Enterprise Imaging Roundtable: Goals and Strategy
Most observers would agree that radiology has done an excellent job of managing medical imaging workflow within the department and extending image access to referring physicians wherever they practice. Nonetheless, as the observations and experiences of the four experts who participated in this roundtable demonstrate, a department-centric image management solution is no longer sufficient and many hospitals and health systems are moving to an enterprise imaging strategy.
As hospitals merge, systems grow and electronic medical records (EMRs) achieve widespread implementation, it has become increasingly clear that the walls of healthcare data silos must come tumbling down. All medical image data being produced in every department must be reportable and accessible, preferably from a single interface.
Getting there will not be easy, but all of the experts who participated in this roundtable have begun taking steps to get there. Radiology Business shares the first part of this roundtable discussion in this issue, focusing on goals and strategy. Part II, focusing on the ideal PACS–EHR relationship, will be published in the August/September issue.
Roundtable Participants:
- Louis Lannum, Director of Enterprise Imaging, Cleveland Clinic
- Kevin McEnery, MD, Director of Innovation and Imaging Informatics and Professor of Radiology, University of Texas, MD Anderson Cancer Center
- Rasu Shrestha, MD, Chief Innovation Officer of University of Pittsburgh Medical Center (UPMC), Executive Vice President of UPMC Enterprises
- Chris C. Tomlinson, MBA, CRA, FAHRA, Senior Director of Radiology, Children’s Hospital of Philadelphia, and Executive Director of Radiology Associates of the Children’s Hospital of Philadelphia
For those who have tried to get their arms around enterprise imaging, its scope has been eye opening to say the least. What have you learned about who is imaging and how it is being stored that has surprised you?
Lannum: Today, we’re storing and providing access to images from about 30 different specialties or departments at the Cleveland Clinic. What surprised me the most is how much point-of-care ultrasound is out there, and how many actual departments—orthopedics, rheumatology, ED, anesthesiology, intensivists—are using point-of-care ultrasound that the images never reach our radiology PACS. They’re stored locally. They’re stored in silos inside of each of the departments and they were never shared.
What surprised us the most, I think, was the lack of workflow in each of these departments compared to radiology, where they manage the workflow very, very well, and associate all the images with the EMR records.
What are you seeing in enterprise imaging trends at MD Anderson?
McEnery: At MD Anderson, the big centerpiece is the PACS, in terms of all the diagnostic imaging that’s being done. We’ve attempted over the years to look at integrating other “ologies”, endoscopy and digital light, but the workflow became too onerous.
In radiology, there was a tight coupling, whether it was between the EMR and RIS or the PACS with billing. The imaging being done in non-radiology sites seems not to have that strict linkage, which continues to be a hindrance. Thus far, the strategy has been to create individual archives to meet the individual needs of the business owner, with the intention in the near future to bring them into a single archive as best can be achieved.
I guess what surprised me is that the number of ologies continues to grow. We now have pathology digital images online in our EMR. That has been quite interesting to view. Overall, the demand at our institution is for access to outside images that are being brought in by patients from a variety of different sources.
What has been the experience with enterprise imaging been at UPMC?
Shrestha: Imaging has definitely evolved from being a departmental solution, a very local solution, to now being an enterprise asset at UPMC. We—enterprise imaging or radiology—were the first to embrace digital, but in many ways, we’ve almost been left on the wayside of being an afterthought, at least in things like meaningful use—we were pretty much left off of the party invite list.
In the mad rush to move from analog to digital, the value of radiologists and radiology services seems to have been sidelined, and we’ve been (in many ways) commoditized as well. We’ve gone from being the center of the universe, which is what we were in the days of film and even the early days of PACS, to being a generous mention, if that. That was surprising.
I think it was very much driven by the volume-based approach that has been the definition of how healthcare has been delivered in the last decade or so. As we move toward this value-based paradigm, it need not remain the case. Speaking as a radiologist, looking back at all of the advancements that we’ve made, that shouldn’t be the case. This equates to opportunities in how we need to reinvent ourselves.
What has been seen with a switch to enterprise imaging at CHOP?
Tomlinson: A number of years ago, CHOP recognized that radiology has a lot of experience subject matter expertise in image management and IT has experience supporting large enterprise platforms. In the 2008–2009 timeframe, we went to the institution and suggested that instead of living in a siloed world that never gives clinicians that longitudinal patient view—with a PACS, a C-PACS, something in pathology, something for the ED—that we move toward a vendor-neutral archive (VNA). It saved $3 million dollars over five years versus the traditional PACS configuration.
