PACS Next Steps: A Q&A with Don Dennison


Don Dennison Discusses PACS Next StepsSince the first PACS was developed in the early 1980s, medical imaging systems have continued to improve. Monitors have higher resolutions for viewing clarity. Images have 3D capabilities. Faster network speeds allow for more efficient image access.

In the 30-plus years since the construction of PACS, healthcare is now seeing its demise — or evolution. In this Q&A, industry consultant and speaker Don Dennison discusses the drive behind PACS’ evolution, what healthcare leaders need to consider and benefits organizations can gain.

Q: What is driving the need to “deconstruct” PACS?

A: I don’t know that I would refer to this trend as a “need,” but rather a trend in response to market forces and technical opportunities.

Essentially, with the rapid adoption of Electronic Medical Record (EMR) systems in the U.S. and other jurisdictions, there is a common belief that images — being part of the patient’s medical record — need to be available within the EMR.

Image-Enabling the EMR

Most PACS designs are optimized to maximize radiology department productivity, specifically around the diagnostic image acquisition and reading workflows. Shifting image viewing to a system that can embed the viewer directly in the EMR client, without installing any software (a so-called zero-footprint viewer), is one popular method for getting images into the EMR. This enterprise viewer is expected to display all radiology-generated images, as well as images from other specialties, in the context of the patient, encounter/visit, or procedure.

Similarly, providers are looking for a system to store and manage imaging data being generated by all departments, not just radiology. By deploying a vendor neutral archive (VNA) in a highly available configuration, they can provide image management services for all departments. Also, by integrating this system with the EMR — for example, by having the EMR provide patient information in HL7 ADT messages — the VNA can act as the image repository for the EMR. Patient, and even encounter/visit, information is in sync between the two systems.

By shifting the enterprise image management and viewing to systems other than the radiology department PACS, it allows the departmental PACS to focus its resources on its primary functions — acquisition and reading workflows.

The above two changes in system responsibilities are probably the most achievable changes today, given the maturity of the applications available on the market and the effectiveness of the methods necessary to achieve this change.

Beyond this, another trend is to move the reading worklist from being provided as part of the PACS (or departmental RIS) to the responsibility of a RIS module in the EMR, or a so called enterprise worklist provider.

Fostering Interdepartmental Access of Images

The intent of this change is to provide more effective and productive reading workflow across multiple facilities. If all the facilities involved in the reading workflow are part of a single organization (owned or strongly affiliated), it is more likely that they would choose the EMR RIS module to provide the reading worklist. If the facilities involved are loosely affiliated, it is more likely that a standalone Enterprise Worklist application is used. Regardless of which approach is more logical from an IT perspective, the reading worklist application’s capability to provide the necessary worklist definition and assignment, to enable the desired reading workflow, is an important aspect to consider.

The driver behind using non-PACS-provided reading worklists is often maximizing resources. This includes subspecialty assignment, load-balancing work across readers across the enterprise, meeting RVU and SLA requirements, and dealing with off-hours reading assignments.

Often, consolidating the voice recognition and reporting application to a shared application across facilities — where procedure-specific report templates, users’ voice profiles, and the interface to the EMR for orders management and results distribution are centralized—is evaluated at the same time as the reading worklist change.

With the enterprise user access shifted to an enterprise viewer in the EMR, the image management shifted to an EMR-aligned enterprise image repository, and the reading worklists provided by either the EMR or an enterprise-wide reading workflow management application, the primary remaining component of the PACS is the diagnostic image viewer.

Utilizing enterprise viewers

As enterprise viewers evolve, some predict that they will prove capable of replacing this part of PACS, at some point in time. Whether this proves true (I predict it will, but the timing of this is debatable), the good news for providers is that they don’t have to bet on which way this will go today.

As the provider will employ an enterprise viewer (for EMR image viewing) and keep their departmental PACS during this transition, they can compare the two systems in their own environment, and when the radiologists deem the enterprise viewer mature and capable enough to read from, a transition plan can be put in place. Whether this happens today, next year, or five years from now, providers can enjoy the benefits of the other changes while they form a long-term strategy for diagnostic image viewing.

Overall, the changes are being driven by a shift to use applications that are much more aware of multi-facility, multi-department, multi-specialty enterprises, and that are able to provide services that allow many systems to be consolidated into far fewer (while still providing clinically effective capabilities).










In addition to the benefits already mentioned, organizations taking this approach are also looking to add new services around the capture, management, and display of so-called Enterprise Images. These are often generated in a clinical setting, but may also be used in diagnosis, for supporting documentation, or for use operations. Developing an enterprise imaging program that includes the EMR, a VNA, and a document management (sometimes called Enterprise Content Management or ECM) system and sets organization-wide policies and governance for the capture and indexing, as well as the presentation, of images (and their related information) is becoming more and more common at leading institutions. Often the patient’s clinical information, along with any enterprise images, is highly desirable by the radiologist on the reading workstation. There is growing belief (and anecdotal evidence) that the correlation of a patient’s clinical information (from the EMR), enterprise images and diagnostic images at the point of diagnosis leads to a more accurate and specific diagnosis and a more precise result.

