The Ann Arbor VA Healthcare System has established a picture archiving and communication system (PACS) that integrates all imaging modalities into a fully digital environment. This article provides a detailed discussion of the system's implementation and capabilities. The author illustrates how such systems can streamline workflow, reduce operating expenses, and improve patient care.
is the PACS System Administrator of Ann Arbor VA Healthcare
System, Ann Arbor, MI.
veryone in the industry is talking about picture archiving and
communication systems (PACS)--how to design a system, the
challenges of implementation, and the impact digital imaging can
have on patient care. At Ann Arbor VA Healthcare System, when we
talk about PACS we speak from experience. Our PACS integrates all
modalities into a 100% digital environment that streamlines
workflow, reduces operating expenses, and improves patient
Current and prior imaging studies are available consistently for
interpretation or review. We have eliminated lost or missing films.
More importantly, Dr. Charles Marn, Chief of Radiology Services,
reports that timely access to a patient's previous studies (from
any modality) improves the quality of diagnostic results we
Our 147-bed facility provides ambulatory and tertiary services
for veterans in the Southeast Michigan area and is a teaching
hospital affiliated with the University of Michigan.
The hospital began with mini-PACS, which included a 0.5 terabyte
archive and two primary diagnostic work-stations, and a
teleradiology system that connected the main facility with an
outpatient clinic in Toledo, Ohio. In 1997, we began to upgrade to
a full-scale PACS.
All imaging modalities were connected to the PACS by 1998, and
the hospital significantly reduced expenses for film and chemicals
from $125,145 in fiscal year 1996 to $20,419 in fiscal year 1999
(Table 1). Labor costs have been trimmed by $26,300 a year. Because
we are a teaching hospital, our percentage of missing films is
high--between 15% to 20%. Therefore, digital imaging has also
eliminated the cost of labor and consumables for retakes. At the
same time, workflow is improved. Imaging studies are now available
in just minutes to referring physicians on any floor of the
hospital. With the film-based system, it might take 2 to 6 hours
for a referring physician to see a film study.
Digital imaging has also improved technologist productivity and
patient throughput. It may also improve the quality of care we are
able to provide. Currently, we are conducting a study that compares
how many previous studies a radiologist examines with a film system
as compared with a digital imaging system. We are also tracking how
long it takes for radiology reports to be made available, for
technologists to process studies, and for patients to complete
Prior to interviewing any vendors, we created a
multi-disciplinary committee to develop a list of requirements.
This step was critical to our success, because it established goals
and objectives that drove the selection and implementation of the
final, full-scale PACS solution.
Our PACS team consisted of the chief of radiology, a hospital
administrator, the chief technologist, and a referring clinician.
The chief of radiology and the chief information officer had the
foresight to recognize the importance of a dedicated, full-time
system administrator. Thus, I was transferred from the IT staff to
facilitate the PACS team.
We developed our PACS team in 1996 and spent 4 to 5 months
outlining our goals. During that time, each department
representative collected input from, and reported to, a group of
his or her peers. All team members agreed on the overall objectives
for PACS and participated in creating a detailed system plan and
Based on the initial needs assessment, I created a detailed
request for proposal. After interviewing several vendors,
Cemax-Icon, a Kodak Company, best met our list of requirements,
which included: functionality of the diagnostic workstation;
enterprise-wide networking expertise; archive storage and database
technology; teleradiology experience; and validation and
integration of imaging modalities.
With the vendor selected, we began to implement PACS. This
gradual implementation was desirable because of management and
training issues, as well as financial concerns. In our case, this
timeline was controlled tightly because of a new construction
project. The PACS was cost-justified as part of a $70 million
clinical addition, which features a new radiology department,
ambulatory care center, diagnostic imaging service, operating
rooms, and intensive care units.
In January 1997, our first phase began when we brought the
portable CR system and two ICU referral-quality workstations
online. In the next stage, which began in May 1997, we integrated
ultrasound and RF systems. In June 1998, we integrated CT and MR
systems as part of phase three. Phase four, in October 1998,
involved connecting angiography suites, a digital chest unit, and
general radiographic rooms to the PACS and coincided with our move
to a new facility. This stage completed our conversion to digital
imaging and now all images are read with on-screen digital
Our current stage, phase five, involves establishing worklist
management on all modalitites except angiography (completed); and
providing remote access to images and the electronic medical record
over the Internet (currently being tested). Figure 1 illustrates
the PACS workflow for radiology.
Each imaging modality was validated before it was brought
online. In 1997, Cemax-Icon had a DICOM validation group that was
responsible for validating the clinical and technical functionality
of its system with leading imaging modalities. This group
established protocols and used phantoms to create a series of
images that were analyzed for integrity during capture, viewing,
and storage. They compared softcopy images with film output to
ensure that the images were identical and to verify that positional
markers and other critical elements were correct.
