The Health-system Challenge: Strategic Deployment of Imaging IT

As health systems prepare for value-based health care, CIOs walk a tightrope between consolidating enterprise IT and providing the functionality that radiologists require

One of the first commercial image-management systems, in its advertising slogan, promoted its ability to deliver any image, anywhere, anytime—throughout the health-care enterprise. That continues to represent the primary objective of imaging informatics: to increase the accessibility and availability of diagnostic imaging for clinical use. 

The RIS and PACS, however, are no longer at the center of this universe. Electronic medical record (EMR) systems are assuming this role, consolidating department-based information systems and providing a single portal to patient records. 

Centralized archives are replacing specialized archive silos. Hospital enterprises are investing millions of dollars to consolidate their IT systems and make them as productive as possible: as interoperable as possible (to exchange information transparently) and as cost effective as possible to operate. 

How are radiology IT departments being affected by changes in a world where the EHR is dominant and the HIE is emerging? Radiology Business Journal discussed this topic with two CIOs and two chairs of imaging informatics. 

They represent a coast-to-coast mix of integrated delivery networks (IDNs). In addition, they understand (and stand on) the cutting edge of health-care IT adoption.

Preload: Preview
  • As health systems rationalize IT infrastructure, individual departments—including radiology—are feeling the repercussions.
  • Imaging IT personnel have tighter ties to enterprise information departments (if not direct reporting), providing more complete coverage of the imaging chain.
  • Departmental information systems, including PACS and RIS, are being disassembled—and in some cases, replaced by enterprise electronic health records (EHRs).
  • Both enterprise and imaging IT experts anticipate that health information exchanges (HIEs) will be used to solve their image-exchange problems—soon.

NorthShore University HealthSystem 

Greater Chicago, Illinois, has one of the most wired clusters of hospital enterprises in the nation. As CIO of NorthShore University HealthSystem (Evanston, Illinois), Steve Smith oversees the IT activities of one of these. This IDN includes four hospitals in the northern suburbs, and its multispecialty NorthShore Medical Group has more than 100 office locations. 

NorthShore implemented an EMR across its entire system in 2003. All of its hospitals and clinics are fully electronic and fully integrated. It has attained stage 7, the highest level of the HIMSS Analytics® Electronic Medical Record Adoption Model (EMRAM), achieved by only 2.9% of hospitals in the United States.

The work doesn’t stop, though, for this IT–expert IDN. During the past five years, the IT department has continued to enhance functionality and improve interoperability. Smith says, “We have been building an enterprise data warehouse because use of data from our EMR is the key to improving the quality of patient care and patient safety.” He points out that the growth of accountable-care organizations (ACOs) requires data input (ideally) from all possible sources—so that the data can be analyzed and used meaningfully, and so that best-practice patterns of care and treatment can be developed from them.

“With the movement toward health-care reform and meaningful use, as well as the eventual proliferation of HIEs during this decade, the ability to manage patients and data to deliver total population health, across the continuum, is very important,” he says. “This presumes interoperability. If I have an ACO that covers 10,000 lives, NorthShore will need to manage each patient’s care and acquire the patient’s data, regardless of where the patient was seen or treated.”

He adds, “It is important that when a patient comes in for treatment, we know about that patient. This will enhance the overall experience, both for the patient and for our providers. Our clinicians need to have data about that patient—his or her history and images—if they are relevant for treatment. We are investigating solutions to make this possible.”

He continues, “I do see the need for a tighter integration between radiology applications and EMR applications, so that we can have more data available. It is no longer enough to place orders for radiology exams electronically and receive reports. Imaging informatics needs to present data in a manner in which they can be seen and used easily by a clinician—and so that we can perform advanced analytics on them.” 

This has resulted in the integration of specialty IT teams into the hospital IT department (a fairly recent phenomenon) occurring over the past year. NorthShore has an IT staff of about 300, of whom eight or nine are radiology IT specialists. The radiology IT staff has autonomy within the group, and the need to maintain on-site staff is presumed. Smith says that by integrating specialty groups, the organization makes staff resources available to handle future integration requirements in the most efficient way. 

One challenge is that HIE tools are not as mature and complete as they should be, Smith says. Few have the ability to transfer images. He says, “How are we going to deal with this? Many hospitals are waiting until after HIEs are developed, but we are starting to think about this now. This adds challenges to maintaining security, which is another very high priority for our organization. If we send images to the cloud, will they be secure? How will a cloud provider protect us from a security breach? Will audits be robust enough?” 

