PACS faces close scrutiny over workflow

While PACS has more than lived up to its promise of improving access to imaging data, installation of the technology may not necessarily speed up healthcare delivery.

Time spent hunting down lost x-rays is a distant memory at filmless hospitals, where images can in theory be retrieved 24/7. But the switch to soft copy does not automatically speed up reporting workflow, according to speakers at a meeting in London organized by the British Institute of Radiology.

No matter how good the PACS, efficiency gains will not be realized without a mature network infrastructure and clear integration with other hospital IT systems, said Dr. Laurence Sutton, clinical director at Calderdale Royal Hospital in Halifax. The Halifax hospital opened filmless in 2001 and had enjoyed four years of problem-free image retrieval prior to this May. Then the short-term RAID became full, leading to a greater number of requests for images from the long-term archive. When a mechanical failure to this retrieval mechanism occurred, image delivery to clinical workstations was compromised.

"If we were to redesign the system to prevent this from happening again, we would have a huge short-term RAID so we could store seven years of images," Sutton said.

Theoretical workflow efficiencies can also fail to materialize because of regular reporting interruptions. Hospital-wide access to e-images means that clinicians are more likely to query reports.

"With protected time, you could perhaps deal with four or five times the number of routine MR scans than you did without PACS," Sutton said. "Of course, that does not happen in reality. PACS has rendered the radiologist available to the whole organization via the telephone."

Putting in PACS can actually compound existing workflow inefficiencies, said Dr. Nicola Strickland, a radiologist at Hammersmith Hospitals NHS Trust. A reporting backlog is only going to get worse once PACS facilitates almost instant access to studies as soon as the scan is complete.

"If you don't get your own house in order before you put in your PACS, you will be in for a very unpleasant time," she said. "If you have colleagues who are not pulling their weight, putting in PACS will only emphasize that. You have to have robust timetabling. People need to be doing their reporting when they are supposed to be doing it."

The use of projected efficiencies to justify the cost of purchasing PACS is also misguided, Strickland said.

"We don't do this for other modalities," she said. "If you want a new MRI or a CT scanner, you don't argue only on cost. I have always felt that because PACS improves healthcare by bringing the imaging environment to the entire hospital-wide system, we shouldn't justify PACS on economic grounds."

Movement in the debate over storage of CT data illustrates how expectations of PACS are changing as technology improves, said Dr. Stephen Davies, a radiologist at the Royal Glamorgan Hospital in Llantrisant, Wales. The vast data burden from multislice CT series prompted many radiologists to suggest that only a subset of data be archived. But growing interest in imaging data from other clinical specialties, such as orthopedics and maxillofacial surgery, should also dictate what is stored on the PACS.

"When we are designing our systems, we need to think outside of the radiology box," Davies said.

Strickland agrees that raw MSCT data should be stored on PACS, having previously argued against this position. Worries that the network infrastructure at Hammersmith Hospital would crash if clinicians tried to pull an entire 1000-image CT study onto a Web browser turned out to be unfounded, she said. The low cost of storage removes another possible hurdle to storing all thin, unreconstructed CT data in the PACS archive.