Industry observers agree: the leveraging of data analytics is in an early and frankly, rather challenging, phase in U.S. healthcare. The leaders of patient care organizations that are plunging into the development of accountable care organizations (ACOs), other risk-based types of contracting, and every form of population health and care management, are beginning to work through myriad obstacles to try to achieve success.
And success in such contracting and population-based work inevitably requires excellent data analytics. But, as patient care leaders are learning, marrying clinical data, including from electronic records (EHRs), with claims-based data, is turning out to be far more complex and challenging than many imagined.
Not surprisingly, patient care leaders are turning en masse to expert consultants to help them sort through the issues and plow ahead. Among the legion of consultants working with healthcare leaders is Mary Jo Morrison, a principal in The Chartis Group, a Chicago-based consulting firm. Morrison, who is based in Minneapolis, has spent more than two decades participating in analytics work, operational improvement, and strategic improvement, in patient care organizations. Most recently, Morrison served as vice president of performance measures at Allina Health in Minneapolis. In that role, she identified opportunities for clinical and operational improvement through data analytics and data mining. She joined The Chartis Group in June.
Recently, HCI Editor-in-Chief Mark Hagland spoke with Morrison about the current state of data analytics in healthcare. Below are excerpts from that interview.
When you look at the current landscape, how do you see the leveraging of data analytics right now?
To be honest, it’s all over the board. The maturity of organizations varies so much in this area. Some organizations are just now beginning to get into electronic health records; they haven’t leveraged the claims data, and haven’t been able to fully leverage the data in the EHR, and they’ve been overly reliant on vendor-based portals, using data that they’re being fed, and having to retrofit that data to their needs. On the advanced end, there are patient care organizations moving forward into significant forays into data science. And that is a much smaller group of organizations. It’s just across the board, across the spectrum. And even in those organizations with sophisticated data scientists and data structure, I see a lack of data governance. But the willingness to forge ahead is exciting. That is encouraging.
What are the biggest challenges facing patient care organizations going into accountable care and population health at this moment in U.S. healthcare?
The biggest challenges are not so much on the technical side. Instead, the challenges are around process and culture. There is a real organizational struggle going on. There is such a plethora of information available, and such a dynamic marketplace around analytics, that people just need to sit down and reflect on what they need, and what they want to do with this data. And they need to ask whether they have a culture that will use this data. So it’s not necessarily the mechanics; it’s that softer side, the human aspect, where the challenges are.
One problem seems to be this: every single patient care leader I’ve interviewed has said there’s no off-the-shelf analytics software that is sufficient to get providers to where they need to get to—that they need to do quite extensive customization in every case.
I think it’s the way people are working with the EHR data. We’re working with an organization right now that wants to reduce their length of stay and see if the ability to reduce LOS will have an adverse impact on readmissions. In that regard, I don’t think the EHR is necessarily too elementary; it’s a wealth of information. The problem speaks to what organizations are doing with the data, applying some statistics or higher-level analysis to it. The EHR is a wonderful tool, but you have to apply some analytical rigor above and beyond it.
Do you agree that there is a broad lack of data scientists and analysts out there right now in the U.S. healthcare industry? Most organizations don’t have the people capable of doing what you mentioned above, and are faced with either trying to groom healthcare professionals, sometimes clinicians, to do analytics work, or with the prospect of bringing data analysts and data scientists into healthcare from other industries. What are your perceptions of that dilemma?
I would agree that that is a real challenge right now. There is a very significant need for people with analytics and statistical capability. But I wouldn’t let that limit leaders to what they’re doing in an organization. Because based on one analytics project, you may see five or six performance improvement initiatives coming out of that. So I would say that you need to take a focused view and ask the right questions that are clinically, operationally, or financially important. Applying good rigor to your analytics around a question or problem, you can get good downstream impact. And, too, you want to make sure that you’re not throwing too much data at people that they can’t assimilate it. So as we go out and work with organizations, what we say is, prepare well for the future. The organizations that are doing it well are forward-thinking and are creating capacity, and using data well will really help you prepare for the next stage. The clarity and focus are an oasis in analytics right now.
So you’re saying, fundamentally, that you need to decide what you want to achieve, correct?
Exactly; decide what you want to achieve, and then build forward a governance structure that will help prioritize, because the decisions made at the top will have a ripple effect, and governance is critical in that. And a little pragmatism and discipline helps, too, because there’s just too much variability in the process.
And even pioneers are saying that they’re learning as they go, per population health and ACO development. We really are still in a very early phase in general, right?
Yes, I would agree, it’s still very new. And if you take that learn-as-you-go mentality, the value of that iterative cycle cannot be overestimated. It’s a real opportunity for organizational development and change.
What should CIOs and CMIOs know about all this right now?
Well, analytics and improvement are a team sport. And governance is a process that requires people to sit down and develop priorities and responsibilities together, and get into change management. And many times, the governance processes begin in IS, and those responsibilities fall to IS. Governance will play a key role in this, and governance that brings in good leadership, is important. You need the CEO, CFO, COO, CMO, CIO, CMIO of your organization involved—you need all of those folks at the table, and that will go a long way.
So the c-suite has to align on priorities and focus?
Yes, absolutely, that’s critical. Especially in the context that value-based care is changing so rapidly.
What will be happening in this area in the next couple of years?
I think that organizations without a doubt will figure out that you really need to take a look at the world from three views. One is that the integrity, the management, and the governance of the data, will be critical. Add to that your analytics capabilities, developed from within and using outside resources. And finally, moving into that data science space. And without a doubt, the market will push organizations to that, if they’re not proactive. But I believe that organizations will figure it out. But going back to that focus and framework, will be very good. If organizations build good data analytics capability, they’ll be well-positioned. Organizations have the data that they need: electronic health record- and claims-based data are fabulous; you just need the focus, the governance and the strategy.