During the past few years, the practice of diagnostic imaging in the United States has experienced an inexorable trend toward managed care including capitation, the emergence of integrated healthcare delivery networks (IHDNs), and the integration of radiology and nuclear medicine into the healthcare enterprise. For better or worse, this transition has moved the practice of diagnostic imaging closer to the model that has been utilized by the Department of Veterans Affairs (VA).

One of the ways in which the VA has responded to Vice President Al Gore's challenge to "reinvent government" has been to move rapidly towards the complete electronic medical record (EMR). The adoption of this comprehensive EMR would permit health care workers to have easy and instant access to all forms of patient information including demographics, laboratory results, admission data, progress and discharge notes, and vital signs, as well as multimedia data such as patient EKGs, EEGs and audiology studies. The EMR also incorporates images obtained throughout the healthcare enterprise, including pathology slides, intraoperative photographs, GI endoscopy, bronchoscopy, and so on. Diagnostic imaging studies, including those obtained in radiology, nuclear medicine, cardiology, and the vascular laboratory, also are considered a part of the patient's electronic medical record.

The goal of this transition to the EMR is to increase productivity while improving the quality of patient care. Our last 5 years of experience with filmless radiology in the Baltimore VA Medical Center and our 3-year experience with the multifacility VA Maryland Health Care System have allowed us to evaluate the relative economic costs and benefits of an enterprise- and network-wide PACS.

The plan to replace an aging suburban VA Medical Center with an ultramodern high-tech medical center, located on the University of Maryland Campus, provided us with a unique opportunity: If we could find a vendor who could successfully respond to our specifications for a viable commercial PACS, we would be in charge of designing the hospital and radiology department prospectively to permit us to operate the first filmless imaging department in the United States.

An extensive survey of the technology available in the late 1980s and early 1990s convinced us that, despite optimistic predictions, there was no available system in operation that could meet our requirements. We subsequently wrote an extensive Request for Proposal (RFP), providing functional specifications that, if met, could provide a viable system. We decided that if a system could not be designed to our specifications, we would drop our plans and open the new hospital with a conventional imaging department.

Fortunately, the Department of Defense (DoD) and the Hammersmith Hospital in London had plans to purchase a PACS the same year, with a less "ambitious" (i.e. less crazy) implementation schedule. The three groups (DoD, Hammersmith Hospital, and VA) constituted a large enough critical mass of interest in filmless radiology to make it attractive (but not, as it turned out, profitable) for Siemens and Loral (known best as a DoD contractor) to join forces and offer a clinically acceptable system. The system, which has subsequently been taken over by General Electric Medical Systems, offers a "centralized architecture" in which images are stored using a high speed central server and are routed "on demand" to video memory on the Macintosh® workstations rather than being stored on the workstation hard drives.

The PACS was purchased in September of 1991 for $7.8 million dollars-approximately $7.0 million for the PACS and approximately $800,000 for computed radiography. The system was installed and first became operational in June 1993, approximately

5 months after the Baltimore VA Medical Center opened.

In addition to the importance of proving clinical efficacy of filmless or

soft-copy operation, one of the most significant questions was whether filmless operation could indeed be cost effective using the "state of the art" PACS technology. Our best guess at the time of initial operation was that a soft-copy department would be somewhat more expensive to operate than a conventional one, and that radiologists (though perhaps not technologists) would be slightly (10 to 20%) less productive using this new technology. However, as we discuss in the next several paragraphs, both of these guesses were proved incorrect.

Although a few well written articles were available in the literature, these generally were based upon expected or potential savings with PACS or were extrapolated from the use of the system in a very limited environment, such as the intensive care unit. Thus, there was little information available to guide our expectations about the cost effectiveness of PACS in an almost completely filmless environment.

Baseline data were collected from the hospital and radiology information systems, supply orders, personnel records, equipment procurement records, and other sources both before and after the transition to "filmless" operation. A substantial increase in volume was attained, from approximately 34,000 studies per year prior to filmless operation to the current more than 60,000 studies per year. This included a two-fold increase in the volume of CT and angiographic studies and the addition of MRI, mammography, and bone densitometry, which were not available at the old hospital.

Despite this substantial increase in the volume of studies, the number of attending radiologists only increased from 4.5 to 5.5; the additional radiologist position was used to provide coverage for MRI and mammography. There was no significant increase in the number of technologists used, despite the increase in the number of studies and the complexity of the studies. The number of general radiographic studies decreased from approximately 74% to approximately 67%, resulting in an even larger increase in the weighted work units in the department.

