ultimately, across the community? And from a process perspective, what specific interventions did you have to make?”
“The University of Kentucky had an RFD out to bid on analytics,” Steltenkamp noted. “So we are in the process of building a data warehouse. I’m sure other organizations are much further down that path than we are. But like any decent university, we thought, we could do this ourselves! But after about three years of some really hard work, there was a self-realization that that was not going to happen. So we went out and said, we needed something appropriately scalable, to drive towards business analytics for healthcare, but also, to accommodate all the research that we’re doing.”
“The reality,” Eastman said, “is that EHRs weren’t built for analytics; they were built to capture information and to present clinical information about a patient at a point of care, and to help with the billing, the collection of revenue. That’s not unique to healthcare. I check into a hotel, and the application they’re using isn’t built for analytics. So there’s a need to take data out of these transactional systems, and organize it, for analytics. That’s the point of a data warehouse.”
“So it’s a journey, then, right?” Kanwal said. “We started with business intelligence. Some are still in the building phases of that. Then, business intelligence getting to the prescriptive phase; and we’re all intending to get into the predictive phase. Obviously, it was built with silos not designed to optimize a data warehouse architecture. And you’re relying on your mission-critical systems to feed it into those systems. And a lot of your internal groups are challenged, they have to do a lot of segmentation of data architectures, because, for example, research needs are different from patient care needs. So, in terms of the strategies for value-based care, what was the best practice you adopted for people, process, or technology?”
“From a people-based standpoint, it really is a challenge to get the right people in place to actually do the analytics,” Steltenkamp emphasized. “It just feels like there are not enough of those folks out there to being that. And it’s not the math department at the university or the statistics department; it’s almost the math department crossed with the philosophy department! So from a people standpoint, it’s a challenge.”
“What’s hard to find is the unicorn of the true data scientist,” MacLaren said. “So what I look for is critical thinking skills, a thirst for knowledge, and innate curiosity; and you’ve got to understand data. But a lot of that can be taught, a lot of the healthcare can be taught, as long as someone understands the clinical context. I think it’s when we have these pockets of analysts out there, being hired by someone who doesn’t understand the right concepts or technology, so those people have to learn through Goggle. So the people part is about networking and leveraging the best possible tools and technologies to enable what they were hired to do. From the process standpoint, there’s a lot of work to be done yet. The technology is actually the furthest along. Meanwhile, in terms of connecting people and process, it comes down to this process of data governance and data management, creating shared concepts around what we’re trying to do. So we’re talking about this length of stay metric, maybe there are multiple definitions involved, but we define what it is we’re doing, what we’re looking for.”
Later in the discussion, panelists turned to the question of the extent to which the leveraging of analytics can promote behavioral change. “Data can help initiate change,” Eastman insisted. “When we can show facts, that obviously helps get people on board, but it’s not necessarily enough. And if I show a report to a physician saying that your cost per patient is higher than others’, you know that he’ll immediately challenge the data. And so, per data governance, you have to e able to explain where the data came from, what definitions you’re using, and how you derived your analysis. So that data lineage is very important.”
MacLaren offered that “I’m actually really excited for population itself. The concept itself is actually a galvanizing force for bringing all these silos of operations together under a common purpose or umbrella. And Sharp has done a lot of these things within various groups, like our medical groups and hospitals, but at the same time, and we’re still in the early phases of this, but seeing the movement of everybody starting to work together under the same umbrella, thinking about how we really engage folks to improve the care to the entire community.”