Home

Cover Story

Table of Contents

E-Newsletter

Article Archive

Editorial Calendar

Datebook

Writers' Guidelines

Orgs/Links

Opinion Polls

Reprints

Forum

August 13 , 2007

Workflow Efficiency — Maybe the Best Answer to Radiology’s Growing Volume Problem
By David Yeager
Radiology Today
Vol. 8 No. 16 P. 20

Imaging’s growth and its evolving technology are rapidly increasing the number of images radiologists need to read. Keeping up requires increasing efficiency in collecting, interpreting, and distributing that information.

In any industry, the organizations that find ways to enhance efficiency tend to be the most successful. Radiology is no exception. Radiologists read more studies and look at more images than ever before, and many radiology departments are struggling to keep up. The ones that streamline their processes will be better positioned to survive, and even thrive, in this era of information overload.

It’s been said that the journey of a thousand miles begins with one step. And with that much ground to cover, every step counts. So, if someone could tell you how to travel that distance in fewer steps, you’d probably be at least somewhat interested in what they have to say.

“We did a study where we looked at our workflow steps, [specifically] the 59 steps that were required just to do a chest radiograph. And so, one of the things that we tried to do was to study those to automate the processes and to minimize the number of steps,” says Eliot L. Siegel, MD, professor and vice chairman of the University of Maryland’s department of diagnostic radiology. “And that [includes] all the steps involving the referring clinician and the clerks and all the parts in the radiology department, including the radiologist and transcriptionist.”

Navigating the Terrain
A significant source of extra steps in the image viewing process is inefficient navigation. In another study, Siegel and colleagues compared a traditional mouse to an alternate navigation device. “We were able to document a 40% decrease in the amount of time required to interpret CT studies by only altering the navigation device, which was unexpectedly high,” he says. “We really were hoping to be able to document a 4% or 5% decrease, and the radiologists really didn’t feel as though they were much faster until we actually told them what their times were with the one device vs. the conventional mouse.” To aid time reduction, Siegel uses an intelligent keyboard that minimizes the number of keystrokes and a special roller mouse. Others are looking at alternative navigation devices as well.

“The keyboard and mouse are antiquated, as far as I’m concerned, for controlling PACS. I log on with the keyboard, and I try not to use it for the rest of my session,” says David L. Weiss, MD, clinical head of imaging informatics at Geisinger Medical Center in Danville, Pa. “I will use a mouse in one hand usually and some other control device in the other hand, and I use a headset microphone so I don’t waste a hand holding the microphone,” he adds. Weiss prefers a shuttle pro multimedia controller for anything that requires scrolling or moving images, such as CT or MR.

The goal is to reduce the number of times the radiologist looks away from the image. Time spent thinking about the navigation device is wasted time.

“I find that every time I look away from an image, I lose my concentration,” says Weiss. “So I need to be looking at the image and not at any visual user interface. Ideally, my eyes should be locked on the image, and I should be controlling both PACS and the reporting system without the need for visual input. Right now, I use my hands and voice commands for that. There is no reason I shouldn't ultimately be able to use other body parts such as feet or head movements.”

Hanging protocols—the way the images are presented—also affect reading time. Earlier PACS systems sent images in standard views or in the position in which they were acquired. Newer systems allow radiologists to tailor their preferences according to how they work.

Integrating Information Systems
Improved efficiency also entails eliminating functions that don’t make optimum use of people’s time. “One area of low-hanging fruit to improve your efficiency in your process, to achieve optimized information throughput, is minimize the peripheral tasks,” says Paul J. Chang, MD, FSIIM. Considering the volume of information attached to each patient, integration of electronic systems can ease some of the information overload burden by transferring relevant information between systems.

And while many healthcare facilities have gotten rid of film, Chang believes they haven’t gone far enough. “You’ve got to get rid of the paper because paper is a baton. Because of the lack of integration, a tech has to stop what he or she is doing and walk physically over to deliver that baton to me, even though everything’s electronic.”

Eliminating paper and integrating information systems allows the radiologist and the technologist to use asynchronous communication tools, which don’t require each person to drop what they’re doing and speak directly to each other in order to perform their respective jobs. When Chang, who recently became the vice chairman of radiology informatics and medical director of pathology informatics at the University of Chicago’s Pritzker School of Medicine, was at the University of Pittsburgh Medical Center, Pitt’s transition to a paperless environment doubled their study throughput.

In addition to streamlining workflow, integrating data systems reduces errors. “If you ever look at a technologist in a CT or MR scanner, they’re spending a lot of time transcribing information that exists in one electronic system, the RIS and the PACS, into another electronic system, which is the scanner. And that information should transfer automatically,” says Chang. “They’re spending 10 minutes typing the same information from one system to the other one. That’s crazy. That’s only going to add inefficiency, and more importantly, that usually adds ... inaccuracies.”

