New technologies recently have been introduced in the display device arena that are vying to overtake the role of the cathode ray tube (CRT) monitors in picture archiving and communications systems (PACS). As proof of this trend, we offer the undeniable fact that liquid crystal display (LCD) and plasma display monitors claimed the 2003 RSNA floor as the predominant type of display in the show. CRTs were few and far between. This prevalence of flat panel displays, however, does not yet exist in the field of deployed PACS, due to three primary factors:

1. Plasma and LCD monitors with high enough resolution to support PACS display requirements are relatively new technology and have not been in ready supply for PACS purchases prior to the year 2000.
2. The cost of purchasing LCD and plasma displays has been seen by many as prohibitive, as it has been typically two to three times the cost of purchasing CRTs of equal spatial resolution.
3. The technology of flat panel displays was considered too new to be deemed reliable, and, therefore, not worthy of a higher investment cost than the CRTs on the market.

This article addresses the issues associated with the financial decision to deploy LCD flat panel monitors in lieu of CRTs for PACS purposes. The use of a total cost of ownership (TCO) model is discussed, with some amplification on the issues associated with monitor deployment in the PACS environment that impact on life-cycle costs. Finally, some differences in the physical nature of LCD and CRT monitors are discussed to provide additional technical understanding of the factors in this TCO discussion.

What is Total Cost of Ownership?

Figure 1: Wall-mounted clinical workstation

Total cost of ownership is the total cost of a particular system or subsystem throughout its life cycle, from acquisition to disposal. The aim of TCO is to identify, quantify, and ultimately reduce the costs associated with ownership. The TCO models were initially developed by The Gartner Group of Stamford, Conn, in 1987 to analyze the direct and indirect cost of owning and using desktop computer hardware and software. The original model included cost categories in the general areas of:

• Capital
• Administration
• Technical Support
• End User (futz factor, formal learning, informal learning, data management, applications development, supplies, peer support)

Note that the "futz factor" is cost-associated nonbusiness use of the technology, such as web surfing or game playing. This is certainly a factor in PACS workstations built on open architecture with Internet access and direct operating system access for running common PC applications.

Another analysis by Forrester Research Inc of Cambridge, Mass, uses a different set of cost categories, namely:

• Hardware
• Software
• Training
• Management (administration, end-user downtime, coworker time, applications development, disaster prevention, disaster recovery)

Figure 2 Diagnostic viewing area with wall mounted 3MP LCD monitors

It is easy to see where each of these models can be applied to a flat panel monitor analysis. Neither, however, captures the importance of the unique cost factors in a PACS workstation environment. We prefer a model that blends categories from each of these into a combination of easily understood cost areas in the PACS environment. Our list of cost categories includes:

• Procurement (hardware, software, shipping, tax, cost of money, hardware upgrades, software upgrades)
• Installation (site preparation, mounting hardware, space allocation, cabinetry, simple network management protocol and network configuration, initial calibration, initial failure rate)
• Technical support (preventive maintenance, QC checks, calibration, record keeping, formal training, informal learning, warranty cost, management software, special tools, supplies, application development)
• Repair and replacement (reliability, spares stocks, screen repairs, back-light repairs, repair and return time, display card repairs, unscheduled monitor replacement, scheduled monitor replacement, obsolescence, ordering lead time)
• End-user impact (image quality, formal learning, informal learning, supplies, user QC operations, downtime cost, environmental factors, user fatigue, futz factor)

Why TCO analysis?

In converting from film-based radiological services to a PACS environment, a complex relationship between people, processes, and technology is affected. Complementary investments in training people, streamlining processes, and acquiring technologies that are easy to manage, service, and support will have a positive impact on the overall transition to the PACS implementation. Therefore, cost management strategies go far beyond simple reduction of direct cost. In fact, the use of the term "cost strategy" implies taking a broader view of cost than just tallying up the initial purchase price of PACS. Such a view should include the overall impact that monitors have on clinical operations.

Since PACS workstation monitors are the principal output device of PACS, they have a tremendous impact on the overall performance of a PACS in terms of optimizing potential clinical benefit. If properly planned, deployed, and maintained, display monitors can provide radiologists, and all other clinicians who look at images and related data as part of the standard of care, with timely, accurate, and reliable output in a manner that film-based operations cannot match. The benefits of PACS on clinical efficiency and effectiveness are now well documented. If display devices are improperly planned, deployed, and maintained, these same clinicians will come to distrust the PACS output, and either require other imaging resources (film) or reduce their standard of care to accommodate the lack of quality imaging services (slower diagnosis and treatment and/or incomplete basis for diagnosis). In either event, the failure of the monitors to perform adequately will be directly reflected in the inefficiencies of radiological and clinical throughput. Higher operational cost, both direct and indirect, will be the result.

Since monitors comprise a major cost in PACS implementation, it is proper that an analysis be performed on any major technology change, such as flat panel technology, which affects monitors' life-cycle costs and performance parameters.

