How can the leveraging of data analytics help transform patient care delivery, at a time of intensive accountable care organization (ACO) development in U.S. healthcare? Healthcare leaders delved into some of the challenges and opportunities facing patient care organizations, during the opening panel at the Health IT Summit in Seattle on Aug. 18, being held at the Seattle Marriott Waterfront, and sponsored by the Institute for Health Technology Transformation (iHT2, a sister organization to Healthcare Informatics, under the corporate umbrella of parent organization Vendome Group, LLC).
The panel, entitled “Analytics Driving Accountable Care,” was moderated by Richard Gibson, M.D., Ph.D., affiliate assistant professor in the Department of Medical Informatics and Clinical Epidemiology at Oregon Health and Science University, and a well-known healthcare IT leader. Dr. Gibson’s fellow discussants were Jeff Hummel, M.D., M.P.H., medical director for Healthcare Informatics at Qualis Health, a quality improvement organization and consulting firm in Washington state; Kirk Larson, regional chief information officer for NetApp and chairman of the board of the Central Valley (California) Health Information Exchange, and Michael Simpson, CEO of Caradigm.
Panelists (l. to r.) Gibson, Hummel, Larson, and Simpson
The discussion ranged widely, from the subject of engaging primary care physicians in analytics work around population health in accountable care settings, to some of the broader social and societal barriers to improving the health of populations.
Early on in the discussion, Dr. Gibson asked Dr. Hummel, “Dr. Hummel, what do primary care providers need to know about analytics?”
“The biggest challenge is the paradigm shift from one-at-a-time patient care to population health,” Hummel said. “Everybody in healthcare has really spent their entire career in fee-for-service [care delivery], taking care of patients one at a time. So just thinking about the concepts of population health is difficult for physicians. When you start coming down into the working ranks of providers, they’re very good, smart people, and good at data, but are always looking to solve a specific problem of, what I do in the moment, with this patient in front of me? So, looking at broad populations is pretty foreign. So small and medium-sized practices have a challenge.”
In fact, Hummel said, “The technology that people need is pretty simple. The problem is not so much analytics, but it’s really presenting the data in a way that care teams can understand it. How many patients do you have over time with hypertension, or with hypertension that’s not under control? And if you’re screening people for a particular issue—what’s the population you’re looking at—diabetes? And what are you screening them for? So the question is, what did we find, and for those for whom we found something, what did we do? And if you can show run charts that show providers how they’re doing around those parameters, they can start to respond. But even putting information into an organized framework they can understand, is a challenge. So I think that primary care providers need help with that information and using it.”
Gibson turned to Larson, asking him, “Kirk, what needs to happen with making the necessary investments in population health tools?”
“It’s good question,” Larson said, “because it’s not always abundantly clear. It’s not something where you can necessarily run an ROI on to determine what return on investment you’ll get from it in 18 months, for example. When I was at Children’s Hospital,” he said, “when patients would present in the ER, the data was a month old, or four months old. I have a three-year-old son with a chronic illness, and I think of him. And think of the difference between data that’s a month or a few months old, versus data that’s real-time, and what a difference it will make for patients like my son, in terms of how effective the data will be for clinicians at the point of care.” But, he said in order to get real return on any investment in analytics solutions, “It’s going to require clinicians being champions for analytics, to get this to happen, based on advocacy for quality of care, based on what a difference we can make if we provide the technology. I think that will be persuasive to physicians and clinicians.”
One of the challenges in making the shift from fee-for-service assumptions to value-based payment care delivery, said Simpson, is that “It’s not going to be about retrospective analytics, though they are valuable. But the people who cost you money last year are not the same people who will cost you money this year. Of my million patients, which 500 diabetics will hit the ER?” he asked rhetorically. “By predicting those things, you as a provider organization can take a proactive step to intervene. Providers understand that retrospective analytics is necessary and important, but you have to move towards predictive analytics. And to do that, you have to take data from the entire continuum of healthcare, whether from the retail pharmacy or outpatient clinic, or wherever, and then you need to drive that towards actual action. So organizations are beginning to focus on how they can get that data, and how they can use it.”
How much do primary care physicians and nurses actually need to know about the details of analytics programs? In response to that question from Gibson, Hummel said, “We’re in the middle of a cultural shift here, so the answer depends. I think your point is well-taken that we need to be looking forward, to act to prevent unnecessary utilization. And yet the challenge is that primary care isn’t really structured to do that. We still are working in pretty much physician-centric practices, rather than a working with a high-performance team model in which everyone on the team shares the care.”