The University of Maryland has also eliminated paper, and they’ve engineered their reading room to optimize efficiency. In what Siegel describes as the radiology reading room of the future, “what we’ve tried to do is minimize the number of steps that the radiologist takes and minimize the amount of time that a radiologist has to go from one computer to another computer,” he says. “And so, at my workstation and at all workstations in the radiology reading room of the future, we have computer-controlled lighting. We have computer-controlled LCD privacy glass. The radiologist has speech recognition built into the same workstation that does the PACS, and it’s the same workstation that does the 3-D visualization, and it’s the same workstation that allows him or her access to the electronic medical record [EMR] and the Internet.” This set-up frees the technologists to focus on acquiring the volume of scanned information and allows the radiologists to manipulate it as needed.

Contrary to the Maryland approach, the University of Wisconsin has its technologists do the 3-D modeling to allow the radiologists to focus on the rendered image, according to Walter W. Peppler, PhD, professor in the department of medical physics, but what is in common with Maryland is that both try to utilize a single physical space for viewing. Minimizing time spent moving between workstations is crucial for keeping up with increasing image volumes.

Better Reporting
Patient care isn’t optimal unless the radiologist communicates his or her findings in a timely manner. By linking relevant systems—PACS, RIS, EMR—to a single workstation, reporting can be expedited and enhanced. “You can consider your reporting software as really a communication system, and that, arguably, could be the cockpit, so to speak, where you’re controlling all of your clinical information. If you live in the reporting world and have other information systems integrated, you can pull data as needed, contextually, from each of those,” says Weiss.

Weiss uses templates, or macros, to speed reporting, but rather than using a macro that simply allows him to make changes by voice command, which frees his eyes but requires him to expend some brain power on the structure of his report, he favors a tool called findings-only dictation. The just-developed findings-only dictation software recognizes clinical terms and automatically organizes them within the report, allowing the radiologist to focus on the study. “As I am scrolling through the liver, I may see something in another organ that catches my eye that requires dictation. So if I need to report findings in a certain order in a macro, it’s not going to work for me,” says Weiss. “I don’t want the macro dictating my viewing habits, I want my viewing habits dictating, no pun intended, how my report is created.”

Back to Basics
Making data centrally available also has implications across the healthcare enterprise, says Peppler, who’s in charge of distributing data across the enterprise at Wisconsin. With more departments using PACS, workflow considerations need to encompass entire facilities. “There’s nothing fundamentally difficult about getting it all in one place, but the other ‘ologies’ aren’t aware of it as much, and so you need to kind of persuade them,” he says. He expects the enterprise view to become a larger focus over the next decade.

Improving workflow in radiology and across the enterprise will require new work processes. “A lot of times in radiology, people focus on what I believe are too granular and too limiting views, or measures, or metrics for efficiency. For instance, patient turnaround time, report turnaround time, throughput within the scanner suite. And the problem with that is, that’s great and these are all important, but if you just optimize just to improve those numbers, you’ll miss the forest for the trees,” says Chang. “You won’t see that, for instance, it does no good to have a patient go through our department within just a few minutes if it still takes them three weeks to get scheduled.”

Chang believes that better utilization of staff—”the pit crew”—can have a significant impact on efficiency. “When it comes to throughput in a radiology department, we’ve got 64-slice detectors or plus, we have multidetector CTs that can scan you from head to toe in a few seconds, and yet we still schedule them in 40-minute blocks. Why? Because you still have to get the patient in the room, you’ve got to get the IV, you’ve got to interview the patient, you’ve got to set up the machine. There’s a lot of potential for improved efficiency. We need to get closer to 15 minutes, if not 10 minutes.”

Considering that the people who perform these ancillary tasks are the ones often overlooked, Chang believes an adequate pit crew should be in place to address throughput barriers before more radiologists are hired. One of the people that Chang credits with helping him to see this particular forest is James H. Thrall, MD, FACR.

Eighteen years ago, Thrall, chief of radiology at Massachusetts General Hospital (MGH), noticed that workflow was low on the radar screen. “When a quality improvement initiative was to be undertaken, someone with a ‘day job’ was typically given the leadership responsibility. Then a committee of other people with ‘day jobs’ was put together to support the project,” he says. The problem with this approach was “quality improvement was tackled as a series of projects. No continuity on process issues from one ad hoc team to the next. No substantial time commitment or availability beyond people carving out time from their preexisting responsibilities,” he adds.

To remedy this shortcoming, Thrall and his colleagues created an internal consulting group—consisting of “first, a person with a master’s degree in organizational development; second, a person with a master’s degree in industrial engineering; third, a person with a master’s degree in educational curriculum development and design; fourth, a project manager; and fifth, a coordinator”— whose only responsibility was to study work processes.

Using flow charts to identify the necessary people, knowledge, technology, and space required to improve or re-engineer a work process, MGH began to see dramatic efficiency improvements. In fact, it worked so well that it’s become a fixture at MGH. And even though it’s been successful at MGH and other hospitals, Thrall believes that too few facilities are putting enough resources into understanding their workflow. “The key part of this that everyone misses is how complex healthcare delivery is, and we have missed the fact that we really need an industrial strength approach to process engineering.”

— David Yeager is assistant editor of Radiology Today.