When it comes to digital radiography (DR), all those in favor shout “I” — as in Improved departmental workflow and Increased patient throughput.
While it is a commonly accepted view that, in the efficiency game, DR trumps computed radiography (CR) and traditional film radiography hands down, is there any information that proves that position? And, while we are on the subject, what other considerations foster an efficient digital environment?
Those questions are what Mt. Auburn Hospital (Cambridge, Mass.) set out to determine. Those issues also were the reasons why the Cleveland (Ohio) Clinic Foundation recently teamed with Canon Research Center America (Palo Alto, Calif.) on their DR project.
Playing in the Ivy League
Mt. Auburn, a Harvard University (Cambridge, Mass.) teaching hospital that counts the Ivy League university and the well-respected Massachusetts Institute of Technology (MIT of Cambridge) among its high-profile neighbors, employs both CR and DR.
Mt. Auburn operates two diagnostic suites — each with one Philips Medical Systems North America (Shelton, Conn.) bucky system — in its main emergency room and an Agfa Corp. (Ridgefield Park, N.J.) Compact 70 multiplate CR reader. The facility also runs a DR room in an ambulatory imaging area one floor above the main radiology department.
Hologic Inc.’s Epex system features a
floating tabletop and flexible detector system.
Currently, the equipment in that room is a Hologic Inc. (Bedford, Mass.) retrofit. While Mt. Auburn has an image network and a print network, it lacks a fully functional PACS (picture archiving and communications system), thus limiting the radiology department’s ability to broadcast views and requiring the department to print dry laser films for archiving in file jackets.
Mt. Auburn was expected to host the first clinical installation of Hologic’s newest Epex unit come mid-July — an option that Dean DeMasters, Mt. Auburn’s radiology department manager, says was negotiated upon delivery of the retrofit.
In a series of time-and-motion studies conducted earlier this year, Mt. Auburn clocked 75 consecutive patients through a two-view DR process and another 75 through a two-view CR process. The “process” was defined as a series of procedures that began the moment patients stepped into the room and ended the moment that images became available on screen in the main interpreting area one floor below. The steps include data entry, cassette identification, position, exposure, QC (in the case of CR) and similar steps for DR — all in an effort to at least limit, if not avoid, variation across procedures.
The comparison surprised even DeMasters, who admits he had expected that DR would come out on top.
“I had figured I might get as much as a two-to-one advantage back when we first started looking at this, in terms of productivity,” he says. “Our preliminary data makes it look more that it is three or four times faster to do digital radiography compared to CR.
“In our actual work experience [with CR] over those 75 cases,” he continues, “we’re looking at nine to 10 minutes time, beginning with the patient exam to those images being available on the monitors in main interpreting for my radiologists to type into the emergency room computer system the initial reading on those films. [With DR], we’re averaging something on the order of three minutes from the patient stepping into the room and starting the process to the images being available one floor below in main interpreting. And I expect that to get faster when my techs are positioning with a floating tabletop [on the Epex], as opposed to a tilt table [on the retrofit].”
The heart of the matter
Meanwhile, at the Cleveland Clinic Foundation’s renowned heart center, Candice Bergsneider, manager of infrastructure applications, Doreen Dackiewicz, R.N., and radiologist David Piraino, M.D., section head, computers in radiology, were collaborating with Jerry May, Ph.D., director of systems development at Canon Research Center America, to study the impact of DR on clinical workflow. Cleveland Clinic looked to the experiment as an opportunity to integrate its new radiology information system (RIS) from IDX Systems Corp. (Burlington, Vt.). The hospital also decided to extend the study and document patient satisfaction.
May, who appreciates that his colleagues allowed him to take the lead, says Canon had surmised that DR — a faster, yet more expensive modality — would increase patient volume, while reducing operating costs, resulting in a tighter, more efficient radiology practice. The company, however, had no figures or formal workplace data to back up its claims.