When looking at the cost categories listed above, some points can be made about particular areas in flat panel technology that have significantly more impact than one may think when compared to CRT technology. The following is a series of expansions on issues of comparison where the life-cycle cost will differ significantly between CRT and flat panel technology. Our use of the term "flat panel" is restricted to liquid crystal display technology for the remainder of this article.

Procurement Cost

Initial capital outlay for CRT displays has been a dynamic issue since the first PACS were deployed in the late 1980s and early 1990s. Flat panel monitors came on the scene in commercially available quantities as recently as the year 2000. The following table is a comparison of the average list price of comparable CRT and LCD monitors for the past 2 years and a forecast of list prices in 2006. The table reflects the higher price for LCD over similarly specified CRTs but also reflects the shrinking difference as production of LCD technology ramps up to meet demand.

In PACS, the standard resolution for a CRT to be used for plain radiographic diagnostic reading has been a 5MP CRT. The price for the same resolution in an LCD monitor in CY 2002 was more than twice as much. At the 2003 annual meeting of the Society for Computer Applications in Radiology (SCAR), Elliot Siegel, MD, and Bruce Reiner, MD, of the VA Baltimore Medical Center reported an interesting finding relative to the equivalent resolution between LCD and CRT monitors. They reported that the image quality and diagnostic efficacy of a 3MP LCD were found to be equivalent to those of a 5MP CRT. This finding makes a good case for the LCD technology due to the fact that the price of a 3MP LCD monitor is less than 20% higher than the cost of the 5MP CRT and the difference will all but disappear between now and CY 2006.

Installation

The outward physical differences (size, shape, weight, and mounting options) between CRT and LCD monitors are not only obvious to the eye but are significant in driving the cost of converting a viewing environment from film operations to PACS operations. Not only do the size and shape of LCDs closely resemble film transluminators (view boxes), but the weight is significantly less than that of CRTs and therefore reduces the need for structural reinforcement to walls and cabinetry to support surface mounting. For example, the table below lists physical dimensions of a 5MP CRT and a 5MP LCD monitor. In a reading environment where wall-mounted view boxes are used, the conversion to PACS operations with large CRT displays will likely place such a burden on space that the number of reading locations in a room may need to be reduced to accommodate the size of the CRTs. Not so with LCD monitors, wherein the typical size of a 5MP LCD display is slightly more than a 14 x 17-inch film and no deeper than six inches. Add some additional depth to accommodate the wall-mounting hardware and cables exiting the rear of the monitor, and you are at a comfortable 10 to 12 inches from the wall in most cases and have the ability to tilt and turn the display. Figures 1 and 2 show workstation designs where viewing areas were created by using a combination of an LCD monitor and view boxes to provide a coordinated multimode viewing capability. Figure 1 is a clinical viewing area using a color flat panel LCD monitor (1.3MP, 1,280 x 1,024, 300 cd/m²), and Figure 2 is a diagnostic reading area in radiology using grayscale LCD monitors (3MP, 1,536 x 2,048, 600 cd/m²). Note the clean wall mounting and the amount of free work surface.

Site preparation and space planning in areas where many simultaneous tasks are performed are especially difficult and costly. Areas such as emergency departments, operating rooms, intensive care units, and nursing stations are typically very congested and space is a valuable commodity. If PACS displays are to meet their goal of providing display information at the right time and in the right place, flat panel monitors open up possibilities for mounting that were much more costly to consider with the larger and heavier CRT displays. Additional considerations are the ability of LCDs to be sealed for use in wet environments such as ORs, examination rooms, and trauma rooms, and to be less susceptible to magnetic field interference.

Technical Support

With image quality being a critical factor for PACS, the continuous monitoring and calibration of PACS display devices is a priority task. There are recommended standards and frequencies for QC tasks developed by the American Association of Physicists in Medicine Task Group 18. Satisfying these tasks alone is a significant cost driver due to the amount of manpower required to check, adjust, and calibrate PACS workstation displays. The average time to check a two-monitor diagnostic workstation is about 20 minutes plus travel time. An estimated time to recalibrate a two-headed diagnostic workstation is 40 minutes including travel time. These times are typical of CRT displays. The time to check and calibrate flat panel displays is reduced because the LCD display technology does not have many of the physical characteristics that tend to drift in CRT technology. For example, geometric distortion, focus, and raster size checks and adjustments are not needed for an LCD monitor. Additionally, the limited experience with LCD displays indicates that the luminance level does not drift as fast as it does with CRT technologies. These checks and adjustments are time-consuming to make, and in some cases, if they are out of tolerance, there is no adjustment and, in the case of a CRT, replacement would be required. Such a failure-based replacement need not be considered with LCD displays, as these technical parameters are not present.

In recent years, improvements have been developed in service tools for streamlining the grayscale monitoring and adjustment on both CRTs and LCD monitors. Several monitor vendors have introduced products aimed at providing automatic adjustment of brightness levels and grayscale calibration, enabling remote access to monitors and display cards via a network connection, and centralizing the management of as many display parameters as possible. Service methods that utilize these types of tools will certainly prove to be worthwhile in terms of maintaining display quality at a lower cost.