Q: What are the key considerations in the strategy to consolidate data within an organization?

A: Imaging data, for the most part, has been held to a fairly low standard of correctness and completeness, in my experience. In most organizations, if the images make it into the PACS, display well enough that the radiologist can review them to complete their report, this is considered “good enough.”

For example, there are lots of DICOM attributes that serve a valuable purpose in properly identifying the study’s patient identity in a multi-facility enterprise that uses multiple MRN (Patient ID) domains/formats. These are often not included in the DICOM objects, or contain no information (a null value), however.

In a consolidated enterprise, imaging data from multiple facilities will be combined and cross-indexed in shared systems. Inconsistency and incompleteness of the data will cause issues when viewing and comparing the data, as well as when trying to apply automated processes, like relevant prior prefetching and routing (the rules often need to be changed to deal with the variability of values within the images).

Before an organization starts putting all the disparate DICOM data from the various PACS at the facilities that have been acquired (or affiliated) in a consolidated enterprise, it is important to evaluate the consistency and completeness of these records.

Establishing Quality Policy

One approach I recommend is to establish an imaging record quality policy that defines exactly what is expected or required to be true of the imaging record before it can be made available in or through the image repository. Tools and methods that measure and report on policy compliance can then be developed or acquired. Once the records that do not comply can be identified, corrective actions can be applied. This may involve correcting the data only, but sometimes requires corrections to the system or processes that are the source of the data, to prevent further data quality issues.









While a complete and consistent patient imaging record has clinical benefits, the normalization of data as part of this process makes the use of analytics for process improvement much easier. If the data is inconsistent and/or unreliable, it is hard to draw and depend of patterns and conclusions from an analytics program.

Matching Patient Images and EMR

Another important aspect is to ensure that the image management system’s information on the patient matches that in the EMR. Most EMR can provide an interface to share sufficient data about the patient, including all the Patient IDs by which the patient is registered, as well as send message to perform automatic record updates, merges, splits, etc. Often, VNA that have been designed for multi-facility enterprise use have capabilities in this area that exceed those of the standard departmental PACS, but the purchasing provider should explicitly evaluate current and potentially new systems against their needs. Simply labelling something a PACS or a VNA does not guarantee the necessary capabilities are lacking, or are included.

Q: What are the key benefits of consolidating systems/data?

A: Obviously, the fewer systems, the less complexity and (presumably) lower costs (purchasing, operations, maintenance, etc.). Another benefit is centralization of talent and control over policy (record quality, secure access, etc.) enforcement.

If an enterprise with a shared EMR can achieve a single system for long-term image archiving that also acts as the central point of data access, combine this with a single enterprise viewer for embedding in the EMR and any other portal (for external referring physicians, patients, etc.), and have a single system to manage all reading workflow to ensure productivity and quality, they will have much more control over their costs and risks. If they can consolidate the diagnostic Radiology image viewer (in a single PACS or enterprise viewer), they can gain additional cost efficiencies and can even increase collaboration among users (rads, techs, etc.) across the facilities.

Benefiting from One System

When attempting to share data across enterprises, it is much easier (and cost effective) to develop interfaces with one system, such as the VNA, than with many departmental systems. Likewise, embedding the enterprise viewer into externally managed portals, such as ones provided as part of a Health Information Exchange (HIE), is much easier to achieve than developing integrations with all the various PACS clients in all the departments in all the facilities.

One of the benefits of employing an enterprise viewer to image-enable the EMR (and other portals), is that the large user community that access the EMR, which often measures in the tens of thousands of people, can become familiar with a single user interface, regardless of where and how the imaging data was acquired. Usability of imaging record discovery and access cannot be underemphasized.

This separation of responsibilities—the enterprise viewer for image record access and the PACS for diagnostic image reading—allows the imaging departments to have the autonomy to adapt to the best practices in their field. As long as the images they produce comply with the imaging record quality policy, they should be consumable by the enterprise viewer and any other specialty system integrated with the VNA.

Supporting Value-Driven Reimbursement Models

In the end, the organization that can provide the best access to patient information and coordination of care, across the most subspecialties and care settings, will be in the best position to have their staff deliver the best patient outcomes. And in the emerging value-driven and outcomes-driven financial model of healthcare, this goal cannot be ignored. Imaging practices will either become involved in this trend, or will find their value (and compensation) dramatically declining.

Don Dennison, an industry consultant, is a speaker and panelist on topics such as medical imaging record interoperability and integration of imaging data within the EMR. He has published numerous articles on various imaging informatics topics. Don serves on the Board of Directors of SIIM and chairs the ACR Connect committee.

Have interviews with industry experts and other information delivered to your inbox by subscribing to the Medical Imaging Talk Blog.

Subscribe to the Medical Imaging Talk blog










Leave a Reply