As May puts it: “People have done ad hoc things, and they kind of know it is faster, but we wanted to nail this down for the Canon system specifically. We had to get hard numbers.”
And they did.
Throughout January and February and into the spring, the team surveyed 39 film images, 65 digital images without RIS and 75 digital images with RIS integration.
The film/screen system was a Marconi Medical Systems Inc. (Highland Heights, Ohio) model with a direct feed to an Eastman Kodak Co. (Rochester, N.Y.) RPX-OMT processor model M6B. The DR system, already in place at Cleveland Clinic, was a Canon Medical Systems (Irvine, Calif.) CXDI-11, interfaced to a Windows NT operating system.
All exams consisted of posterior-anterior (PA) and lateral views on outpatients only. Time measurements were organized into the following categories:
• ExamBegin (EB) — The time from when the patient entered the room to positioning for the PA exam, including time needed to enter patient demographic information;
• ExamComplete (EC) — The time needed to expose the PA exam, reposition the patient and complete the lateral exposure;
• Development (DEV) — The time for film/screen system development or, in the case of DR, the time to present an image for viewing;
• Quality Control (QC) — The time it takes to QC the exam until the time the patient leaves the department. For DR, that meant approving the preview image that appears on the touch screen; for film, that meant looking at the films on a light box.
Their findings?
“It takes about five minutes to do a film X-ray all the way through and it takes half that — two-and-a-half minutes — to do a digital X-ray,” May discloses. “With RIS integration, where you download the information for the RIS system, you get an additional 30 percent savings from that.” (See table.)
“I think the study confirmed what we had been thinking,” Bergsneider offers. “We knew it was going to be faster; the question was just: How much faster was it?”
“When you make your pitch for digital radiography, it is more expensive than film, up front. What we want to show is, even though you have the up-front cost, you can process patients more than twice as fast, so your throughput goes dramatically up,” May adds. “Therefore, in terms of dollars-and-cents savings, you can show specifically now that there is a savings that mitigates the cost of the equipment.”
DR and PACS: Made for each other
DeMasters plans to repeat some of his data collection once the Epex — with its floating tabletop and flexible detector system — is up and running. He also intends to go DR in the two emergency room suites and to stick with Philips for those conversions, “since Philips has the technology in place to upgrade all of its bucky diagnostic systems to digital without any major renovations to the system. Their existing buckys will accept their new bucky holder for their digital plates.”
DeMasters hopes, too, that the hospital will see its way clear to purchase a long-awaited PACS, thus taking advantage of DR’s full range of capabilities.
“We have paid the cost for this high-tech equipment, and then we still have to archive on film, so we have a film budget of approximately $350,000 a year,” DeMasters adds. “I could eliminate an awful lot of that with a PACS system.
“In our situation, running a CR reader separate from an integrated PACS, each image has to be sent separately to each location; I cannot broadcast them,” DeMaster continues. “That means if I do a two-view chest and I want those to go to main interpreting, I have to send each view to main interpreting. I also want them to show up on some clinical review monitors that I have in the emergency room, so I have to send them there. I want to print them on a dry laser so I can archive the film, so I have to send them there. Then, this is a teaching hospital, and we have another set of clinical review monitors next to our film room so that house officers and referring physicians can come up, so I have to send them there as well. With a PACS, that would be one step to send them everywhere.”
Many industry observers wholeheartedly agree that a PACS is what makes digital X-ray all that it can be. Unfortunately, they admit, most hospitals find a PACS’ cost prohibitive. Often, a radiology department has to approach a PACS purchase as an education issue, involving personnel throughout the hospital system.
Every piece of film a hospital does not have to buy saves about $1. On average, 60,000 procedures a year translate to 150,000 sheets of film or $150,000. Multiply that 150,000 by a three-year return on investment, and a hospital readily can see savings in operating costs that could be directed elsewhere within the system or eliminated altogether.
