Over the past 20 years, picture archiving and communications system (PACS) development has been predominantly radiology-centric. This is not surprising since the needs of the radiology department were the driving force in purchasing most PACS and there have been numerous technical issues to address prior to tackling the broader issues of enterprise-wide image distribution.

In order to achieve even a minimal level of functionality, those who deployed a PACS had to deal with many implementation and performance problems. These included modality interfaces, radiology information system (RIS) integration, archive architecture, display configurations, lack of industry standards, work-flow issues, functional requirements, and the limited capabilities of computer technology. Meanwhile, the PACS industry had to cope with changing operating platforms and electronic storage options, evolving networking technologies, and dramatic changes in the corporate merger and acquisition game.

With so many issues and changes, PACS developers had to focus on solving the issues inside the radiology department. To an outsider, it may have appeared that radiology functioned as an island and the benefits from the sizable PACS investment were largely confined to radiology. Some benefits were realized in the emergency department, ICU, OR, and other image-access areas of the hospital, but the most significant benefits were in the areas of improved radiologist and technologist productivity, some film savings, and easier access to images stored in the local archive. PACS at that time were difficult to cost-justify and both users and developers had to struggle to improve their systems' performance and functionality.

The state of technology development was partially responsible for PACS' limited impact on the extended clinical environment. Although it was recognized that medical images were needed well beyond the radiology department and hospital clinical areas, there was no cost-effective way to distribute electronic images to the broader medical community. Consequently, radiology continued printing film to service patients and referring physicians.

A factor that reinforced the perception of PACS as a department-only issue was a lack of an institutional or enterprise vision. Other information systems had been deployed in hospitals, but they too had been department-centric. Few hospitals had a coherent information systems strategy.

The PACS vendors had a hand in this approach also. A majority of the early entrants to the PACS marketplace were companies already providing products and services to radiology that saw PACS as an opportunity to expand or protect their primary businesses. PACS were purchased much like imaging modalities, with similar types of terms and conditions and service contracts. So in many respects, vendors accepted the narrow view of PACS and proceeded to develop stand-alone radiology solutions. PACS solutions had rules-based routing capabilities, and proprietary compression algorithms, and were not easily scaled or upgradeable. They did not provide, moreover, for wide-area distribution of images.

PACS EXTEND THEIR REACH

In more recent years, PACS have evolved rapidly, keeping pace with the speed of technology change. PACS have become less expensive, more functional, and more standardized from a hardware and software perspective. PACS also have extended their reach with electronic distribution of images to the world beyond radiology. This has caused a rethinking about PACS, and a realization that PACS are a significant health care issue that can have a positive impact on the entire health care delivery system. Considering the role of the medical image in the treatment and management of patients, PACS have the potential to improve the timeliness and efficiency of diagnosis, treatment planning, therapy, and outcomes assessment.

Again, many factors have contributed to this evolution of thought. Many information technology (IT) departments have upgraded their institution's network infrastructure, providing the necessary bandwidth, fault tolerance, redundancy, and security to support an enterprise PACS. IT also has begun to take responsibility for the central archiving of all DICOM medical images from multiple departments. Industry standards (DICOM, HL7) have evolved to a point where systems are attaining a higher level of interoperability. Now, most PACS offerings can service the entire hospital with a high degree of reliability. All of these factors have significantly improved the hard and soft benefits of PACS to the entire institution.

Despite these advances, however, a real-time enterprise radiology service, or the capability to distribute and access medical images across the span of the extended medical community, continues to elude many health care organizations. An institution will make a much greater impact on film usage and film-related costs, turnaround times, and quality of service to referring physicians and patients if it has a wide-area image distribution solution. Without such a solution, the full benefits of PACS cannot be attained.

CRITICAL MASS

Electronic wide-area image distribution requires an integrated global network operating on low-cost computing platforms with the same base circuit and memory technologies that process and store information. With appropriate interfaces to users and the physical world, such a network has the potential to reduce film (and paper) usage dramatically. Greater use and acceptance of the Internet and web technology have improved the ability to inform and educate in the health care arena. As this trend continues, one of the most important benefits to health care will be a breakthrough in the ability to communicate remotely in a real-time and interactive manner.

There are two primary laws that will make the cyberspace connection feasible. The first is the familiar Moore's Law, which states that the number of transistors on a microchip will approximately double every one and a half years. A simple translation of this is that computers will get faster, smaller, and cheaper over time. We have all witnessed this happening, and industry experts project that this phenomenon will continue for at least another decade.

Table 1. Three generations of computing.

