The PACS — picture archiving and communication systems — have been in existence for more than 45 years. One of the first was created in 1972 by Richard J. Steckel, M.D., and the large-scale introduction occurred 10 years later at the University of Kansas. In the early ’90s, U.K. physician Harold Glass, M.D., pioneered the use of PACS to transform London’s Hammersmith Hospital into the first filmless hospital in the United Kingdom. This emerging technology was based on a need to transition from hard copy to digital imagery, which removed many constraints of moving film around departments and physical file storage. The result was a consolidated technology model for storage and administration of digital radiology images.
Over years of trial and error, the legacy PACS system evolved to handle rudimentary functions required to run an imaging department. These functions were developed across an array of department users including technologists, radiologists, and administrative and file clerks.
PACS functionality for these users included:
• Image receipt, storage and retention based on an evolving DICOM standard;
• Image quality control for editing, manipulation and maintenance;
• Rudimentary PACS-driven workflows for technologists and radiologists;
• Exam visualization for radiologists;
• Document management and storage for exam diagnosis reports prior to automated voice recognition;
• Integration with scheduling systems and radiology information systems (RIS) as they were introduced; and
• Later use of portals for referring physician access to reports and imagery.
These functions signaled a transition from paper to digital workflows for all users within the radiology department.
Digital radiology. The emerging Digital Imaging and Communications in Medicine (DICOM) standard defined the digital file type utilized for communication between image acquisition devices (modalities) and the PACS. Although initially very loose and challenging when working with multimodality and PACS-to-PACS environments, the standard eventually became a stable solution for image portability. However, this portability was only feasible between DICOM-compliant devices in an organization. In parallel, advanced PACS needs for clinical document storage resulted in the development of applications to scan, convert and send non-DICOM compliant physical documentation to the PACS. These documents included paper orders, worksheets and additional clinical notes. Other non-DICOM compliant imaging devices such as ultrasound required the use of inline capture hardware to screenshot and send selected still images as seen by the stenographer. Although the conversion processes provided a feasible alternative for non-DICOM devices and clinical documentation, it required additional cost and manpower to transition to a filmless and paper-free radiology workflow.
In the meantime, other image producing departments began similar initiatives to convert and consolidate within a filmless department. This led to the development of department-specific systems such as the cardiology picture archiving and communication system (CPACS). Although functionality was fundamentally similar to a PACS, differences in departmental workflows and image visualization requirements meant that the radiology PACS existed as mutually exclusive and siloed imaging systems.
Canning the Solution. For the most part, PACS primarily served as the end-of-the-road for radiology imaging until the advent of HL7, when it could begin receiving scheduled orders and sending results back to associated EMRs, EHRs and foreign RIS. The exception to this was native integration with same-vendor RIS systems but meant all capabilities were defined by the best interests of a single vendor. The PACS closed-architecture system became a “canned” and “one-size-fits-all” solution to radiology. While PACS was advancing standards in radiology technology, limited development flexibility meant users were required to augment the vendors’ way of doing business instead of the software accommodating their unique use cases.
As PACS evolved as a powerful tool for transitioning to a filmless and paperless radiology system, the incidental creation of siloed technology meant limited cohesiveness with ancillary hospital service lines. Together with single vendor development limitations, limited workflow flexibility, resounding cost and operational overhead, a breakdown of PACS functionality was imminent. This breakdown has been defined as the deconstructing and stratification of enterprise imaging.
EI Pushes PACS to the Limit
Much has been said regarding the modular or “deconstructed” approach to imaging. The philosophy supporting this approach is grounded in an assumption that investing in a layered stack of solutions provides greater technical agility, standards-based integration and better solution choices at the functional level. Especially true for enterprise imaging, this approach typically parses functionality at the archiving, visualization and workflow layers of traditional PACS systems. The rise of smaller, innovative companies has filled gaps where many larger vendors have been slow to develop and overcome. This rise of innovation advanced traditional PACS-type functionality by independently focusing on the specific functionality layers, which include browser-based and thin client viewer solutions, wide adoption of DICOMWeb and IHE standards adoption, agnostic storage and greater worklist functionality.
Hybrid environments. New workflow and visualization functionality is a key advancement beyond the traditional PACS by layering applications to simultaneously support multi-departmental and enterprise workflow demands. These new capabilities can now span multiple departments and service lines. Many organizations now also approach image archiving as a data management solution (vs. simple storage management), which includes robust tools to support any number of archiving, routing, normalization and lifecycle management requirements. This led to the arrival and capability for consolidated archiving solutions (vendor neutral archives) to service an entire enterprise of service lines. In addition, the interoperability required for multi-vendor integration resulted in highly functional and well-documented APIs to support peripheral solutions. These peripheral solutions enable organizations to exploit the use of artificial intelligence (AI) engines, image sharing and ultimately patient applications that have been independently speeding to market.
A deconstructed challenge. Some important forces that must be considered in adopting a deconstructed model include resource availability, technical collaboration, governance and organizational behavior, and vendor management. Designing and establishing a multi-vendor solution requires simultaneous and synchronized resources from the provider and all vendor parties. Many organizations are resource-constrained and thus struggle to implement the necessary interfaces, integration points and robust test environments to ensure implementation and ongoing support meet the stringent demands of a critical clinical system — especially in the acute care environment. During the selection and design, it is important to have the internal expertise to guide a collaborative approach across all participants as each “deconstruction” can be technically represented differently. Due to legacy systems, peripheral technologies and functional requirements, collaboration will ensure that each vendor understands the integration points, performance and ultimately the data management model to protect integrity.
Many organizations also struggle with creating effective adoption models, which can become a hurdle even for strong technical and improved clinical solutions. Success requires governance that stays engaged, adheres to the proven science of organizational behavior (people don’t like change), and commits to strong communication and support models. This requires effective leadership and a lot of courage. As with any multi-vendor ecosystem, vendor management must be coordinated and aligned. With deconstruction comes the responsibility of managing all vendors with a shared commitment model and service level agreements. Everyone has experienced system failures across solutions where vendors point fingers and under-commit to issue resolution. Imaging is a mission-critical platform, and collaborative issue resolution across technologies and vendors in the deconstructed model ensures minimal impact to care delivery. This complexity has influenced many provider organizations to rethink deconstruction and reconsider best of breed vendors who offer a simplified approach to their support model.
The Road to Reconstruction
The path to deconstruction was comprised of a need for flexibility, economics and integration. However, governance and supportability of multi-vendor environments can prove to be risky without existing leadership structure and expertise. While it was early predicted that the deconstructed model would replace PACS entirely, many organizations have determined that deconstruction is an overwhelming and difficult path to pursue.
Although the myopic design of PACS missed the industry signals for change and clinical consolidation, its shortfalls led to a natural and progressive deconstruction. And the innovation required by the deconstructed model has become “proof” that trailing PACS vendors can no longer rest on the laurels that maintained early profitability. Without re-engagement, the road is moving under their feet and they can no longer stand still. Progressive traditional PACS vendors are now investing in development, mergers and acquisitions, and rethinking ways to holistically engage user groups and clinical use-cases, much as they did in the early days. As a result, this restoration of focus is challenging multi-vendor deconstructed models with a more elegant single-vendor enterprise imaging solution, the “Reconstructed PACS.”
Dave Whitney and Jef Williams are recognized thought leaders in healthcare systems, design and integration. They both provide subject matter expertise in the fields of medical imaging, clinical workflows, business collaboration, and are frequent publishers of prognostic change.
April 24, 2024 
