PACS: Dead, Dying or Re-born?

The great American humorist Mark Twain once quipped, “Reports of my death have been greatly exaggerated.” If digital image management technology had a voice (and ergonomists would likely agree that it should), it also might pronounce its longevity.

Some naysayers have scribbled an obituary for PACS, claiming good riddance to a radiology-centric, efficiency-obsessed system riddled with baffling buttons. Others refuse to bid adieu, and underscore the central significance of imaging in healthcare.

To better understand if PACS is on the edge of obsolescence or if a new and improved model is about to emerge, Health Imaging spoke with three thought leaders about digital image management.

PACS 3.0: Meaningful innovation for meaningful use

Paul J. Chang, MD, Professor of Radiology, Vice Chair, Radiology Informatics; Medical Director, Pathology Informatics; Medical Director, Enterprise Imaging, The University of Chicago Medicine

“We’ve had two PACS deaths already,” says Chang, adding that the third generation is an expected and welcome evolution of digital image management.

First-generation PACS seemed excruciatingly painful in retrospect. The sole emphasis was on translating film to digital in an environment that lacked networks, technology, archives and more. This radiology-centric, thick-client model has ceased to exist.

“The battle cry of PACS 2.0 was workflow,” says Chang. Radiology had solved the film-to-digital transition and began to leverage digital image management systems to improve workflow. PACS entered the age of the -ologies and employed an enterprise-centric model. Web-based systems, thin clients and EMR integration were the hallmarks of PACS 2.0.

“We’re seeing the end of this cycle and the beginning of PACS 3.0. The fundamental difference is the drivers in radiology have changed dramatically.” As PACS 1.0 and 2.0 were developed and disseminated, reimbursement was a given. Radiology could focus on maximizing its investments through maximum efficiency and throughput.

Although PACS may not be dead, fee-for-service is on its last legs. In new reimbursement models, radiology’s value will hinge on its capability to manage imaging, rather than reading studies.

“In a capitated model, radiology is a cost sink not a revenue generator. This requires us to change our view. Anything we talk about when we design the next-generation PACS has to directly result in measurable improvements in the value proposition, with value defined as quality, efficiency and safety simultaneously.”     

To navigate this disruption, radiology needs to understand that its stakeholders extend beyond referring physicians, patients and local hospital administrators and now include entities beyond the firewall. New stakeholders include third-party payers, who are aligned with radiology in accountable care organizations, as well as government agencies that require analytical data.

In addition, patients have morphed into healthcare consumers who demand value. “If a radiology provider can’t demonstrate its differential value and provide the best service, the patient can go elsewhere.”

Many in radiology, says Chang, aren’t ready for these changes and need PACS to address these new realities. Third-generation systems will require an archive neutral approach that leverages the cloud and emphasizes interoperability. Chang foresees radiologists working in highly distributed, federated environments that will require interaction among multiple vendor neutral archives.

“It’s easy to send current images and priors, but it isn’t enough. We also need clinical context.” PACS will need to interact with the EMR and other information systems to provide data, such as lab values and pathology results.  

The third element of PACS 3.0 is robust business intelligence and analytics. Reimbursement will depend on meaningful services that add value. Traditional data, such as turnaround time, are insufficient, says Chang. Correlation between imaging and clinical outcomes will be a key metric. This requires an unprecedented level of integration with the EMR.

“This is a huge challenge and opportunity,” Chang concludes, before offering a few words of wisdom for RSNA attendees. “If you want to see where the zeitgeist is truly moving, visit the small vendors.”

The next intersection: Imaging informatics & ethics

Keith J. Dreyer, DO, PhD, Vice Chairman of Radiology at Massachusetts General Hospital (MGH); Associate Professor of Radiology at Harvard Medical School, Boston

Radiologists have failed to ask all the pertinent questions, says Dreyer. For decades, radiologists have asked: Is X going to make me more productive; Is Y going to save me money. “Those are fair questions to ask, but they aren’t the only questions.” Other relevant questions as imaging informatics hurdles ahead include:

• How will this system help improve the overall patient care process and its quality?
• How will this system allow me to become more viable and valuable in that process?
• How will this system improve my services and interactions with referring physicians and their patients?

Radiology has been guilty of a self-serving approach to technology deployment, charges Dreyer, as many have focused on digital projects with a myopic aim—improved productivity. Dreyer offers a few words of wisdom for radiologists who spurn Meaningful Use and claim that radiology is sufficiently digital. “At the federal level, the emphasis is not on radiologist productivity. The emphasis is on improving overall patient safety, access and quality while also managing costs. I’m not convinced, nor are the feds, that our efforts to date have been so altruistic.”

A patient-centric imaging informatics approach differs substantially from the conventional model. Rather than pursuit of the fastest workstation, radiology should consider technologies and methodologies more directed toward patient care. For example, access to prior patient images and EHR data across enterprise boundaries or the capability to produce structured multimedia reports with consistently coded, retrievable objective measurements and recommendations for further actions would dramatically improve patient care. However, the architectures to support such functions differ significantly from current monolithic systems designed to optimize throughput, says Dreyer.

Next-generation PACS need to help radiologists remain a vital part of the patient care process, which means giving both patients and physicians what they need. At the informatics level, patients need to be able to securely move images to caregivers outside of siloed hospitals. “That’s not how PACS of today have been designed or implemented.”