Smith is a firm believer that timely image transfer among HIEs will reduce duplicative diagnostic testing, but questions of how to transfer images safely, cost effectively, and efficiently remain. Related to these is another topic on Smith’s mind (as well as the minds of NorthShore’s imaging-informatics professionals): developing a long-term imaging/storage strategy. 

Imaging, in this context, represents the entire scope of images used in health care: diagnostic radiology, cardiology, pathology, clinical photography, radiation oncology, and more. NorthShore is gathering an increasing number of electronic diagnostic-level images from an increasing number of sources. What should be kept, and for how long? What protocols are needed to establish relevance? Should images be purged? Will this be more cost effective, overall, than retaining and managing them?

“All of these issues relate to imaging informatics and image management. They are what a CIO thinks about,” Smith observes.

Legacy Health

On the watch of John Jay Kenagy, PhD, in June 2013, Legacy Health (Portland, Oregon) reached the pinnacle of HIMSS Analytics EMRAM stage 7 for its six hospitals and 23 ambulatory-care facilities in Northwestern Oregon  He’s pleased by that, although he’s the first to say that progress toward this goal was in motion when he joined the organization, in 2012. 

The focus, today, for Kenagy and his team of 275 IT professionals, is on supporting the IDN’s primary goals: providing better-quality health services at a lower cost and with better access for patients. Data drive better care coordination and better management of diseases and treatment for patients with chronic illnesses. Data can make possible better management of the utilization of emergency departments, and data can support planning to manage and accommodate the patients who have improved access to primary care as a result of health-care reform. 

Good utilization management is the linchpin for all health-care stakeholders in the shift to value-based care, Kenagy says. It is the responsibility of both CIOs and CMIOs to make data easily available and usable to help facilitate better decision making—from the point of care through the entire enterprise. 

Increasing the awareness of physicians about the value that business analytics can provide is a key vision. “There is a wealth of information contained in our EMR and in specialty systems, such as radiology PACS,” he says. “The use of business analytics by health-care providers is really in its infancy. We’re in the Stone Age, but once clinicians understand the potential of data mining and analysis, they get excited—because they realize how these IT tools can help them, with respect to treatment and patient management.” 

Kenagy compares the potential effect on patient care of data-mining an IDN’s PACS to the impact that PACS technology itself had on radiology departments’ operations (and on the expectations of the physicians they served). “Radiology departments’ service changed dramatically for the better once they went digital, delivering reports within hours, instead of days. Imagine what decades of information about patients’ images and outcomes can provide, especially when that information is integrated with the data contained in an EMR,” he says.

“Today, when one of our primary-care physicians is using our computerized provider order entry system to place a prescription order for a patient with diabetes, that physician can determine what drugs the patient is taking that may cause interactions. What if the physician had access to a database that evaluated similar patients and could tell him or her what was most effective? This is best-practice clinical decision support: an IT tool that was developed by radiology,” Kenagy adds. 

“The field of medicine has been so positively affected by technology, particularly innovations in radiology,” he continues. “Sophisticated imaging has reduced the number of surgeries that would be performed if these exams didn’t exist, but it has also added to the cost of health care.”

He says, “Sophisticated imaging is making types of radiation-therapy treatment possible that couldn’t be done without it. This also adds to the cost of treatment. If we know that the treatment is appropriate, however—that debilitating side effects, with their own set of costs, may be minimized or eliminated, and that the outcome for the patient is better—the cost can be justified. Imaging informatics helps make this happen.”

Consolidation of archives into a vendor-neutral archive (VNA) is on the horizon, to be implemented in this decade, Kenagy says. So is integration with HIEs for image exchange. Right now, the IT team is working on projects that will enhance patient engagement. Legacy Health patients can access their patient records, but Kenagy would like patients to be able to schedule their imaging exams at the time and facility that best suit them. 

With advanced analytics, a radiology department could determine whether modalities could be used more, as well as whether hours of operation at ambulatory clinics should be expanded or shortened. This could have a positive impact through more productive scheduling of radiology staff (and better planning for the use of additional staff). 