The second law, which is probably more significant in attaining the goal of wide-area image distribution, is Metcalf's Law. Metcalf's Law states that the total value of a network is proportional to the square of the number of subscribers, while the value to the subscriber is proportional to the number of subscribers. To illustrate this law, consider the fax machine, which was invented in the 1940s. Why did it not become popular and widely used until the 1980s? The answer is Metcalf's Law. What good is a fax machine if the sender is the only one who has one? Distribution assumes that other entities can receive what you are sending. Therefore, a minimum critical mass of communications technology is necessary to make wide-area image distribution practicable.

The World Wide Web offers the most potential for change at all levels of health care through standardization and universal access. This will include online information, access to reports and images, linking human and machine-created information, and computers that are body-networked through the use of body sensors that will transform analog body waveforms into digital signals, enabling real-time onboard monitoring. When a PACS and the Internet attain interoperability and the number of access points exceeds the minimal critical mass (in this case, the number of computers and workstations needed to eliminate current manual/film distribution operations), then the requirements for wide-area medical image distribution will be satisfied. Note that interoperability is another key requirement of such a system.

WIDE-AREA DISTRIBUTION

Once the stage has been set for the electronic wide-area distribution of images, each institution must determine which technology best meets specific requirements. Each institution must assess the options that are available to them, their particular barriers to implementation, whether an electronic system can be cost-justified, and, if so, how they get started.

These questions can be answered through an analysis of the institution's wide-area distribution needs and a review of the technology currently available. The institution should consider Internet-based and alternative technologies that can provide an optimal or complementary solution to electronic wide-area distribution. Alternate/complementary technologies may include CD burners, paper printers, and old-fashioned manual distribution. Although this article focuses on the electronic management and distribution of medical images and the issues associated with implementing these systems, it is important to recognize that these other options exist and to understand where they might fit or be added appropriately in the technology mix.

In brief, CD burners offer an inexpensive and effective method for wide-area image distribution. Some institutions use CDs as marketing tools where they print the referring physician's name and logo on the disc along with patient and study identification. The image data saved on the CD also include the viewing application so any Pentium® 3 or higher PC can load the study and application and view the study in soft copy. CD burners range from a few hundred dollars up to $45,000 for an automated, robotic system with attached image workstation.

Paper printers have become a widely accepted method of providing images to referring physicians and patients. There are both black-and-white and color printers. Physicians can make annotations on the paper images and save them in their files. There are a variety of printers available, ranging from a few hundred dollars to more than $18,000.

In some cases, manual/film distribution remains the best method for wide-area image management. When procedure volumes are very low and there is poor or no Internet access and/or a low incidence of PCs, manual/film distribution may still be the most cost-effective solution.

ASSESSING THE NEED

It is a well-accepted principle that no two hospitals are exactly alike. That principle holds even stronger when considering the medical imaging service and the extended medical community it supports. Therefore, it is important to assess where and how images are being distributed today and where they will be needed in the future. This assessment will direct you toward decisions and solutions, and should cover three major categories: the current method of wide-area distribution of medical images, current and near-future PACS plans, and demographics of the wide area you are servicing (including geographic, technology infrastructure, and computer literacy characteristics).

The current method of delivery of images to referring physicians generally falls into one or a combination of the following scenarios:

• The hospital sends films routinely by mail or courier to all referring physicians or only to physicians requesting film.
• The hospital makes films available to the physician at the radiology file room.
• Patients carry a copy of the study by hand to the physician's office.

Most hospitals use a combination of these methods, which creates a tremendous management and tracking problem for the radiology department. These methods are slow, prone to lost or misplaced film, and are manpower heavy. When cost and time factors are analyzed, they add up to a significant cost per transaction and to a slow and cumbersome imaging service.

In many cases, the cost of implementing a web-based distribution system can be justified by reducing or eliminating the costs of the manual/film-based service. To determine this, each location has to calculate the cost of its current wide-area image distribution system. Analyzing the costs of any of these methods requires a detailed assessment of each incremental step. To complete a requested transaction, the steps include people, equipment, supplies, and services. The cost factors are: film and film processing costs, duplicate filming costs, courier or transportation costs, and labor associated with tracking and logging films when they leave the department, when they are returned, and when they must be followed up on when not returned.

In addition, there are film-handling costs when file room staff have to locate a study for retrieval and refile the study when returned. This process can be complicated when file room staff cannot find films (lost films are typically 5% to 7% of all studies requested) or when films are difficult to locate (in as much as 30% to 50% of all requests, the search will last between 20 and 90 minutes).

In some cases, films are retrieved and copied for release. The retrieval and refile costs are added to labor costs, processor and maintenance costs, and distribution costs associated with the duplicate film.

Depending on circumstances, transportation or courier expenses can vary widely. It is easy to see how a hospital supporting hundreds or thousands of referring physicians, multiple imaging centers, and health clinics can be spending considerable amounts of money on a system that is slow, inefficient, and operating by the last century's standards.

