The big question: Where is deconstructed PACS going?

February 06, 2017
By Michael J. Gray

The replacement radiology PACS model popularly known as deconstructed PACS is not sustainable over the majority of the market that has yet to replace their current PACS solution. As time marches on into 2017, the primary features of today’s turnkey radiology PACS solutions that make them look less attractive than the deconstructed PACS solutions are slowly, but surely, evolving, and those major negative differentiators are disappearing.

What is deconstructed PACS?
The concept of deconstructed PACS was introduced a few years ago. This was after the vendor-neutral archive (VNA) had established its foothold in the medical imaging market and the universal viewer was introduced to image-enable the growing number of electronic medical record systems. At that point, we were witnessing the arrival of significant new data management and exchange features (VNA) and highly performant display capabilities (universal viewer) that contrasted sharply against the lumbering old technology represented by the turnkey PACS solutions. It was a natural step for the early adopters to consider pairing their existing VNA and universal viewer with other best-of-breed components to build a “next generation” replacement PACS for radiology.

Deconstructed PACS is simply a semantic variation of best-of-breed PACS. The former descriptor stuck, most likely because the latter did not make an appealing acronym. Deconstructing a turnkey radiology PACS solution into its major components results in five components: Acquisition/QC; Workflow and Worklist; Diagnostic Display; Data Management (Archive); and Clinical Distribution and Display. Many of the early adopters recognized that they already had a superior archiving solution in their VNA.

They had a superior clinical viewer in the universal viewer that image-enabled their EMR, and in some cases, they had (at least) acceptable workflow/worklist capabilities in their EMR, or superior capabilities in a standalone workflow/worklist application. If one recognizes the image acquisition capabilities of the VNA, these early adopters were only one component shy of a higher-performing and more feature-rich replacement radiology PACS than what was available as a traditional turnkey PACS solution. All they needed was one or more specialized diagnostic display applications that they could integrate with the components already on site and they had their replacement PACS.

There are some powerful arguments for the deconstructed PACS model. The true vendor-neutral archive is a far better data management solution than the “archive” component of the turnkey PACS. The VNA can assure data compatibility between disparate PACS. It can support a very sophisticated user-defined information life cycle management strategy. Its cost can be amortized over multiple imaging departments. Most importantly, the better VNA packages included access to the file structure and database schema, making it possible for the customer to basically eliminate an expensive migration from one VNA to another by simply migrating the database directory to the new VNA and pointing it to the image data on the existing storage solutions.

The deconstructed PACS strategy allows for the integration of specialized diagnostic applications with the VNA. This flexibility to choose the most appropriate diagnostic display applications presents a better solution for sophisticated radiology departments that might want one diagnostic application for general radiology, another specific to mammography, another special application for nuclear medicine, and perhaps a fourth for advanced ultrasound. The deconstructed PACS strategy provides more choices for optimizing the clinical display application, the one that will image-enable the EMR. Rather than being strictly radiology-centric, the clinical viewer could support radiology, cardiology and DICOM as well as non-DICOM native data object formats.

The deconstructed PACS strategy also allows for a greater choice for the workflow/ worklist component of the PACS. In addition to a richer set of features and functions, an enterprise workflow application could support inpatient-context access to relevant clinical data known to the EMR, like the structured data being managed by the enterprise content management solution, DICOM studies from other PACS or image data repositories, and non-DICOM images captured by mobile devices and outside images forwarded through image sharing solutions.

There are several negatives to the deconstructed PACS model. First and foremost, the deconstructed PACS “system” is composed of multiple pieces of technology (software applications) developed by and supported by different vendors. These independent pieces require sophisticated and optimized interfaces to support their integration into a functional and highly performant system. The independent components each have their own remote system monitoring dashboards, making it difficult to troubleshoot. For these reasons, a deconstructed PACS model requires a sophisticated IT department with deep experience in project management, help desk and system monitoring. Lastly, the deconstructed PACS model almost always requires a significantly larger investment than a comparably configured turnkey PACS solution.

PACS vendors’ reactions
The vendors that were still offering the traditional, turnkey PACS solutions that notably featured decade-old infrastructure and fat client diagnostic displays were understandably “cool” to the concept of deconstructed PACS. Their initial counter strategy was negative selling, focusing on one or more of the negatives listed above. During the past two years, while that negative selling strategy was failing with some of the large IDNs that chose the deconstructed PACS route, many of the turnkey PACS vendors were executing their technology catch-up programs.

Their first step was the seemingly overnight conversion of their PACS archive into a VNA. The second step was replacing their fat client clinical viewer with either an in-house or OEM-supplied version of the zero-client universal viewer. Several spent the time to develop an improved workflow/worklist application. As of RSNA 2016, at least one of these vendors has replaced the old fat client diagnostic display application with a zero-client, server-side rendering, pixel-streaming diagnostic display application.

In addition to these upgrades to their software application stack, many of the turnkey PACS vendors have made the effort to optimize the interfaces that would allow more efficient connectivity to both their own individual applications and those offered by other display and VNA vendors. In two-plus years, many of the turnkey PACS vendors were back to being competitive with upgraded applications that fit nicely in a deconstructed PACS model, or a mostly new and improved turnkey PACS (if the customer didn’t mind waiting another year for the complete system upgrade to be completed).