Costs

One major cost of the system is the depreciation of the initial expenditure on the expensive equipment. With regard to the best figure for capital depreciation, the Department of Veterans Affairs uses an 8.8-year depreciation on medical equipment, while computer equipment is often depreciated using a 5-year time period. Another significant contributor to the cost of the PACS is the service contract, which is currently approximately $500,000 per year. No additional service personnel were required other than those provided by our vendor for the service contract.

Savings

Savings in the radiology department were achieved in a number of different ways. Current film costs at our institution are less than $10,000 and includes the cost to print mammograms, to print films for patients who are sent to other hospitals and outpatient departments, and to print films for conferences that are held at the University of Maryland. This is only 5% of the estimated $200,000 that would have been required to support our current volume had we stayed in a film -based environment.

The combination of the use of computed radiography and soft-copy interpretation has resulted in a major decrease in the need to retake images. These rates have decreased from approximately 5% to 0.7%, an 86% reduction, resulting in estimated savings of $30,000 per year.

Other decreases have been related to the almost complete elimination of film including a decreased need for film folders and a decreased need for film chemistry. These savings are relatively modest, totalling approximately $15,000 per year.

One of the greatest, and to some extent unexpected savings associated with PACS has been the ability to recover space in the radiology department. The current film file room occupies approximately 2,500 square feet and is divided into two halves. The back half is used for the PACS computer room (although this is more space than is actually required). The front half stores films from the 6 years of operation prior to June 1993 and a limited number of new films, such as mammograms generated during the past few years. Using relatively conservative figures, construction costs have been estimated at $180/square foot for the front half of the file room, and $20/square foot/month. Consequently, recovery of this space for other purposes, such as an MRI scanner and offices, would avert $230,000 in construction costs and $307,200 per year in space costs.

The biggest savings associated with PACS at the Baltimore VA Medical Center have been in personnel costs. Given the 76% increase in volume, we have conservatively estimated that at least two additional radiologists would have been required at Baltimore, even assuming a major increase in productivity per radiologist. Additionally, we estimate that, using national standards for technologist productivity, we would have required three to four additional technologists to keep up with the volume. In our case, there was an elimination of all but one of the film room clerks required to maintain the film library and transport film throughout the medical center. We believe that this has probably averted over $100,000 per year in costs as well.

Cost benefit analysis for the departments of radiology and nuclear medicine

Using relatively conservative estimates of cost savings from the data described above and a depreciation time frame of 5 years, we estimate that the additional cost of PACS is approximately equal to the savings generated by the system. This conclusion has been supported by a recent study of the PACS at the Baltimore VAMC performed by a group of investigators from Johns Hopkins. This determination is consistent with a number of authors who have suggested that, although PACS may pay for itself, it is unlikely to result in considerable savings in the imaging department itself. According to these data, our initial guess that filmless operation in radiology and nuclear medicine would be more expensive than a conventional film based operation was, in fact, incorrect.

Economics of PACS for the entire hospital

The hospital-wide savings associated with filmless operation are more difficult to quantify. The average clinician estimates that he or she saves approximately 50 to 70 minutes per day because of the improved image accessibility associated with the PACS. Our own estimates are that they probably save a more conservative 10 to 12 minutes per day. However, when applied across the entire institution, even these more conservative estimates result in savings of approximately two to four clinician FTE (full time equivalent) positions per year, representing a savings of between $200,000 and $500,000 per year just in clinician time. A number of other, even harder to quantify cost benefits are associated with the PACS. These include decreased waiting times for radiology reports (typed report turnaround has decreased from 24 hours to 2 hours), which has the potential to impact on length of stay, increased clinician efficiency and accuracy in patient care, decreased medicolegal risks, and savings associated with a decreased rate of lost studies. Ignoring these latter

benefits, the savings of $200,000 to $500,000 are still considerable.

Economic benefits for a network of hospitals

As of June 1995, the four Department of Veterans Affairs hospitals in Maryland have been combined into a single IHDN: the VA Maryland Health Care System. These four hospitals-Baltimore, Ft. Howard, Perry Point, and the Rehabilitation and Extended Care Facility-all share a single hospital director, Hospital Information System (HIS) and Radiology Information System (RIS). The imaging departments also have been combined into a single "virtual" imaging department that shares the radiology/nuclear medicine commercial picture archiving and communication system, located in Baltimore. These additional facilities have added a volume of more than 33%, resulting in more than 80,000 studies per year for the combined network.

