SAN ANTONIO - The automation of manual tasks plays an important role in realizing the benefits of PACS, according to Dr. Steven Horii of the University of Pennsylvania Medical Center in Philadelphia.
"Every time you interpose a person doing a manual task as an interface, you are helping to bury your productivity gains," said Horii, associate director of UPMC’s medical informatics group.
Horii discussed the process of reengineering workflow and UPMC’s extensive experience with digital image management at PACS 2004: Working in an Integrated Digital Healthcare Enterprise, sponsored by the University of Rochester School of Medicine and Dentistry last week.
One of the pioneering institutions for PACS research dating back to the early 1980s, UPMC (then known as Hospital of the University of Pennsylvania [HUP]), transitioned from a research PACS environment to a clinical system (PathSpeed, GE Healthcare, Waukesha, WI), adopting what was then the hospital’s third generation of PACS technology. This move was driven by a number of factors, including an increasing proportion of digital images, increasing demand for those studies, and positive results from its work under a National Institutes of Health (NIH) Program Project grant, Horii said.
What went rightLooking back on the experience’s successes, Horii said that the institution worked well with the vendor, especially in developing the initial plans for storage capacity. The interface to the RIS through a Mitra broker also performed well.
Another positive element was the development of customized worklists to suit the highly specialized department’s workflow, Horii said. In addition, the institution was successful in having radiologists and technologists trained and up to speed very quickly, which required considerable vendor time, Horii said. They also succeeded in gaining built-in access to the PACS database for research purposes, a benefit for an academic practice.
In another positive move, UPMC tied film reduction to the use of the system; the vendor had to pay for film used beyond a certain quantity. The hospital also had a negotiated hardware upgrade path, with a low incremental cost for essentially replacing all the hardware. This came in very handy during a migration from Macintosh-based systems to PC-based workstations, Horii said.
"All of that hardware was replaced at extremely low cost, because we built that into the cost for us," he said. "And the hospital was willing to pay for it because it was a negotiated amount."
What went wrongAnalyzing the institution’s mistakes, Horii said that there was a lack of detailed planning.
"We didn’t have enough workstations," he said. "We tried to do too much at once, and we did not look to automate all of the steps we should have automated."
In addition, no attempts were made to reduce or eliminate paper. Also, converting a department from Macs to PCs meant re-training, he said.
The institution did not project sufficient storage volume, and there was no initial support for DICOM modality worklist, Horii said. They also did not insist on robust database management tools, which would have helped in data mining and looking for broken studies.
A good historical record was also not kept, which can be useful in performing research, he said. Reading room design was also an area that could have been improved on; a couple problems were lack of appropriate planning for air conditioning and the placing of lightboxes on a wall directly across from the workstations.
As for the institution’s experience with interfaces, Horii said that electronic interfaces were not nearly as much of a problem as were non-electronic ones.
"The failure to improve workflow was often directly related to problems getting information to move between systems," he said. "So I still had to have people doing things."
In that regard, some important questions for prospective PACS purchasers to ask are:
- Do you know what all the pieces of paper your department handles do?
- Do you realize the cost in time and money of managing those pieces of paper?
Some sections of the radiology department are paperless for the radiologist, but this was difficult to achieve, Horii said.
If they could do it all over again, the institution would go with a more phased-in approach that would allow detection of workflow problems (such as not enough technologist QA workstations), Horii said. Also, they would ask for better systems management and QA tools, he said.
As time went along, the institution realized a dramatic improvement in system performance, and a significant decrease in mysterious crashes. Hardware changeovers went very smoothly, thanks to the vendor’s willingness to temporarily run parallel systems, Horii said.
Also having a big impact was the inclusion of DICOM modality worklist by the imaging equipment vendors.
"It changed our technologist productivity from negative to positive," he said.
Some trends were not possible to plan for, including: emergence of multidetector CT, greater-than-projected volume growth, hospital IT outsourcing, and the rapid growth of Web-based viewing, Horii said.
In conclusion, Horii said that the most difficult problems to solve resulted from operational interfaces, including paper to an electronic system, paper to people, imaging systems to PACS, and PACS to RIS.
"Automation via the IHE initiative is going to be the key to solving all of this," he said.
By Erik L. RidleyAuntMinnie.com staff writer
March 15, 2004
Related Reading
PACS success relies on QA/QC procedures, March 12, 2004
Italian hospital finds that PACS doubles costs, March 5, 2004
SCAR members offer solutions for data deluge, February 24, 2004
Workstation auditing tools open door to better design, February 20, 2004
Mouse roars in workstation navigational study, February 9, 2004
Copyright © 2004 AuntMinnie.com