Enterprise distribution and vendor-neutral archives may be reasonable starting points as they extend the data model beyond the siloed department to the provider-centric enterprise; however, neither offers a complete solution. From the patient perspective, enterprise boundaries are artificial and restrictive to improving the healthcare experience. Radiology needs robust two-way communication with integration across institutional boundaries that allows for order entry, scheduling, decision support, image sharing, physician and patient communication pathways with critical alerts and notifications—capabilities that will allow radiologists to extend their (virtual) presence beyond their departments throughout the entire patient care process.

“Clearly, this requires a different architecture, such as cloud-based computing,” says Dreyer. Cloud-based-systems enable secure, standards based, bi-directional image and communication sharing at the flip of a switch at the patient, provider or organizational level. In contrast, current image transfer mechanisms are arduous afterthoughts that require opening DICOM ports, validating and authenticating information or worse yet, massive reliance on CD imports and exports.

Dreyer urges radiologists to consider how the next stages of Meaningful Use will affect imaging informatics. The Stage 2 final rule includes requirements for healthcare providers to be able to display image data within their EHRs. It’s a fairly straightforward requirement for enterprise systems like MGH, where 100 percent of the hospital’s images have been available in the EHR since 1997. A private practice comprised of 10 radiologists who operate freestanding imaging centers that service dozens or more offices, each with a different EHR, faces a drastically different challenge. This challenge is exactly where radiology needs to position its IT efforts going forward, he says.

Dreyer concludes, “As you walk the halls of McCormick Place looking for new IT solutions to improve your departmental productivity, be mindful of how that technology will affect your ability to more deeply participate in all aspects of the patient care process.”

The smart PACS era

Eliot L. Siegel, MD, Professor of Diagnostic Radiology and Nuclear Medicine, Associate Vice Chairman for Informatics, University of Maryland Medical Center (Baltimore); Director, Baltimore Veterans Affairs Medical Center Radiology

Nearly 20 years after the first transition to digital image management, PACS co-exists with a host of sophisticated technologies, such as advanced artificial intelligence and high-performance computing. PACS, says Siegel, seems relatively immature.

In 1993, when Siegel and colleagues at the Baltimore VA Medical Center deployed PACS, they were thrilled with the futuristic system that replaced film. Yet, despite the bevy of bells and whistles, modern PACS offer little more than the first system—digital image access.

Despite the slow pace of progress, Siegel is optimistic at the prospect of relevant developments. The most significant change will be the advent of much more intelligent systems that actually do what radiologists want (and need) them to do, he predicts. The list of unvoiced, unmet needs bridges the entire imaging process.

Improved pre-reading tools are essential. PACS, says Siegel, should automatically mine the record before the radiologist reads the exam. Smart searching replaces, or supplements, processes that often don’t occur in digital departments. In the film era, radiology was a social hub. Physicians dropped in to review cases; clinical communication was a built-in process. No digital equivalent has yet replaced the social network that characterized film-based departments.

Academic radiologists leverage residents and fellows to review charts and talk with patients to extract key information. Most nonacademic radiologists face Solomon’s choice:  review the chart to gain clinical context or cut corners and interpret the study without the context of the record. “The PACS of the future should act as an artificial fellow and provide this functionality to make radiologists more effective, efficient and accurate.”  

This functionality hinges on better integration between PACS and the EMR, which would enable access to laboratory information, problem lists, progress notes and discharge summaries at the PACS workstation.

Vendor neutral archives also need to step up the game. It’s fairly typical for radiologists to prefer a best-of-breed workflow for specific applications, which requires either storage on both systems or use of the notoriously slow and inefficient DICOM query/retrieve. In an ideal world, all image datasets would be easily accessible on a single system.

A related issue is the plethora of non-DICOM formats for specialized applications, such as PET/CT or functional MRI. These data usually require redundant storage on the subspecialty workstation because vendors often use proprietary formats that cannot be stored in the archive.

Siegel also is looking for improvements in critical results notification. Radiologists may close the case after a manual or semi-automated system notifies the clinician of the results. “Our responsibility doesn’t stop there. Studies show as many as 20 to 25 percent of cases are associated with the clinician not following up despite an alert to a critical finding. As radiologists, we may believe we’ve finished once we make a diagnosis and alert the referring physician. Whether or not that finding turns into action is information we should have.”     

Radiologists unconvinced of the need to lobby for such improvements may be wise to consider the relationship between quality and the bottom line in future payment models. Currently, there is a great deal of disparity among imaging reports. Future reimbursement models will call for radiologists to structure reports to include specific points, which require more sophisticated informatics systems that monitor such details.

As PACS has evolved from its radiology-centric roots to an enterprise tool and radiologists struggle to maintain their relevance, radiology departments may want to advocate for end users’ needs. Two items near the top of the list? Enterprise-wide access to full image datasets and a streamlined imaging history of the patient. “If my report refers to a lung nodule on cut 17 of the CT exam, a hyperlink should launch the viewer and navigate the physician to the exact location.”

Finally, while vendors have outfitted PACS with hundreds of features, most radiologists use only 25 percent of them, suggesting a need for additional training to help radiologists optimize existing PACS tools.

Siegel’s advice for stakeholders heading into the Windy City: Question why systems work the way they do, and how they might work smarter.