In a new white paper from the Falls Church, Va.-based CSC, “Transforming Healthcare through Better Use of Data,” CSC experts argue strongly for the leaders of patient care organizations to move forward to establish the strong “foundational practices and capabilities” needed to make the most out of the data they already have and to meet the many requirements of healthcare reform mandates, payer demands, and meaningful use.
Among other areas, the white paper covers the topics of data governance, data acquisition, data sharing, data standardization, data integration, and data analytics. The paper’s authors urge healthcare leaders to think very clearly about process before plunging headlong into big data.
Jared Rhoads, who co-authored the report along with Lynette Ferrara, and his colleague Dan Foltz, spoke recently with HCI Editor-in-Chief Mark Hagland regarding the findings in the white paper, and their implications for healthcare IT leaders. The Philadelphia-based Foltz leads CSC’s Health Intelligence Division, while the Rhoads works in the company's Waltham, Mass.-based Global Institute for Emerging Practices. Below are excerpts from that interview with Foltz and Rhoads.
Organizations are confused and overwhelmed by data right now. What do you see as the key foundational steps?
Dan Foltz: From a governance perspective, there are organizational and process steps that have to be taken first. Organizations need to start managing their data as a strategic asset, just as they would manage human resources or real estate or other assets. And there are countless areas where that’s not occurring, and it involves defining data standards, processes, quality. I was at an organization last week with hundreds of systems, and they have the basic problem of identifying who their physicians are. Or you want to develop reporting on visit-related metrics? I was visiting another hospital where they had seven different definitions of what a visit was. In and of themselves, these seem like trivial things, but they all compound problems, and make it very hard to trust—or even produce—the data and analysis to begin with.
Is some of this project management, and some of it data management?
Foltz: We think what’s important is this notion of data stewardship. If you look across possible data domains, if you break it down into the types of data in a hospital—patient data, which is further broken into demographics, billing, laboratory, and then physician data, and so on—well, you need data stewardship—people who are necessary to pull together the different data and establish definitions and standards, and to develop standards for quality.
How about the problem that some of the stakeholder groups in patient care organizations, such as physicians, have historically had some of the least profound involvement with data management or “stewardship”?
There’s no single answer, but what we would say is that department chairs, for example, are the least likely people to get involved in deeper data discussions, even as they are held most responsible for outcomes in their departments. So there’s a culture of ownership that has to be created; and there are different people who would serve as stewards of that data. Then you have to come up with different data stewardship processes, to discuss who’s responsible for the quality of the data, to start to set expectations for what has to be done. And that’s irrespective of the information systems themselves.
What you’ll see is that organizations are trying to build data warehouses very quickly, for example, but no one has reached any consensus on what it will do or include.
So what are the proper roles of CIOs, CMIOs, and other informaticists, in all this?
It’s a complex question; and you can’t get there through a single step. So we recommend incremental steps. External drivers like meaningful use will obviously set some of your priorities for you; and then you map those requirements back to what types of data will be required. And what use models will be considered important? For example, infection prevention and control: go into most hospitals, and follow the data through their infection control processes, and it involves a lot of manpower or labor going into mapping processes or doing root-cause analysis. So you can very quickly map something out and say, here’s what we could do together, and you’ll be able to do it faster and more accurately. So the CIOs, if they can facilitate those discussions and that priority-setting, that gives them the language and the tools to hold more meaningful discussions with clinical leaders like CMOs, to discuss what they hope to get out of the process.
And this is culture change and change management 101, right? We had a process recently where we involved over 100 people, and involved 50 different use models, everything from infection control management to emergency room flow management, to producing better outcomes for different clinical conditions, such as otitis media or the overuse of antibiotics, or case studies for clinical research. So you build this long list, and you prioritize it. And that’s a process that the CIOs can deploy teams to facility the activity for, and which can move you forward in those areas.
Jared Rhoads: The asset we’re talking about is data, but we’re talking so much about people and processes, and it’s clear we’re talking about getting the right people together talking; and getting the right skill sets involved is so important, too. So for example, you need people very comfortable with databases who are also comfortable with statistics; and of course, any nurse or clinician comfortable with informatics is that much more valuable, so those skills are so needed.
So what happens to traditional roles like CIO and CMIO?
Foltz: I’m very active in the American Medical Informatics Association, and I was speaking with a leader there who just told me last week, the time has come for the emergence of the CMIO as a true leader. And I’ve talked to some CMIOs who are driving very large budgets, versus some who have a staff of two; that is changing, and a greater recognition of the importance of that role is coming. And I remember talking to someone who runs the lab system in one clinical department, and she was conveying to me that over the years, the demands for access to data from the lab system have skyrocketed over the past several years. Yet nothing in her job description involves her fulfilling those demands. So she’s acting as a good steward, but she’s working extra hours now to fulfill those demands. And those kinds of things tend to be ignored by senior management until they become a problem.
How do you pull all the different skill sets and people needed to do all this?
There’s no one answer, but what some organizations are starting to move toward is establishing some sort of center of excellence—perhaps highly decentralized or virtualized. But this all starts with setting up governance, doing a review of all the data sources, of all the data standards, developing some sort of IT strategy; and then we recommend step-by-step, incremental processes. The days of the big-bang data warehouse implementation are over.
But will people have the time to do it that way?
Well, meaningful use will drive this forward; I’m sure this is creating a dilemma for a lot of organizations.
Rhoads: Dan really hit the right note in terms of the concept of centers of excellence, as we described in the white paper. You get a few champions from a few cherry-picked areas, and you get them to set up a special competency center, for example. But it’s tough, there are a lot of competing priorities.
What can the members of an organization’s c-suite do?
Foltz: They need to be visible sponsors, and to put forward a vision, a stated intent, around data, and talk about how improving data and data processes will improve patient care and efficiency, and create a plan to make things happen. Most patient care organizations put out a yearly list of priorities, and things that don’t make that list tend not to be taken care of; so this needs to be up there. And it can be wrapped in under the umbrella of meaningful use, of course.
Rhoads: They can set the tone to follow through, and not just jump on something that might be exciting, but to do it in accordance with the policies that have been set up. That will avoid going too far down a road and then having things change and you end up with a wasted effort.
Foltz: They have to be circumspect in investing in things that will create sustained effort. And oftentimes, when things end up just happening in one department, they end of getting buried.
Rhoads: And you can coordinate such efforts.
Is there anything you’d like to add?
Foltz: I would add one other dimension. We’ve been talking about challenges, but there’s excitement involved, too. We’re moving into an age of greater collaboration, with health information exchange, accountable care organizations, and medical home models—these are all great things, but the data needs to be high-quality, needs to be standardized. I’ve run roundtables at academic medical centers on the topic of these research networks for comparative research networks, and they’re popping up all over the place, and yet they present new demands on academic medical centers. And those things are a great opportunity, but also a great challenge, because we need to be heading towards a more standardized approach to all this.
Rhoads: And just to build on that, a lot of organizations have great data assets right now that they need to learn how to deal better with; but there are a lot of newer sources of data coming down the pike—you’ve heard of the concept of big data—and you’ll have to prepare yourself, because very shortly, there’s going to be a flood of even more data, and organizations need to be even more ready for that.
Foltz: Think of all the diabetics out there checking their blood sugars every day—how do you turn that into useful data, instead of just a flood of data flowing everywhere? And then there’s all the genomic data that will be coming. I think we’re at the point where getting full genetic will become much more practicable. And there’s the full genetic data, which is huge, but then there are also just the genetic signatures, with the important information about it; but where all the data sits is another matter.