This integrated imaging service was accomplished by converting the two "higher volume" facilities to 100% use of computed radiography for general radiographic examinations. These studies, in addition to CT exams, are sent to the commercial PACS at Baltimore for archival and, in many cases, for primary interpretation, providing subspecialist expertise to supplement the general radiologists located at those sites. Pooling of imaging studies in Baltimore has permitted a decrease in radiologist coverage by 50% at those outlying facilities without an increase in turnaround times for reports. Patients can be more easily triaged prior to referral to the tertiary referral center (Baltimore), and when those patients are transported, their images can be reviewed prior to their arrival. In this scenario, the use of the PACS has resulted in a significantly improved ability to provide back-up and subspecialty coverage for the radiologists at outlying facilities.

The resources required to add an additional facility to the PACS network have been relatively small; costs are less than 20% of the cost of a separate system at those facilities. The major costs thus far have been in the acquisition of computed radiography equipment which, in addition to the networking equipment, represents an investment of approximately $200,000 to $250,000 per institution. The cost savings at those institutions are estimated at approximately $200,000 per year, despite the fact that they have not yet made the transition to filmless operation.

VISTA imaging

The Veterans Affairs Hospital Information System Technology Architecture (VISTA) PACS is currently utilized within the VA Maryland Health Care System for non-radiology/nuclear medicine images such as pathology, GI endoscopy, and dermatology. However given the fact that it uses standard Windows 95 and Windows NT workstations which are already deployed throughout the enterprise, it has the potential to play a major role in the distribution of imaging studies throughout the enterprise. This is particularly true for those facilities outside Baltimore which do not have the commercial PACS. In these medical centers, the VISTA imaging system has the potential to provide most or all of the functions of the commercial PACS to the clinicians using the affordable desktop PCs already being used for office automation and the electronic medical record.

The VA's VISTA imaging system developers are currently working on software and hardware systems that are optimized to help the radiologist to make primary diagnoses. These two to four monitor, 2,000 pixel workstations, along with the new software designed for radiologists, are being tested in the Wilmington, Baltimore, and Washington D.C. VA Medical Centers. The potential advantages of such an internally developed PACS would include an extraordinarily high level of integration with the remainder of the electronic medical record, as well as relatively low costs for the software and minimal "mark-up" on the required hardware costs. The biggest challenges of such an internally developed PACS, however, will be in the maintenance and support of the system and in the development of a system that is sophisticated enough to meet the demanding radiologist requirements for speed and accessibility.

Digital imaging and the Department of Veterans Affairs

The VA has, within the past few years, been reorganized into 22 geographic networks throughout the country, known as Veterans Integrated Service Networks (VISNs). These networks are in a particularly good position to take advantage of the potential benefits associated with digital radiology. One major factor is that VA facilities are considered to be located on federal rather than state property. This permits sharing of images across state lines without worrying about issues of state licensure. The medical centers within the VISNs are being encouraged to minimize duplication of services. In our VISN, the Baltimore VA Medical Center has a particularly strong neurosurgery and neuroradiology program, while the Washington D.C. VA Medical Center has a strong cardiac surgery program. These individual strengths encourage the development of "centers of excellence" within each VISN; these centers are able to share patients and their electronic medical records among the facilities. Teleradiology and PACS technologies have been utilized to increase image accessibility and to maintain or improve the quality of care while keeping costs low. Short-term plans of many of the VISNs include a major acquisition of digital imaging equipment, specifically large scale teleradiology and PAC systems.

In the medium- and long-term, the VA will continue its trend toward the use of teleradiology and PACS to improve image accessibility and access to expert radiologic services. Although sharing between VISNs is currently minimal, it will likely increase in the future. In order to prepare for this, the Department of Veterans Affairs will need to formulate a plan to maximize the intraoperability of the various commercial PAC systems. During the next few years, the role of the VISTA imaging system and commercial systems will become more clear. It is likely that these systems will begin to work synergistically to combine and utilize the relative strengths of the commercial and internal VA systems. It is also likely that there will be greater sharing of radiology services and digital imaging equipment among the various government health care organizations, including the VA, the Department of Defense,

the Indian Health Service, the National Institutes of Health, the Immigration

and Naturalization Service, and the Federal Prison System. PACS and teleradiology technologies, combined with an improved national network infrastructure, will result in continued reinvention of the practice of diagnostic imaging within the Federal Health Systems. This will undoubtedly lead the way towards revolutionizing the practice of radiology and nuclear medicine in general.

The promising results at the Baltimore VA Medical Center and in the VA Maryland Health Care System need to be further studied, and other large scale PACS projects that are hospital, outpatient, and network based in both academic and community settings need to be analyzed to determine the relative costs and benefits of the technology. AR

Dr. Siegel is Director of Imaging for the VA Maryland Healthcare System in Baltimore, MD. Both Dr. Siegel and Dr. Reiner are at the University of Maryland School of Medicine in Baltimore, MD.