Workflow considerations
Mt. Auburn reports that when it came to image quality, their radiologists expressed no preference between CR and DR. “I think, theoretically, that the DR system should provide better images, but in an everyday, working environment, the radiologists like both sets of images, and they don’t point to one and say ‘This is clearly better all the time,’” DeMasters relates.
At the same time, though, DeMasters suggests that DR images take longer to get to their respective destinations than CR images.
“These are big images; they do get hung up in the network more often than I would like, but I do not think that there is anything inherent in the system that is a problem going forward,” he adds. “I think DR images right now get hung up more often than the CR images do, but I think that is a function of broadcasting them to multiple locations at once. If one is not available, it will hang the whole transaction, whereas the CR images — when you are sending them to one place at a time — if they cannot go through to one, it does not keep the others from getting there when you do those ‘sends.’”
The two sites also touched on DR’s infrastructure requirements — for its communication channels as well as for the physical space it demands.
“Obviously, when you take images which actually are large files and ship them between sites, you need a very large pipe on your network,” comments Cleveland Clinic’s Bergsneider. “Like most institutions, we are scrambling to get the biggest pipe we can get at any point in time. We’re kind of chasing technology in that respect. Here, radiology is on a separate network, because we can predict what our point A and point B are going to be. We cannot be that predictive when it comes to an ordering physician or a referring physician wanting to see the actual exam, so there is always a lot of discussion about the network.
Bergsneider says that when Cleveland Clinic integrated any of its imaging modalities with its RIS, training included some “personal hand-holding.” Cleveland Clinic went modality-by-modality, sometimes several modalities at a time, because they were located together and the technologists were able to rotate during the day in order to be trained during actual exams.
“For the Hologic room, we had an on-site person for well over a month, working with our techs and also optimizing the reconstruction software so that the images came up looking the way our radiologists wanted to see them,” DeMasters recalls. “Between doing that and rotating all our techs through that room, we got a considerable amount of training. We will certainly have applications again when we upgrade to the Epex unit, and there will be a new round of adjusting the look of the images with the newer software that comes with that unit as well.”
DeMasters also was able to tailor rooms for digital and include no darkrooms in Mt. Auburn’s new construction.
“I was actually physically limited on the size of the rooms to the available space; all of them are a little smaller than you would ideally like for any X-ray system,” he adds. “I think the Epex design looks as though it may fit in a slightly smaller room than a standard table with a wall bucky, but if you look to the full capacity of the system, you still need a standard X-ray room size.”
Ideally, DeMasters says Mt. Auburn should have been able to know exactly what equipment was going in the room, so as the room was built, the facility could construct the conduits and floor troughs and wall channel.
“We did that,” he adds, “but we did it for a model which then did not become available, so we wound up having to fit this retrofit system in a space designed for [the previous Hologic model]. We built [the ER rooms] specifically to fit the Philips bucky diagnostic systems.”
The ‘end user’
To all the considerations that Mt. Auburn and Cleveland Clinic put into their studies, their planning, and their efforts to improve departmental workflow, Cleveland Clinic added one more. As an extension to the workflow study done with Canon, the hospital separately polled patients — the “end user” — on the issue of customer satisfaction.
Patients scanned with DR who, in the past, had been through a conventional chest X-ray and could compare the two procedures, were asked if they had a preference.
“We basically found out that the patients did like the digital test exam better,” reveals Bergsneider. “Sixty-seven percent thought it was better; 33 percent just rated it the same. There was no one who said it was worse. A good part of it is they are in and out of the department so quickly. The wait that they have is the wait for the exam to be taken, not the additional wait after for the tech to say, ‘Your pictures look good,’ or, in the worst case, ‘you have to do a retake.’
“From my perspective,” she adds, “I would have thought that the numbers would be a little bit higher, that patients felt that the digital chest exam was better and there would have been less than 33 percent that rated it just the same.”
She pauses, then considers the patient’s point of view.
“You are asking people this question, people who are there because they are concerned about their health. To be accurate about their [survey] response is not high on their list.”