Duke University Health System

Duke University Health System (DUHS), Durham, North Carolina, consists of three hospitals: 924-bed Duke University Medical Center (Durham), 369-bed Duke Regional Hospital (Durham), and Duke Raleigh Hospital. Neuroradiologist Christopher J. Roth, MD, is vice chair of radiology for health IT and clinical informatics and is director of imaging IT strategy for Duke Health Technology Solutions. 

Roth has been grappling with the impact of implementing a new enterprise-wide EMR. Academic-center facilities went live in 2013, with implementation continuing into the spring to incorporate the two community hospitals.

“This has been a monumental change, not only for imaging informatics, but for all specialized departments’ IT groups,” he reports. “Think of the impact of turning off more than 150 clinical applications and putting them all on one platform. Everything is centered on the EMR as we adopt an enterprise approach to providing patient care. We haven’t had much of an enterprise approach to imaging—a concept that includes not just radiology, but cardiology, pathology, endoscopy, ophthalmology, clinical ultrasound, clinical photography, and imaging in operating suites. 2013 was a pivotal year for IT.”

Roth’s job is to consolidate all of the imaging that is done throughout the entire system and provide physicians with seamless access to patient images via EMR. Image access also needs to extend to outpatient sites—and be just as easy, at those sites, as it is in the rest of the system.

As a result, Roth is overseeing major changes in imaging informatics. The existing stand-alone RIS versions will cease to exist, once the EMR implementation has been completed. The new EMR has RIS functionality, as an extension of imaging informatics to radiology. Unlike the freestanding RIS, the new RIS functionality is fully integrated into the broader EMR infrastructure and the broader continuum of care. 

The enterprise PACS is also being reconsidered as three components. The first segment consists of the archive and image-management components, which include lifecycle management, security, storage, and distribution. 

The second segment consists of the image viewer and associated image-manipulation tools, including advanced image-processing applications. Daily aspects of radiology-workflow management, such as peer review, critical-results communication, real-time imaging collaboration, and multimedia imaging reporting—important considerations of a successful PACS revision—are under reconsideration.

DUHS acquired a VNA and is migrating images from PACS, as well as from image-generating specialties (such as cardiology) for which images currently reside in archive silos. In addition to centralizing images, having a VNA will reduce operating and maintenance expenses. 

“The process of consolidating images is complex because it means not only transferring images from the heavy-hitter archives for cardiology, radiology, and ophthalmology, but also adding the images that may be on CDs, locked a closet, in smaller departments—in addition to adding images from a single modality (such as ultrasound) in a single clinic,” Roth emphasizes. “These are the kind of images we need to get into the system and affiliated with a patient’s record; care decisions were made based on those images, and we need to have them as documentation.”

This project has necessitated reassigning some enterprise IT staff. Supporting imaging informatics, whether for radiology or another specialty, is complex. Radiology IT still has dedicated staff separate from enterprise-imaging IT, in part because academic radiology departments have specialized needs for clinical care, innovation, research, and teaching that might not be as highly prioritized as enterprise needs during an EHR rollout. 

Roth explains that in reassigning IT staff resources (instead of restricting IT professionals to individual applications), the department is finding that it has more capabilities than it expected. To date, neither the radiology department nor the enterprise IT department has needed to add staff to support radiology: A more granular approach to reassigning responsibilities has resulted in more capacity than was previously assumed.

Being able to share and transfer patient images among health systems and patients is very much on Roth’s mind. Right now, the new EMR doesn’t transmit images to (or exchange them with) patients, state HIEs, or facilities outside the organization. This will be a future project.

In the world of value-based care and declining reimbursement, ensuring that radiology is at the dead center of care for all patients is one of the greatest ongoing challenges in which imaging-informatics tools can play a major part. “Our radiology departments are providing services to an increasing percentage of the patients in an enterprise,” Roth explains, “but I’m not sure that the services provided by radiology or radiologists are better integrated with their care.”

He adds, “In a value-based world, collaboration with referring physicians is going to become paramount. Imaging informatics can help make this happen at an electronic-integration level, and radiologists need to make this happen on a personal-integration level.”

Roth believes that imaging IT tools can assist radiologists with their interpretations by expediting the reporting process and by providing supplemental information. “Clinical decision support software, for example, provides real-time advice on the appropriateness of the exams that physicians order and alerts them if a similar procedure was recently performed,” he says. 

