Dr. Siegel is the Director of Imaging at the VA Maryland Healthcare System (VAMHCS) and an Associate Professor at the University of Maryland School of Medicine; Dr. Reiner is the Director of Radiology Research at VAMHCS and an Associate Professor at the University of Maryland School of Medicine, Baltimore, MD.

In the current era of dwindling reimbursement, undersupply of diagnostic radiologists, and increasing volume of studies, imaging departments are asking radiologists to interpret more studies than ever before. The use of picture archival and communication systems (PACS) has resulted in improved departmental efficiency in most cases, especially when associated with the re-engineering of departmental workflow. Technological development, such as improved network infrastructure and speed, faster workstations with more reliable and brighter monitors, improved image presentation and navigation software, image enhancement, computer-aided diagnosis, and integrated speech recognition have received a good deal of attention in the research community as candidates for improved radiologist efficiency and productivity.

However, surprisingly little attention has been paid to the potential of improvements in radiology reading room design as a means to enhance the performance of radiologists. ¹ Experience and research performed at our facility, the Baltimore VA Medical Center, suggest that relatively small investments in room design and workstation ergonomics can result in major gains in productivity and accuracy with a concomitant decrease in radiologist fatigue.

The Baltimore VA Medical Center opened in 1993 as a replacement facility for an older hospital previously located in the northern part of the city. Although the hospital was designed originally as a high-technology replacement medical center with an emphasis on the infrastructure to support digital imaging and a paperless electronic medical record, the concept of filmless operation came relatively late in the planning phase for the facility. The government rules concerning hospital design and construction precluded "last-minute" changes to the departmental or reading room design, despite the fact that we knew that we were purchasing an enterprise-wide PACS for a conventionally designed department. This resulted in the need to place 5 four-monitor diagnostic imaging workstations in a reading room that was originally intended for film.

Our PACS has been in operation for almost 9 years and radiologists have been reading using soft-copy interpretation for the past 8.5 years. This means that we have the dubious distinction of having more experience not only with soft-copy interpretation but also with the consequences of a poor PACS reading room design than any other facility in the United States. Our experience has allowed us to share with visitors and others how to and, perhaps more importantly, how not to design a filmless reading room.

Room layout

The initial design for the radiologist reading room called for a single large reading area with workstations for four or five radiologists to interpret primarily the conventional radiographic studies performed in the department. The room was located adjacent to a large area designated as the "file room." This design was modeled after most "general read" areas in other imaging departments (figure 1). A large central reading room is typically necessary in a film-based environment to accommodate not only the radiologists interpreting films but also the film room carts brought with batches of studies to be interpreted and the many clinicians who "round" in the radiology reading room using light boxes or film alternators. No partitions were used to isolate the radiologists from each other since glare from overhead lights or adjacent view boxes is typically not a problem in a conventional reading room.

However, in a filmless, soft-copy reading environment, a large central reading room for general radiography is no longer necessary. Radiologists can read from any location in the department or potentially anywhere in the hospital (or outpatient center), since images are available or can be retrieved to any location on the PACS network. The other traditional reason for a large central reading area was to serve as a central location for radiographic consultations with the clinicians. However, as shown in figure 2, the consultation rate for general radiographic examinations fell from 1 consultation per 7.6 examinations to 1 in 42 (and has continued to decrease since we published the study). ²

In our department, this dramatic decrease in one-on-one, in-person consultations has also altered the requirements for a central congregating place for clinicians and consultations. This change has also been confirmed by other facilities with filmless imaging departments. As the number of consultations has decreased, there is even less need for a large, centralized reading room. In a soft-copy reading environment, consultations are much more likely to take place via phone, e-mail, annotation of images by radiologists, or via fast report turnaround times. Fortunately, report turnaround has dropped considerably, to the point where studies from the ER, for example, are often interpreted prior to the patient returning to the emergency room from the radiology department. Consequently, the radiology report itself, rather than an "in-person" consultation, becomes the means of communication of the imaging findings. Room partitions may be unwieldy for clinician traffic in a
conventional radiology department, but they can be very helpful in reducing unwanted glare and noise in a filmless room since clinicians visit less often.

Our current reading room environment is still a hybrid that features both light boxes for film and PACS workstations for soft-copy interpretation for each radiologist. This is despite the fact that old films from our own hospital are never reviewed now that we have been filmless for almost 9 years, and the fact that films from outside institutions are infrequently submitted for evaluation. Our new room design will include only a single viewbox to be shared by all radiologists.