It also created the imaging infrastructure with the idea of having a central archive and whatever viewer diagnosticians want hanging off of that centralized VNA for their diagnostic work, and one enterprise viewer that links to the EMR—or is capable of producing all imaging records across the institution. We did that early on and built the enterprise imaging governance groups that engaged a lot of the big producers of imaging (radiology, cardiology, pathology) and the big users of imaging (surgical modalities, the ED, urology, orthopedics, neurosurgury, neurology).
We thought we were going to have trouble getting people to the table, that the clinicians who were doing ultrasounds outside of the big diagnostic modalities wouldn’t want to integrate to the enterprise platform. What we found out is that people have big messes—like not archiving from the medical-legal perspective or not having images available to the larger enterprise. As soon as we held up an enterprise platform, everybody came to us and said, “Hey, can you put us in first?”
The scope of enterprise imaging in the past was defined in terms of radiology for the most part and the ability to distribute images throughout the enterprise. Where are we now and what is your institution’s endgame?
McEnery: If we think back 15 to 20 years ago, it was clearly radiology-centric. I think imaging actually helped redefine the enterprise in and of itself. The ability to have images move around a hospital is a given and now institutions want the images to move around the geographic footprint of the institution, whatever that is.
As an example, every one of our radiologists has a home workstation. The idea of working from home, getting access to enterprise imaging, is a given—just a fast internet connection, and you’re a part of the enterprise. The idea that a patient can be seen at one location and then move to another location is better for the patient.
How imaging has changed is actually changing with our enterprises. Our hospitals are getting together and the imaging is one of the glues that holds the fabric of an institution together. In talking to colleagues, in most institutions, that’s one of the key things that is looked at right from the get-go: How do we get the imaging to serve the needs of all the enterprise and even new participants of the enterprise?
I honestly don’t know what our endgame is. The institution continues to grow, continues to redefine its mission, and I think, from an MD Anderson standpoint, our endgame is to effectively reach out to patients wherever they may reside in the country or in the world. I think that is eventually how enterprise imaging will be defined. It will be defined as an institution and not as a geographic footprint.
Where is UPMC headed with enterprise imaging?
Shrestha: The endgame from our institution’s perspective—and, I think, in large part the goal for the industry at large—is value-based imaging. This massive transformation that we’ve seen in the last decade or two in moving from analog to digital, we led that charge. We were early to embrace the move to digital, but what we’ve seen in that process is massive proliferation of data silos.
Even within radiology, we had our own silos, which may or may not have been needed, but that’s how they proliferated. We debated for a long time, whether we should have a RIS-driven workflow or a PACS-driven work flow—it should really be a patient-driven workflow.
Outside of radiology, we’ve got all these other data silos that have proliferated. The key message is that data doesn’t reside in any one silo in terms of what’s most important for the patient. The patient doesn’t live in the silo. The patient’s data lives across multiple different silos, so how we get to the right context becomes a key driver of value-based imaging. It is, we believe, the right thing to do for the patient.
As the payor–provider integrated delivery financial and health care delivery system that UPMC is, that’s the end game for us. When it comes to value-based imaging, we say: enough talk, more action. We’re thinking through the specifics of what appropriateness in utilization really means. How does appropriateness criteria translate to better care? How does it really translate to better care? How do we define value? How we make sure we’re actually able to get to that endgame has been a big initiative here for us at our institution.
Where is CHOP headed with enterprise imaging?
Tomlinson: Whether you call it, univiewer or enterprise viewer, there’s a launch from the EMR that produces images longitudinally across the patient’s record—not just radiology or cardiology but whatever images everybody produces. Clinicians know who performed it by viewing a coding algorithm that shows, for instance, that the ED produced these ultrasounds.
Nonradiologists put in a result. It’s not a full report like a radiologist would create, but anyone who does imaging needs to have a basic order and a result—“I saw this,” or “I wanted to visualize that”—something other clinicians can see so that we’re not duplicating services, to the cost issue raised by Dr. Shrestha.
That’s one component—having everything available to all clinicians at all times, so that we’re not repeating services, especially on these risk-based contracts that we have now. Another is having images available across institutions.