Most of our imaging systems had already been validated, with the
exception of our new MR and RF units. The DICOM validation group
provided a phantom and a series of imaging protocols and we
acquired the test images, which were then evaluated.
This validation effort was critical to ensuring the clinical
accuracy of all images, as well as the smooth integration of each
imaging modality. It also allowed any non-conformance issues
to be addressed prior to linking the modality to the PACS
network--which ensured the integrity of every patient study in the
Storage and networking issues
In 1997, we moved our existing image files from our initial
archive to our new 17.5 terabyte digital linear tape (DLT) archive.
At this time, we began to take full advantage of the workflow
benefits of our new large-scale digital archiving system.
Film studies that are necessary for comparative purposes are
being digitized as needed, however we are not converting our entire
film library. Previous research has shown that film studies are
rarely accessed after 18 to 24 months. At our current volume of
60,000 studies per year, we estimate our archive requirements to be
1 to 2 tera-bytes a year in storage. Newer tech-nologies, such as
spiral CT, and the increased demand to provide primary diagnostic
services to remote facilities will expand our storage needs, so we
have built scalability into our system.
Our storage system supports hierarchical storage management and
an object-oriented database to speed retrieval times. In our
system, the patient is the object. We can enter a patient ID or
Social Security number and rapidly locate every study and report
for that patient.
We have recently added Modality Worklist, which eliminates the
need for technicians to key a patient's name and other demographic
information. With this feature, patient demographics are linked to
imaging studies through a HIS/RIS (hospital/radiology information
system) gateway. For example, a CR system operator can access a
list of all scheduled studies for a given day and click on a
patient's name, then all of the patient's demographic information
is attached to the image and transmitted to the PACS. This workflow
enhancement improves full-time equivalent productivity and
significantly reduces manual archiving reconciliations.
A bi-directional interface between the PACS and HIS/RIS system
enables radiology reports to be accessed from any PACS workstation
throughout the hospital. This interface also transmits study
schedules to the archive, where previous studies are pre-fetched
and posted to the primary diagnostic workstation for comparison
with the new study.
In any major system conversion, human factors must be
considered. For example, we tailored our training efforts to the
needs of each audience and described how each staff member would
benefit from the PACS. We explained to physicians and nurses in ER
and ICU, for example, that the new PACS would provide faster access
to images and eliminate lost or missing films. We emphasized to all
audiences that the primary purpose of our PACS is to achieve
improved access for medical staff and optimized care for
We also stated that we were not going to run parallel systems.
Once training was completed and the modality was brought online,
everyone would be required to use the new system as it was
intended. There is little motivation to learn a new system if you
know the old system is still going to be accessible. So it is
important to notify everyone up front that they will be required to
convert to the new system.
It is helpful to identify a "champion" within each group that
has a good understanding of the system and can help encourage and
train others on an informal basis. If users are enthusiastic, they
will quickly overcome the initial discomfort that accompanies the
adoption of a new technology, which will speed learning and create
Advantages for radiologists
By far the most important achievement of our PACS is that any
image from any modality can be retrieved from any diagnostic
workstation in seconds. Dr. Marn and other radiologists report that
access to comparative information can influence a diagnosis.
Because the PACS is connected to HIS/RIS, patients' previous
studies of the same or multiple modalities will be pre-fetched
automatically and packaged with current studies. Images are also
available on an ad hoc basis. If a radiologist sees a suspicious
nodule at the bottom of the lung, he or she can immediately request
the patient's most recent chest radiograph.
Dr. Marn outlines additional advantages our PACS offers
radiologists: 1) customizable options allow radiologists to create
parameters they find most productive for viewing different types of
studies. A radiologist can create automatic instructions for the
system to retrieve a chest CT with every thoracic spine MR exam,
and create hanging protocols that direct the order and location for
the placement of each study and image; 2) digital enhancement tools
enable radiologists to adjust contrast, enlarge selected areas, and
use other viewing techniques to check subtleties in an image. In
addition, on-screen measuring tools are so precise that Dr. Marn
notes he now includes more measurements in reports; and 3) the use
of "bookmarks" allows radiologists to save their place when
interrupted for STAT reads, which saves time and reduces
frustration. This is also a valuable teaching tool that allows
radiologists to share research and teaching cases in an extremely
In addition to implementing digital imaging, we have also
replaced paper patient records with electronic medical records. Our
goal is to make reports and images available to remote clinics and
physicians' offices and homes by using the Internet and an
interface to electronic medical records.
Enterprise-wide distribution will bring us closer to our
referring physicians, while equipping them with diagnostic
information they need. That is what PACS is really about--better