Repair and Replacement

Cost associated with the repair and replacement of PACS displays can be divided into three categories: (1) Simple repairs, (2) Repair and return, (3) Replacement. These categories apply to both the display card and the monitor of a PACS workstation. Simple repairs are usually restricted to the adjustment of or replacement of user replaceable parts, such as cables, knobs, or plastic covers. Repair and return services are usually performed under the terms of the warranty or service contract and require the removal of the broken device until it returns from the manufacturer's repair depot. In this case a loaner may or may not be provided as part of the transaction. The same goes for the outright replacement of a monitor or display card. The workstation will be without that device until either a newly purchased replacement arrives or a spare is pulled from local stocks and installed in its place. In the case of the latter, the replacement transaction goes on in the background while the workstation is operated with the spare component installed.

Two major technological differences in the construction of CRT and LCD monitors have impact on the primary cost of monitor ownership.

1. In LCD technology, the backlight outlasts the typical CRT cathode/phosphor. (Some LCD vendors are currently offering a 5-year backlight warranty option.)
2. The LCD backlight is replaceable at a cost of $500 to $1,000, whereas the replacement of the CRT envelope costs nearly as much as a new monitor. A TOC analysis for LCD vs CRT should take these differences into account.

In the case where a spare is pulled from local stocks, the cost of ownership must include the overall cost of such spares and their management. An alternative method is to contract with the monitor vendor to maintain a hot spare in their inventory for rapid shipment to your site, a sort of spare monitor insurance policy. While the first costs more than the other, it produces a replacement with less downtime. These additional levels of cost should be considered for each individual monitor type included in the PACS inventory. There may be several workstation locations that may be deemed critical enough to require such support. Examples would include the following workstations: radiology diagnostic, OR, emergency department, ICU, and QC workstations.

User Downtime

While an indirect cost, user downtime carries the most potential for enterprise cost impact for PACS monitors. This cost category takes into account the increased cost incurred by the enterprise due to the reduced throughput of radiology and clinical areas supported by the PACS workstations and the risk to clinical care, which may have cost associated with it. Examples of these include:

1. Reduced throughput for one or more radiologists for the period of the downtime.
2. Increased OR time for cases run in an OR with an inoperative workstation.
3. Slow patient processing in the emergency department due to nonfunctional PACS workstations.
4. Radiologic technologists are busy delivering laser-printed films to the emergency department, causing delays in getting morning ICU chest films done in preparation for morning rounds.
5. Radiologic technologists are catching up on ICU chest films instead of imaging early morning patients for the day's orthopedic clinic visits.

The cost of this category can outweigh all other costs associated with ensuring a high degree of uptime for PACS workstations and, more specifically, their monitors. Any technologies and processes that reduce user downtime will have a strong return on investment if the impact to clinical throughput is completely evaluated.

Conclusion

The discussion of total cost of ownership for LCD monitors must revert back to a discussion of PACS and its overall required functionality and impact on clinical operations. PACS can be described as consisting of several functional layers: a database and storage layer, an image acquisition layer, a distribution layer, and an output and display layer. These layers come together in a variety of architectural formats with the common goal of supporting clinical image management functions. No matter what the architecture of the PACS, filmless operations and simultaneous display of images and data are usually the goals of PACS, and in this context, the need to have a well-defined and fully functional output and display layer is critical. Failure to satisfy the requirements of adequate output and display of the stored images and related data creates either a general or a local failure of the PACS to perform its basic duties in support of clinical activities.

A TOC model for PACS monitors must take into account as many cost-related issues as can be described, either quantitatively or qualitatively. Several cost categories can be examined in order to develop the total cost of ownership for LCD monitors and to expose their potential cost advantage over CRT technology. These include:

• Procurement cost
• Installation cost
• Technical support cost
• Repair and replacement cost
• User downtime cost

In these cost categories, there is a combination of direct and indirect costs, which must be accounted for and which will extend the period of cost consideration from the point of acquisition through to the point of replacement. Additionally, the cost evaluation will include factors that will reinforce the business evaluation of the PACS itself. The display device is, after all, only the final component in a long line of components and operations that come together to enable a major transformation of the health care delivery system. It is this level of truly strategic thinking that is required if an effective total cost of ownership analysis on flat panel technology is to be performed. Remember the words of Ben Franklin: "For want of a nail, the shoe was lost; for want of a shoe, the horse was lost; and for want of a horse, the rider was lost&."

John Romlein, MSE, is vice president, eastern region, Xtria Healthcare, Digital Solutions, Frederick, Md

John C. Weiser, PhD, is chief scientist and principal, Xtria Healthcare, Digital Solutions, Frederick, Md

References:

1. Assessment of Display Performance for Medical Imaging Systems. Report of Task Group 18, American Association of Physicists in Medicine.
2. Fetterly K, Hangiandreou NJ, Langer SG. Monitor QC. Decisions in Imaging Economics. 2003;16(4):46-52.