We have a relationship with a hospital in New Jersey where our pediatricians and ED folks staff their NICUs and EDs for the pediatric component. We made sure that we have those images available in our health system as well as in our health system, so that there’s no duplication of imaging or double-radiating being done if that patient comes to our institution.
The endgame continues to evolve, but our goal is that again, in this risk-based world, we’re not duplicating images—even across other institutions that we partner with—or that send us patients. Those imaging studies, wherever they’re done institutionally, or whatever modality or specialty they’re done in, must be seen across that patient record, wherever they may be.
How is Cleveland Clinic defining enterprise imaging?
Lannum: Our enterprise imaging strategy is very much coupled with our EMR strategy. We realized a long time ago, that when a physician is in the EMR, he is still missing 50% to 60% of the data that has been acquired on that patient. A big part of the missing content was images.
As we went down this road with enterprise imaging, our goal was to complete the medical record, or as much of the medical record as we could. Clinical content exists in almost every department—photographs from multiple specialties, clinical ultrasounds, endoscopy procedures, ophthalmology procedures, wound care program. Those images were not accessible.
Today, at the Cleveland Clinic, they are. I can actually go to our EMR launch a patient’s record, and within that record, I can see a longitudinal display of every image that was acquired on that patient during that episode of care. Today, across 35,000 workstations, all of our clinical physicians are accessing images as part of the patient’s medical record. That was important to us.
As we continue to go down this road, we are looking at other opportunities to bring images into that space. We took a VNA and truly made it an enterprise imaging library for medical images.
What was happening before, especially at the point of care, is a lot of the physicians that were doing ultrasounds were actually printing the last image of that ultrasound, scanning it in and sending it to medical records as a scanned document, because they needed proof that they did the scan. They no longer have to do that.
Today, when our anesthesiologists go into the ORs, there’s a workflow. I know that the patient is in that OR, he associates that patient with the ultrasound, he sends it on to our medical imaging library and it is now part of the medical record.
What is the strategy and timetable to achieve enterprise imaging objectives and what is radiology’s role will be in achieving them?
Shrestha: Timing wise, we’re off to the races. Our institution’s primary objective is achieving value-based imaging, which is a lofty goal.
Specific to radiology’s role, we should lead the charge. We need to evolve from being mere diagnosticians to really embracing the role of a physician consultant, which I think is the right model to push value-based imaging to the forefront. This leadership needs to come out of radiology on three different fronts: clinical leadership, IT leadership and executive-level leadership across a healthcare institution.
Clinically, we need to look across the value chain, and that’s what we’re starting to do at UPMC.
It’s not just about us sitting behind that workstation and churning out report after report, the journey starts way up front: The scene of the crime is where studies are being ordered and acquired. We read the exams, report on the results, manage the archives—maybe sharing the studies for more collaboration that may happen—and clinicians view the results. Then, we need to track the outcomes. The entire value chain needs to be defined, and we need to have the right metrics to capture value throughout.
From an IT perspective, DICOM came into being way back in 1983. In imaging, we were quick to embrace DICOM, for good or bad, and it became a sort of de facto standard. The industry at large still struggles with interoperability and standards like SNOMED, RxNorm, LOINC and how to do all of the mapping that the different vendors adhere to. As we look at other ologies, we have an opportunity to say, “Here’s how we did it.” As the VNA evolves from being the single-bucket strategy that it started off as to more of an enterprise content management solution, as the cloud, image-exchange and storage use cases come together, we’re showing leadership.
We need to lead that charge in how to best manage all of the data that exists outside of the EMR— including wave forms, movie files, voice data and genomic data—from an IT perspective, an enterprise perspective. Defining that value, capturing the value, looking at all of the specific metrics, tracking them and then quantifying that value in an amicable way across clinical quality metrics, business growth metrics and service metrics—those are the right ingredients that we need to put into what we’re cooking up here in terms of value-based imaging.
CHOP’s enterprise imaging strategy
Mr. Tomlinson: The strategy we talk about has to do with the longitudinal view. It’s both uncomfortable as well as exciting in that you’re opening up the radiology imaging silo that had previously been so protected by radiology. It makes a lot of sense for radiology to partner with our IT friends on how to do enterprise imaging and not just protect our precious PACS.
How do we turn this thing into an enterprise platform? Think about the subject matter expertise radiology brings to the table around this: we’ve been doing image workflows for years.