Although we originally utilized four-monitor workstations to emulate film alternators more closely, and our radiologists who interpret conventional radiographic images prefer four monitors, our research has demonstrated that the use of two-monitor workstations results in comparable radiologist efficiency when compared with a four-monitor configuration (figure 3).

Room lighting

Background room lighting, which was thought to be relatively unimportant in the original design for a film-based reading area, becomes critically important in a soft-copy environment. This is due to the very low levels of light associated with a typical high-resolution 5 megapixel (2,000 by 2,500 pixel) PACS monitor, which has an output typically in the range of 60 to 70 foot-lamberts. This figure is approximately 1/10 of the light that is associated with conventional lightboxes, which can range between 500 and 1,000 foot lamberts. We have performed studies previously that have documented the importance of an optimal balance between monitor light and background ambient room lighting. ³ Decreased radiologist productivity (increased interpretation times), decreased accuracy, and increased fatigue levels were seen when using monitors that were less bright when compared with those with higher luminance. Other studies performed in our reading room have documented that ambient room lighting is also very important in radiologist performance. For example, the use of window/level workstation tools increased from 45% to 72% to 91% as the background light levels in our reading room changed from off, to half on, to completely on (using overhead fluorescent lights). Fatigue levels increased dramatically as background light levels increased as well (figure 4). As was the case with decreased monitor luminance, higher ambient light levels were also associated with significantly decreased interpretation accuracy. These studies underscore the importance of striking a balance between ambient room lighting and monitor brightness. The use of newer generation, higher brightness, active-matrix LCD displays is likely to permit radiologists to once again increase the background lighting levels in reading rooms, freeing radiologists (as was the case decades ago with developments in fluoroscopy equipment) from the constraint of reading "in the dark."

The lighting in our radiology reading room currently uses overhead industrial-type fluorescent fixtures that are comparable in brightness to the film viewboxes. These lights have on/off switches located near the entrance to the room and cannot be dimmed or individually controlled by the radiologists. Additional lighting is provided by conventional light boxes that were originally intended for film display prior to the implementation of the PACS. The radiologists place old films, often hung upside down or sideways, on the light boxes, which results in a modicum of light for the radiologists. But this solution is a very poor substitute for adequate individual task lighting. The suboptimal brightness of the early PACS monitors, coupled with the lack of true task lighting in our reading room, has resulted in an increase in the number of complaints of eyestrain and fatigue comparison with that expected in a film-based environment. Other factors that might have contributed to the increased fatigue are monitor flicker, small cursor size, and the more active role required for image manipulation.

In order to improve ambient lighting, it is important to remember four primary objectives in the radiology reading environment: 1) general illumination levels for computer tasks, 2) illumination for reading tasks using localized light sources, 3) balance of brightness levels in the user's field of view, and 4) control of monitor reflection. In order to accomplish these objectives, a combination of indirect overhead lighting and local task lighting, using dimmable sources, can be used to provide maximum flexibility for each radiologist. Moveable partitions can be helpful to further fine-tune the control of local and general lighting (as well as to reduce ambient noise levels). A number of sites have repainted the walls of the reading room with dark colors in an attempt to further reduce reflected light from the walls. We are not aware of any studies that suggest a specific color or color combination to reduce fatigue and improve productivity in a radiology reading room.

Temperature and ventilation

The contribution of improved air conditioning and individual temperature and ventilation controls is typically underestimated in the design of filmless radiology reading environments. Improved air handling is especially important, in some cases, due to the high heat output of high-resolution computer monitors and workstations and because of the greater sensitivity of the PACS equipment to temperature and humidity than film and film-based viewboxes. Radiologists who are already subject to increased fatigue as a result of the transition to soft-copy interpretation may be even more vulnerable to the effects of increased heat and poor ventilation. When the Baltimore VA PACS first became operational in the summer of 1993, we quickly discovered that the reading room air conditioning and ventilation systems were overwhelmed, and temperatures exceeded 100 degrees Fahrenheit when the door to the reading area was closed. Monitor life expectancy at that time (due to a combination of increased temperature and inherently unreliable first-generation PACS monitors) was a surprisingly short 3 months and radiologists' coffee break frequency and duration soared until additional air conditioning capacity was installed. Our experience underscores the importance of adequate planning for air conditioning and ventilation in the reading rooms. Our plans for the new, redesigned reading room include not only better individual control of local and general lighting, but also individual ventilation controls similar to those available in most automobiles. We believe that a small expenditure in improved air handling will result in improvements in productivity and decreased fatigue, although we have not yet rigorously tested this in our laboratory or reading room.

