PACS primer focuses on fundamentals, foresight
SAN ANTONIO—For facilities looking to implement their first PACS product, a solid grounding in the fundamentals of clinical information system deployment is paramount—as is an understanding of the evolving environment of healthcare IT. According to a host of presenters as the Digital Healthcare Information Management System (DHIMS) conference this week, installing a new PACS requires that system administrators not only focus on the needs of their digital imaging department, but also on its interoperability and integration with other healthcare systems.
Steven C. Horii, MD, professor of radiology and clinical director of medical informatics at the University of Pennsylvania Medical Center in Philadelphia, noted that a PACS needs a great deal of information from a RIS if it is to gain all its purported advantages.
“Much of the automation in a PACS requires information from the RIS,” he said. This includes features such as the DICOM Modality Worklist; the pre-fetch of prior studies (if the archive design includes a long-term, near-line component); hanging protocols for radiologists; radiology reports (if they are presented with the studies); and routing information (if the PACS design uses this).
In turn, a RIS (in a hospital environment or integrated delivery network) commonly will receive its information from a hospital information system (HIS).
“The disadvantage of this model is that the common information system has to interface to all the other information systems,” he said. “Since much of the HIS, RIS and IS development took place before standards were proposed, interfacing was [and still is] largely a custom effort.”
However, Horii noted that HIS and RIS products are evolving. They are gradually (mostly because of customer and other vendor pressure) becoming more open, he observed.
“With the deployment of PACS, the RIS [and HIS as the facility-wide information systems] have become essential parts of PACS operation if the productivity gains promised by PACS are to be achieved,” he said.
Herman Oosterwijk, president of Dallas-based healthcare IT education firm OTech, presented DHIMS attendees with an introduction to DICOM and Health Language 7 (HL7).
“Both DICOM and HL7 are flexible by design, which has encouraged broad implementation,” he said. “However, the same flexibility can interfere with interoperability between components in a PACS, or with a PACS capability to communicate with a HIS or RIS.”
In order to minimize potential disruptions to the clinical environment, Oosterwijk advocates that system administrators set up a test system that mirrors the PACS environment. It should include a modality simulator, a worklist simulator, test database/archive as well as DICOM viewers and a DICOM sniffer that captures DICOM communications (including images).
This test-system environment can then be used to work out interface issues, configure new modalities to bring online, examine the effect of software patches/upgrades, and provide temporary PACS capabilities in the event the production PACS going offline.
“If the primary PACS goes down, critical images can be routed to the test system as a failsafe measure,” he said. “To take advantage of this, the simulated database/archive software should be identical to what is used for the main PACS. To control costs, the simulated database/archive can be run on lower-capacity hardware than the production PACS.”
Leonard Avecilla, a Dallas-based clinical information system consultant, advised first-time PACS adopters to confer with other digital imaging providers within their institution—such as cardiology, neurology and orthopedics—as these specialties may also be users of the system.
“Although PACS was once believed a radiology-centric system, it now has to be considered as an enterprise asset,” he said.
John Koller, president of Larkspur, Colo.-based KAI Consulting, recommended that a strong IT infrastructure be in place prior to PACS deployment, as the expectation among users is that the system will be highly available—with a target uptime of 99.999 percent.
“The infrastructure is the foundation of the integrated digital healthcare enterprise,” he said.
Steven C. Horii, MD, professor of radiology and clinical director of medical informatics at the University of Pennsylvania Medical Center in Philadelphia, noted that a PACS needs a great deal of information from a RIS if it is to gain all its purported advantages.
“Much of the automation in a PACS requires information from the RIS,” he said. This includes features such as the DICOM Modality Worklist; the pre-fetch of prior studies (if the archive design includes a long-term, near-line component); hanging protocols for radiologists; radiology reports (if they are presented with the studies); and routing information (if the PACS design uses this).
In turn, a RIS (in a hospital environment or integrated delivery network) commonly will receive its information from a hospital information system (HIS).
“The disadvantage of this model is that the common information system has to interface to all the other information systems,” he said. “Since much of the HIS, RIS and IS development took place before standards were proposed, interfacing was [and still is] largely a custom effort.”
However, Horii noted that HIS and RIS products are evolving. They are gradually (mostly because of customer and other vendor pressure) becoming more open, he observed.
“With the deployment of PACS, the RIS [and HIS as the facility-wide information systems] have become essential parts of PACS operation if the productivity gains promised by PACS are to be achieved,” he said.
Herman Oosterwijk, president of Dallas-based healthcare IT education firm OTech, presented DHIMS attendees with an introduction to DICOM and Health Language 7 (HL7).
“Both DICOM and HL7 are flexible by design, which has encouraged broad implementation,” he said. “However, the same flexibility can interfere with interoperability between components in a PACS, or with a PACS capability to communicate with a HIS or RIS.”
In order to minimize potential disruptions to the clinical environment, Oosterwijk advocates that system administrators set up a test system that mirrors the PACS environment. It should include a modality simulator, a worklist simulator, test database/archive as well as DICOM viewers and a DICOM sniffer that captures DICOM communications (including images).
This test-system environment can then be used to work out interface issues, configure new modalities to bring online, examine the effect of software patches/upgrades, and provide temporary PACS capabilities in the event the production PACS going offline.
“If the primary PACS goes down, critical images can be routed to the test system as a failsafe measure,” he said. “To take advantage of this, the simulated database/archive software should be identical to what is used for the main PACS. To control costs, the simulated database/archive can be run on lower-capacity hardware than the production PACS.”
Leonard Avecilla, a Dallas-based clinical information system consultant, advised first-time PACS adopters to confer with other digital imaging providers within their institution—such as cardiology, neurology and orthopedics—as these specialties may also be users of the system.
“Although PACS was once believed a radiology-centric system, it now has to be considered as an enterprise asset,” he said.
John Koller, president of Larkspur, Colo.-based KAI Consulting, recommended that a strong IT infrastructure be in place prior to PACS deployment, as the expectation among users is that the system will be highly available—with a target uptime of 99.999 percent.
“The infrastructure is the foundation of the integrated digital healthcare enterprise,” he said.