Regarding the [EHR]workflow we’ve set up for imaging done outside of radiology and cardiology, we have different flavors, but we have one construct. Folks get to customize their workflow to a certain extent, but everybody uses the same workflow with some different order questions and different configurations. That allows radiology and IT to expand their role and really demonstrate leadership in enterprise imaging—and it’s a very sought-after design.
I have struggled with folks wanting us to solve other problems that maybe aren’t imaging problems. We get into the document management world with pictures and video of patients, not necessarily records, but maybe a patient that’s having some neurological event and the parents take a video. I know Mr. Lannum takes a much more inclusive role—Hey, any content, I’m happy to own it—but some places don’t have the infrastructure [the Cleveland Clinic] has. Some have a separate document management system or Epic’s Haiku and Canto infrastructure to be able to do actual pictures, if you will.
Success also breeds this question: Where do I end in the enterprise imaging strategy and where does document management begin? What belongs in the imaging record versus what’s stored in what they call the blob server in the Epic world, where lot of content is stored. Working on an enterprise governance model can help sort through those issues along with what to do with outside images and patients bringing in [images and documents] on unencrypted USB drives.
Cleveland Clinic's enterprise imaging strategy
Lannum: Governance was extremely important to us. I worked for 15 years as the administrator of radiology information technology, and when we built a VNA, I moved to IT. Enterprise imaging became an IT department.
We created a governance group called the imaging counsel. The imaging counsel had IT, the CIO, the COO, the CMIO, radiology, with the chairman and their PACS people. We also have our distance health chairman at the table, colleagues from cardiology and a number of other different stakeholders as we go into departments.
As we integrate a department into the strategy, the chairman or designate of that department becomes the de facto temporary member of the imaging counsel while we’re doing an implementation. Questions about what goes into document imaging, what images you want to include, what it looks like in the EMR are addressed inside the imaging counsel. Decision-making, direction and vision is given to the IT department, along with our colleagues in radiology, and we built this enterprise strategy.
It all comes back to governance. If you’ve got a governance model, it helps you achieve that goal. You can get radiology involved, you can get every department involved at a level to where this is truly an enterprise initiative.
If we didn’t do that, we would have to talk to each individual department separately, and each one brings a different bias and a different way of doing things to the table. In bringing them together as a counsel, those decisions and those discussions happen, and then all of a sudden a collaborative effort is made to reach our imaging enterprise goals and coordinate that with our EMR goals. I believe that that’s why we were successful in building this strategy, because we have brought to the table the executive leaders necessary to help us execute on the vision.
What is radiology’s role in achieving the MD Anderson enterprise imaging strategy?
McEnery: Radiology’s role in the enterprise will no longer be defined by the images we produce. Imaging availability will be ubiquitous and expected, but there will be a transition to what was described as the value-based proposition. The role of the radiologist and what radiology delivers to the enterprise and the EMR is to increase the clinical efficiency of the entire enterprise.
More and more, radiology is going to be tasked not with how fast it can deliver images to the desktop, but how fast it can deliver information to the desktop to guide the clinical process. I think all ologies that are involved in imaging are going to be tasked in that goal.
Looking back 15 years, a chest/abdomen/pelvis CT scan was 80 images. Today at our institution, it’s over 1,000. Images used to be a centimeter thick; now they’re 0.625, and we’ll do it three different ways. At MD Anderson, we’re talking not so much about how many more images or ologies we can get online, but how can we make those more valuable? How can we integrate them into the EMR?
I think the radiologist needs to be at the table, but needs to be at the table in fulfilling the goals and concerns of clinicians in the enterprise, not as an entity, but as the entirety of a group of individuals trying to bring the best value of care to people.
The timetable and our primary objectives are moving, honestly, away from how much we produce, but how we produce it, and what ways can we make the clinicians more efficient in ordering studies and how the studies can be done more efficiently. The integration with the EMR actually begins to bring out that the radiologist’s workflow is not defined necessarily by PACS anymore, but how does the radiologist workflow positively impact the clinician when they work in the EMR?
Read Part II of this discussion in the article Enterprise Imaging: The Optimal EMR–PACS Relationship.
Vice President, Executive Editor, Radiology Business
Cheryl began her career in journalism when Wite-Out was a relatively new technology. During the past 16 years, she has covered radiology and followed developments in healthcare policy. She holds a BA in History from the University of Delaware and likes nothing better than a good story, well told.