Lead author Dr. Patrick Flanagan, of Santa Clara Valley Medical Center in San Jose, CA, and co-authors from the University of Washington found that the image transfer approach used by the university's Harborview Medical Center (HMC) led to a repeat study rate of only 17% over 500 consecutive trauma transfer patients -- compared with rates of 29% to 58% in the literature.
"Using a combination of the Internet and CDs to transfer images during interhospital transfer is associated with a lower repeat CT rate than in the literature, suggesting that regional PACS networks may be useful for reducing cost and radiation exposure associated with trauma," the authors wrote (JACR, September 2012, Vol. 9:9, pp. 648-656).
VPN for image sharing
Like many hospitals, HMC was faced with the problem of handling images for trauma patients transferred to the facility, an increasing phenomenon due to the regionalization of healthcare, according to the authors. A 2000 study estimated that repeated laboratory and radiology testing resulted in an additional cost of $600 per transferred trauma patient.
CDs are frequently sent along with patients, but they are often incompatible with the receiving hospital's PACS network and can take time to upload. Internet-based teleradiology is seen as a more elegant solution, particularly with the development of regional image exchange networks.
HMC decided to take a two-pronged approach to the problem, setting up an extensive Internet-based VPN to provide direct connections for image sharing with outside hospitals. At the same time, the enterprise adopted technology to make CDs more easily importable into its PACS network for those facilities without access to the VPN.
All trauma patient transfers that arrive at HMC are processed by a centralized transfer center, which coordinates the transfer of outside imaging studies along with all other aspects of the transfer. Whenever possible, the transfer center encourages referring hospitals to use Internet-based image transfer.
Images that arrive from outside PACS networks are received at a separate PACS (Centricity RA600, GE Healthcare) that was installed specifically to temporarily store images arriving via the VPN. The progress of outside images arriving from this PACS can be tracked through an electronic whiteboard that is accessible over the HMC intranet.
Trauma patients that are en route have their images automatically transferred from the separate PACS to HMC's main PACS (Centricity RA1000, GE). CDs are the preferred image transfer means for patients arriving from centers without VPN connections. These images are uploaded to the PACS (RA1000) in the emergency radiology area using a desktop PC and version 2.7 of DICOM Open Lite Box software (Sorna).
Because previous studies that had investigated CT repeat rates in transferred trauma patients did not involve Internet-based electronic transfer of images, the researchers sought to assess how their approach affected imaging repeat rate, cost, and radiation dose.
In a retrospective study, the team identified 500 consecutive trauma patients who had been transferred to the trauma center from an outside institution between June 1 and July 15, 2009. To be included in the study, patients had to arrive at HMC within 48 hours of their initial arrival at the outside institution.
From a review of the PACS log, the team identified all CT and radiographic exams that were performed on each transfer patient at the outside institution and transferred into the HMC PACS, as well as all imaging that was performed at the trauma center.
Studies that were performed at the HMC emergency department were further classified as completion studies or repeat studies. Completion studies were studies of a different body region than what was examined by the outside study. They also included imaging of the same body region and modality if the repeat study was performed for a change in clinical status or for follow-up imaging based on a recognized finding on the outside study, according to the researchers.
The authors defined a repeat study as a local study performed after an equivalent outside study of the same modality was conducted but did not meet the criteria to be considered a completion study.
CT scan costs were determined by combining both the professional and technical component rates from the U.S. Centers for Medicare and Medicaid Services' (CMS) published rates for the state of Washington in 2009. The researchers relied on published values for radiation effective doses for each body part.
After nine patients were excluded for being nonacute transfers, 491 trauma transfer patients were included in the final study. Of these, 261 (53%) had outside radiography, 318 (65%) had CT exams, and 383 (78%) had some form of imaging that was imported into the trauma center PACS. At HMC, 331 patients (67%) went on to have radiography, while 257 (52%) had a CT study and 383 (78%) had radiography or CT (or both). Of the 318 patients who received outside CT scans, 92 (29%) also received at least one further scan at the trauma center.
Low repeat rates
There were 69 repeat local CT studies performed on 55 patients, representing 17% of all patients who had imported CT studies and 11% of all trauma transfer patients.
"Overall, patients who underwent repeat CT scans were older and more severely injured than patients who did not undergo repeat imaging," the authors wrote.
Reasons for repeat CT included the following:
- Inadequate outside CT: 36 patients (52%)
- Unknown: 21 patients (30%)
- 3D reconstructions required for surgery: nine patients (13%)
- Images inaccessible on the PACS: three patients (4.3%)
The total cost of the CT studies that were transferred for the 491 patients was $244,373.69, or $768.09 per patient. The estimated value of CT studies that were repeated at the trauma center was $20,495.95, or $84.65 per patient.
As for dose, the patients who had their CT scans imported into the trauma center had an effective dose of 11.3 mSv. Repeat CT imaging accounted for an average additional 1.0 mSv per patient; the average effective dose from additional completion CT studies led to 4.8 mSv per transferred patient, according to the authors.
The researchers pointed to several probable reasons for the lower repeat CT rate, including HMC's policy of informing referring hospitals of minimal standards required for CT studies and the presence of an onsite radiologist who examines all incoming CT studies to determine acceptability. In addition, the institution rarely encounters CD incompatibility issues in its practice, according to the researchers.
Also, the reduction may be due to the ability to call the outside hospital and request the transmission of additional CT reconstructions after patient arrival.
Since the study was performed, HMC has reduced its use of CDs to the point where only 10% of images for its transfer patients are on the media. The goal is to eliminate CDs altogether, according to senior study author Dr. Martin Gunn, from the University of Washington.
"When [CDs] are used, they slow the process and are seen as not ideal," Gunn wrote in an email to AuntMinnie.com.
HMC's VPN is also an interim technology, Gunn added. Many radiology facilities are moving toward systems that use image transfer protocols that are compatible with the Integrating the Healthcare Enterprise (IHE) Cross-Enterprise Document Sharing (XDS) standard, rather than peer-to-peer systems such as VPNs.
And while commercial software is now available for image sharing, Gunn acknowledged that such solutions are more expensive than what HMC developed on its own. He advised sites looking at such applications to make sure they use vendor-neutral architecture, support industry-wide image sharing protocols such as IHE-XDS, and meet the newest security regulations.