Figure 1. Schematic of the data flow in an example of an electronic wide-area distribution system.

Another major consideration not mentioned so far is the cost to referring physicians. This is the time it takes them to access films in the hospital. Typically, a physician will walk to the file room, wait in line, review the study, and walk back to the clinical area, several times a day. The cost in physician time ranges from $60 to $80 per study. Multiply this by the number of studies reviewed each day, times the number of days in the hospital per year and you have calculated a huge cost. Since most physicians are fee-for-service, hospital financial people usually do not consider this a cost. However, every hospital is working to improve services to physicians to attract more patients and delays of this type have a negative impact on the physician's productivity.

MANUAL DISTRIBUTION COST MODELA

To demonstrate the costs associated with the manual wide-area distribution of images, the chart below presents some typical costs. These numbers are averages from a number of in-house surveys and published cost analyses in the United States. This cost model is not specific to any hospital but it clearly represents the concept.

Since a radiological service utilizes multiple modalities with a varying modality volume mix, no two departments are alike. In addition, they use varying amounts of film and film size per study. For this example, we will use an average of $7.50 for the cost of film, chemicals, and jackets per study for an all analog department.

Modalities and volumes used in cost model.

The modalities and volumes in our model are as follows: 

Table 2 below presents the costs associated with a radiology service that is all manual/film-based, utilizing couriers to distribute films. In this example, all studies are sent to the referring physician and assumptions are made regarding the demographics of the wide-area. The object here is to demonstrate the potential cost of operating a manual/film-based wide-area image distribution service.

Table 2: Costs associated with a film-based wide-area image distribution system.

IS PACS A PREREQUISITE?

When assessing the need for an electronic wide-area distribution system, the question arises, is a PACS requisite to begin distributing images over the wide-area electronically? Today the answer is no. There are products and fee-for-use services on the market that provide a fully functioning electronic distribution system. Many hospitals are starting to distribute to the wide-area prior to implementing an in-house PACS. The strategy here is to demonstrate the benefits of electronic image management, to generate support and enthusiasm for managing images electronically, and to cut costs.

However, there are some downsides to this approach. At the most basic level of electronic distribution of images, all studies distributed electronically must be scanned into the system by a file room full-time equivalent (FTE) with a film digitizer. This basic system operates much like a teleradiology system, except that it sends scanned images to a web server for wide-area distribution.

This method is slow and still requires study retrieval and refiling. Another disadvantage is that it does not populate an archive with digital images for future retrieval and comparison. In this case, a radiology department would realize savings in duplicating the film and eliminating courier costs, but most of the other costs would remain.

At this primary level, there would be quality control and security issues as well. Without a radiology information system interface and a database manager to control the distribution and storage of studies, the system would have to rely on the capabilities of the web server to manage the distribution service. Careful evaluation of the prospective system is essential.

A second level configuration would significantly improve the functionality of the service but will be more expensive to implement. In this case, the department will need an electronic archive, RIS interface, film digitizer, technologist's workstation, radiologist's workstation, and interfaces to the modalities. This system  will perform faster, easier, and with greater assurance of security. It also has greater potential for savings because it can eliminate many of the costs associated with the film-based system.

DEMOGRAPHICS OF SERVICE AREA

The last major element in evaluating and understanding the costs of an electronic wide-area distribution system is the demographics of the wide area itself. The hospital must assess the number of referring physicians within the service area, the number of films currently being distributed, the current method of film distribution, the availability of network infrastructure and bandwidth, and the accessibility of PCs in the physicians' offices. There must be a significant benefit to the physician for him to change his method of image access and display.

Internal hospital resources must also be evaluated. Does the hospital have the marketing capabilities to launch and the expertise to implement an electronic wide-area image distribution system service? Can radiology or the information services department support a level one or two configuration?Are file room FTEs willing or capable of operating a film-digitizing service? Does an electronic wide-area distribution system fit into the hospital's strategy for growth and development? Will the selected system function with the present or future PACS?

CONCLUSION

Electronic wide-area distribution systems present a significant opportunity to improve the radiology service to referring physicians and to reduce operating costs. However, as in any high-tech acquisition and implementation, great care and investigation are necessary to achieve the results anticipated. The same adage applies to electronic, wide-area distribution systems as in PACS: "Specify, specify, and specify." Development of an unambiguous technical specification and minimum performance criteria prior to contracting will help foster a relationship of partnership and cooperation with the system provider and lead to greater success in realizing the benefits of electronic imaging.

Gary Reed is president of a PACS consulting company located in Lebanon, NJ, gary@irpacs.com.