“Critical-results–reporting software delivers alerts of unexpected findings, and software exists that tracks recommendations for additional testing and when it should be performed; it can send alerts in a timely manner. It is an exciting time to be in imaging informatics because there are so many new technologies available to make care better,” he adds.

University of Wisconsin Hospital and Clinics

PACS remains very much its own entity at the University of Wisconsin Hospital and Clinics in Madison. In fact, the enterprise imaging group has been working with its PACS vendor to implement a registry and repository for Cross-Enterprise Document Sharing (XDS) and XDS for Imaging (XDS-I)—as well as to develop a patient-identifier cross-referencing, or PIX, actor for an HIE implementation. Gary J. Wendt, MD, PhD, enterprise director of medical imaging and vice chair of informatics in the radiology department, says that a major objective is to facilitate image exchange throughout an HIE.

“It’s not sufficient just to move small amounts of text,” he says. “Physicians need to have access to imaging exams. The two highest cost points in providing medical care are prescription drugs and high-tech imaging. If health-care organizations really want to provide cost-effective care, they need to be able to move the data around to be able to avoid the ordering of duplicate studies. This is a huge waste of resources for both providers and payors, not to mention the safety factor of exposing a patient to an unnecessary radiation dose.”

Making image exchange possible among unaffiliated entities is a major challenge. Contained within that challenge is another: Making sure that all data, including imaging, flow seamlessly, and that there are minimal barriers to moving those data. 

“If data move around easily, it will produce better information for the recipient. For example, a physician should not receive a warning that an exam that he or she is trying to order was recently performed at another hospital. That physician should have that exam, which is relevant to the patient treatment being planned, accessible from the patient’s records that he or she is working with,” Wendt explains.

The radiology IT team has joined the hospital IT department, but it remained a separate segment when the XDS/XDS-I transition occurred, about a year ago. Wendt believes that this organizational structure has helped maintain a more effective system. 

“To keep the enterprise imaging and associated subcomponents operational 24/7, you need staff with end-to-end expertise. If there is a problem, such as not being able to get images from a CT scanner to display on PACS, it’s invaluable to have a staff member who can troubleshoot the entire imaging chain: a person who knows the scanner, the configuration, the network, the display devices, and the PACS infrastructure,” Wendt says. 

He explains that in some large hospital IT departments, responsibilities might be compartmentalized. “Somebody may support an application, a virtual server, or a network,” he says. “It may take much more time to assemble a team that collectively can resolve the problem than it would take a specialized staff member (with comprehensive radiology IT knowledge) to resolve it.”

The other problem that deeply concerns Wendt is lack of awareness of what cannot be done with FDA 510(k)-approved IT systems. EMRs are not regulated by the FDA, nor are file servers, Web servers, time-management software, and much of the software used for hospital IT functions. 

Mistakes can be made while educating generalist IT staff that FDA-regulated IT systems (such as RIS, PACS, and 3D image-processing software) must be kept separated from nonregulated devices and platforms. Adding to the challenge of IT infrastructures is providing vendors with necessary access to the systems. 

“Vendors must have a fair amount of access to the systems; particularly, they need virtual-management capabilities to determine whether the system is set up appropriately,” Wendt observes. “In many virtual environments, that is not something that the typical IT group has been willing to do. Hospital enterprises planning consolidation of specialist teams with established departments need to be aware of this.”

A fairly recent technological trend that has had an impact on the enterprise-imaging group is the movement from a physical-hardware platform to a virtualized platform. The group is now focused on disaster sustainability, rather than disaster recovery. Wendt attributes this to using virtualized machines with very high availability. 

Development of software to provide security for mobile devices is another recent project: The enterprise imaging team worked with the hospital’s EMR provider to develop an app for smartphones, enabling users to take photos of whatever they wish to document clinically. As soon as images are acquired, they move to a server, are converted to DICOM formats, and are transferred into the hospital PACS. Neither images nor patient information ever reside on a mobile device.

Health-care systems across the nation are grappling with many of the same issues, when it comes to image management and availability: interoperability and high availability, with the ever-present need for cost containment. While imaging IT is experiencing greater levels of integration into enterprise IT, the discipline remains vibrant and alive, with the potential to make significant contributions to health care. 

“Imaging is a very technology-dependent and technology-savvy department,” Kenagy says. “It will meet the challenges of change that Legacy Health is undergoing, and undoubtedly, it will implement imaging-informatics changes that will have, as PACS did, a profound influence on our organization.”