Best-of-breed component vendor reactions
The vendors promoting the deconstructed PACS strategy had to foresee the difficulties they would soon face in selling a (relatively) expensive, multi-piece, multi-vendor solution to health care systems that did not have the requisite IT staff or financial resources. As indicated in Geoffrey Moore’s Technology Adoption curve, there are significant differences between the innovators/early adopters and the early majority/late majority.

The former groups are more inclined to take reasonable risk and are willing to pay a premium for the technology they feel they need. The latter groups are much less likely to take any significant risk, and they generally do not want to pay for technology they know they do not really need. Faced with this potentially gloomy forecast, the deconstructed PACS vendors went to work expanding their offerings. The vendors that offered universal clinical display applications went about adding features and functions that one would expect to find in a diagnostic display application, including hanging protocols, 3-D rendering and fusion capabilities and the ability to handle digital breast tomosynthesis image studies.

The vendors that already had a zero or near-zero diagnostic display package went about developing their own VNA technology, or at least building distribution partnerships for the VNA component. The vendors offering stand-alone VNA technology went about developing, buying or partnering for clinical and/or diagnostic display technology. Both the component display and the VNA vendors frequently partnered with the vendors that offered the stand-alone workflow/worklist technology to create what certainly began to look like a turnkey PACS solution comprising best-of-breed components. How ironic, indeed, that the vendors who launched the anti-turnkey PACS movement would eventually seek to look like a turnkey PACS provider.

This packaging of best-of-breed components into a somewhat turnkey solution makes sense. Whether the individual components came by way of in-house development, OEM arrangement, purchase of a company or simply a strategic reseller agreement, the technology packaging allowed for tighter integration of the software components, unification of the disparate remote system monitoring solutions and a much more coordinated and efficient project management process. Any vendor that can now offer all five components — Acquisition/QC, Workflow and Worklist, Diagnostic Display, Data Management (Archive) and Clinical Distribution and Display — effectively can market themselves as a turnkey PACS solution provider.

My market forecast is something of a summary of what I have already presented here.

• The deconstructed PACS strategy appeals to innovators and early adopters, and based on the adoption rate of deconstructed PACS over the last two years, the deconstructed PACS market segment is nearing saturation as we move into the early majority segment of the technology adoption curve.
• The deconstructed PACS strategy is not as appealing to the early majority and late majority segments of the market. These market segments do not need high-end features/functions and component flexibility, and they certainly do not want to pay for features that they will not use. These market segments have limited IT resources and the experience to self-manage complex multi-component systems.
• Turnkey packages with fewer separate parts are more attractive to the early majority and late majority market segments. Their optimal comfort level is with “one throat to choke” and the large established vendors. Fewer separate parts and separate vendors certainly suggests fewer software compatibility and interfacing issues. In these market segments, value-based pricing is a strong plus. Acceptable technology that allows for competitive pricing is a sought-after characteristic of the deal. Larger vendors, especially those that offer imaging modalities, can bundle the turnkey PACS with the larger modality deals. Finally, fewer vendors contributing components to the turnkey package usually means lower deployment and operating costs.
• The traditional turnkey PACS vendors are working their way toward being able to offer turnkey PACS solutions comprising new software technology that looks a lot like best-of-breed technology that launched the deconstructed PACS movement.
• The deconstructed PACS component vendors are morphing into PACS providers that will soon be able to offer turnkey PACS solutions comprising best-of-breed software applications for the larger early majority and late majority market segments.

What should any health care system do in preparation for a radiology PACS replacement project? The first subject I always bring up in the kickoff meeting with a new client is qualifying and quantifying the organization’s IT resources. This process is basically an honest and careful self-evaluation of the IT department by its leadership. All of the questions should focus on determining whether the department has the staffing and experience to support a best-of breed solution.

After determining the IT resources, the next step is to do your homework. Research both the vendors and the solutions in both categories: turnkey and best-of-breed component. Ask the vendors to identify their product road maps and their recent history of meeting their previously stated road map goals.

Develop an enterprise imaging strategic plan that covers both the larger imaging departments like radiology and cardiology that have already gone digital, as well as all of the smaller imaging departments that may or may not have converted to digital, and may or may not have a department PACS. The plan should also include a strategy for dealing with images that are already being captured by mobile devices during office encounters. The process of developing the enterprise imaging strategic plan should carefully consider all of the following:
• The composition of the upcoming replacement PACS solutions.
• How to deal with DICOM as well as native data object formats.
• All of the interface issues.
• How to achieve the synchronization of the various databases in the various data repositories (i.e. IOCM).

Once the enterprise imaging strategic plan has been developed and adopted by the organization, it is never too early to begin conversations with both incumbent vendors and potential new vendors aimed at establishing a more rational pricing of their software licenses. Specifically, the value of an ordered imaging procedure such as radiology is significantly different than the value of an encounter-based imaging set such as those captured in emergency and dermatology.

Wait, if necessary, and if possible, for the right package to become available from the right vendor. That means waiting for the turnkey PACS vendor to catch up to your system expectations with those technology upgrades, or waiting for the component vendor to come up with a more consolidated best-of-breed solution.

About the author: Michael J. Gray is a consultant specializing in the digital management and distribution of medical image data, and the founder of Gray Consulting. Gray’s areas of expertise are market analysis, technology analysis, strategic planning, equipment utilization, needs assessment, workflow analysis and vendor analysis/selection.