In fact, Hummel said, there are deeper issues involved. “Often,” he emphasized, “the social disorganization, mixed with a complex burden of multiple chronic diseases, is the problem, for the high utilizers. They’re not taking all their meds, they’re not even sure what they do, and they may not even have a place to stay tonight. And you may not need just another nurse, but a social worker and a behavioral health team, that can address those interventions. But the amount of social intervention required to create change in that regard is of a magnitude compared to what we can do now. And the focus on volume” in today’s medical practices, “on getting patients in and out as fast as possible, is a barrier to progress, too. The technology out there is amazing, but it’s the ability of the delivery system to respond in a meaningful way.”
Who should the “ideal champion” to advocate for analytics implementation and use in an organization be? In response to that question from Gibson, Larson said, “I’ll start by saying who the ideal person is not: it’s not OK. And that is not a criticism of IT; I’m not saying that people in IT are not capable. But because this is such a far-reaching thing, the best people really are the clinicians. Because when you strip down the question to, who do you want to be the champion of analytics, or whatever? You really want the people who will be using the system—physicians and nurses, etc.—to be the champions. They’re the ones who will use it in their workflow. And when you get their input really from the ground level, people who are there from the ground level saying, we have this need—that’s important. So the people best able to advocate for these systems are the people who use it every day, and that’s the clinicians.”
In response to an audience question about how important patient empowerment is to the success of analytics-driven population health efforts, Hummel said, “I interpret that question as patients’ taking on of engagement in the care process, and the care team’s support of that. Remember, this is relatively new. I began practice working with a physician who had been practicing since World War II. And what we’re seeing is an understanding that patient engagement and self-management is absolutely key, particularly for many of the chronic illnesses. The ability to reduce the cost of the morbidity over long periods of time—lifestyle and health changes on the part of the patients are incredibly important. But change is slow.” But, he added, “When patients themselves walk in and say, it’s time for my hemoglobin a1c, and it’s time for my eye exam, and the more we can get patients to see that it’s about them, the more we’ll see change happening.
“One factor working in our favor is the expectations of the younger generations,” Larson noted. “The reality is that the kids who are 3, 6, 10, they can unlock your cell phone right now, right? And on Facebook, each day, people post tens of millions of comments about their personal health. This is normal for the younger generations. And we’d darn well better be ready for this: this is what the younger generations are living in, and this is their world. And we need to address this.”
Acknowledging the social determinants of health
Speaking to broader issues, Gibson said that “I’m a big fan of analytics. My big concern is that a lot of patients don’t have what they need to really manage their care. In a middle- and upper-middle-class environment, we assume that people know what good diet is or what good sleep is. In Europe, they have a lot lower medical spend but a lot higher social spend. And even though people know they should eat better and not smoke, they need help understanding how to deal with that.”
“This is a direct effect,” Hummel noted, “of such phenomena as food deserts and income equality, and we’ve got to change that” as a broader society.
What’s more, Larson added, “In my work with Central Valley Health Information Exchange, it’s worth noting that it covers a broad swath of rural California. And once you get outside some of the metro areas, English isn’t spoken at all, people don’t have computers or broadband or even Internet access; but what everyone does have is smartphones. So we as a society need to think through some of those societal issues. Meanwhile, for us as an HIE, as we thought about how to give patients access who may or may not speak English or have Internet access, we fell back on the realization that they do have mobile phones. So what are we going to do as an HIE to be successful, in identifying pockets of opportunity. People are shocked when I give that example because California is the most populous state in the country, but the reality of the diversity of the state is different from how people imagine it.” What is key, he added, is to address patient needs and assets as they actually are, and to leverage technologies, such as mobile devices, that they have the access and ability to use.
“We do need to address the root causes of poor health in our society,” Gibson stressed. “That said, we are spending $3.1 trillion a year on healthcare so it’s incumbent on providers to deliver care that is cost-effective. Now, with precision medicine, we can attune individual interventions to individual patients, while trying to standardize treatment. So I’m not discouraged that we’ve spent this much on electronic health records; we need to be able to track outcomes, and at the same time, we need to focus on spending effectively on interventions that make a difference.”
Fortunately, Simpson noted, “Where we are today with information technology is a world away from where we were ten years ago. And I really think we’re poised to be able to make some major changes. The technology will help us identify gaps in care, and we’ll need to address those gaps that we didn’t even know we had before.”