Sound

Another key factor that was not considered in the initial reading room design was the impact of ambient noise. After we made the transition to a soft-copy department, we quickly learned that computer workstations generate a good deal of background noise that can be distracting during image interpretation and dictation. Although one-on-one consultations have decreased considerably, telephone communications with clinicians have increased. In a single, large reading space without carpeting or other sound-absorbing partitions, we believe background noise can have an adverse effect on radiologist fatigue and productivity. We plan to attempt to quantify the effect of background noise on radiologist performance in future studies. The recent, partial introduction of speech-recognition systems in our department has made us much more aware of distracting background sounds such as a noisy ventilation fan, the phone, overhead hospital paging system, or other radiologists dictating in the same room, all of which can decrease the accuracy of these systems. We believe that the use of acoustic dampening materials, such as carpets and sound-absorbing panels, will ameliorate some of these problems. The impact of background music, white noise, or even active sound cancellation has not been documented adequately in the radiology literature and these are consequently interesting avenues of future research in our laboratory.

Integration of information systems and ergonomic workstation design

As the hospital moves toward an electronic medical record and radiologists become increasingly reliant on computer information systems and other electronic systems, it becomes increasingly evident that these systems must be integrated. Radiologists at the Baltimore VA currently require access to the PACS workstation, the Internet and Intranet, a speech-recognition or digital-dictation system, the hospital "paperless" electronic medical record, e-mail, office software (such as word processing), and the telephone (figure 5). These functions should be able to run on a single multi-tasking workstation that should be designed to allow easy access to all of the features on demand.

Despite the fact that our corporate counterparts have documented the importance of optimizing the ergonomic design of the workstation user, the radiology literature has paid scant attention to the importance of this in reading room design. The radiologist's chair, workstation table, keyboard, mouse, and monitors should be designed to maximize comfort and efficiency. The architectural literature makes specific recommendations concerning optimal viewing angle and distance for computer monitors. ¹ A large body of literature exists regarding airplane cockpit design that clearly documents the importance of ergonomic factors.

Other reading environments within and outside the radiology department

In addition to the main reading room, which is used primarily for the interpretation of conventional radiographs (computed radiography, digital radiography, and fluoroscopy), the Baltimore VA has soft-copy reading areas in other locations in the department, such as angiography, neuroradiology, cross-sectional imaging, and nuclear medicine. Each of these locations has its own unique challenges with regard to lighting and sound depending on its proximity to other working areas in the department, often with limited ability to control room lighting, noise, and ventilation.

As we move outside the imaging department to workstations located in the emergency department, the intensive care units, and the operating rooms, it becomes apparent that these environments are even more difficult to control (figure 6). For example, one of our PACS workstations is located in the admitting area of the emergency department where lighting and sound are clearly impossible to control as is true, of course, of the operating rooms and the intensive care units. In these areas, restricted physical access to these workstations, which will be mandated under the HIPAA regulations, will require additional technologic developments such as a radiofrequency-controlled identification card that can automatically sign users onto and off of a PACS workstation in a relatively "public" location. We have performed a study that has documented large variations in background sound (both average and instantaneous) in the numerous areas throughout the hospital that have PACS workstations. As some radiology departments are considering relocating from the imaging department to a more distributed model with radiologists located in the trauma areas or intensive care units, these background sound and lighting concerns become more important.

Conclusion

The transition from a film-based to a soft-copy, filmless environment presents us with the opportunity to redesign not only our workflow, but also our reading environments. In our opinion, this redesign is likely to result in substantial improvements in radiologist performance resulting in reduction of fatigue, increased productivity, increased diagnostic accuracy, and possibly increased job satisfaction. Surprisingly, this opportunity has received little attention in the diagnostic imaging literature and presents the radiology research community with fertile grounds for future investigation. AR

Acknowledgment

The Baltimore VA Medical Center Department of Diagnostic Imaging gratefully acknowledges the partial support of our work by General Electric Medical Systems (Milwaukee, WI) who have provided us with a research grant to design and implement a second-generation PACS reading room to test the impact of various environmental factors such as lighting, acoustics, ventilation, optimization of workstation ergonomics, and alternative display devices.