Digital Mammography Saves Technologists’ Time, Increases Radiologists’
Nuclear Medicine Use Increases 22% in 13 Years, Report Shows
NYU Medical Center Rolls Out Intuitive Voice Recognition System
Digital Mammography Saves Technologists’ Time, Increases Radiologists’
A recent study found that although digital mammography saves technologists’ time compared with film-screen mammography, it increases physician time from 1.4 minutes to 2.3 minutes—a 57% increase. Technologists spent an average of 21.6 minutes acquiring analog mammograms, compared to 14.1 minutes for digital mammograms, resulting in a 35% time savings.
Shown here is the average time in minutes, for interpreting screen-film (green) and digital (blue) mammograms for each physician. Rad=radiologist. |
“The use of digital mammography for screening examinations significantly shortened acquisition time but significantly increased interpretation time,” states the report, which was featured in the July issue of the American Journal of Roentgenology.1 “In addition, more technical problems were encountered that delayed the interpretation of digital cases.”
Researchers at Northwestern Memorial Hospital (Chicago) tracked how long it took several technologists from examination initiation to release of the patient for 100 screen-film and 100 digital mammography screenings. The reason for a decrease in time for technologists: Digital acquisition takes less time because medical staff can stay in the room during the procedure. Also, film-processing time is eliminated.
To evaluate interpretation time, seven radiologists—four experienced breast-imaging radiologists and three breast-imaging fellows—reviewed 183 hard-copy screen-film cases and 181 soft-copy digital cases. “It simply takes longer for a radiologist to read a digital case,” says Eric Berns, PhD, research assistant professor in the Department of Radiology at Northwestern University and lead author of the study. “That’s due to the length of time it takes for them to navigate through the workstation to read a digital image.”
Berns believes the design of workstations could be improved. “There needs to be a more intuitive design,” he says. “It needs to be self-taught—you should not need a lot of training. So, ease of use is one key feature. The other key feature is that when there need to be adjustments or when the radiologist needs to read images that don’t fit exact protocols, the system should easily adapt, easily incorporate random images or extra images into the images that the radiologist is reading.”
He also notes that the new digital systems hinder optimal workflow for radiologists. “The biggest issue is workflow, hands down, in terms of how these systems integrate into PACS, integrate with other digital modalities, with other digital mammo systems, with other vendors,” Berns says. “Integration and workflow are frustrating points.”
Berns says that film priors are an encumbrance as well. “If we could digitize our priors, it would be a great help,” he says. “One of the challenges is how to digitize and then label images so that they’re presented in the right orientation. There’s a lot of labor and cost that goes into that. But it’s inevitable, and it needs to happen.”
The report noted that its conclusions on screening mammography may not apply to diagnostic mammography.
Reference
- Berns EA, Hendrick RE, Solari M, et al. Digital and screen-film mammography: comparison of image acquisition and interpretation times. AJR Am J Roentgenol. 2006;187:38–41. Available at: www.ajronline.org/cgi/content/abstract/187/1/38. Accessed September 20, 2006.
Nuclear Medicine Use Increases 22% in 13 Years, Report Shows
A report recently released by IMV Medical Information Division (Des Plaines, Ill) reveals that 19.7 million nuclear medicine procedures were performed across 17.2 million patient visits in 2005, up 15% from 2002, forming an average annualized rate of increase of 5% per year over the period. Cardiovascular studies, including cardiac perfusion, accounted for 57% of the patient visits.
“Nuclear medicine utilization, not including PET procedures, has been driven by cardiovascular applications, which have grown from 35% of 1992 procedures to 57% in 2005,” said Lorna Young, senior director of market research at IMV Medical Information, in a press release. “Nonhospitals, such as cardiology practices, have focused on cardiovascular procedures, where 80% of their nuclear medicine patient studies are cardiovascular. Although both the hospitals and nonhospitals are equally likely to perform cardiovascular procedures, hospital sites are more likely to perform tumor localizations; radionuclide therapy; bone scans; and liver, respiratory, and renal studies.”
The report’s other findings include the following highlights:
- The average number of nuclear imaging cameras installed in each site is 1.8.
- Two thirds of the camera installations at nonhospital locations occurred during or after the year 2000, compared to only 45% of those installed at hospitals.
- Replacement activity is ongoing: more than two thirds of purchase activity at nuclear medicine sites involved replacement units.
- Nuclear medicine sites are expanding network capability to encompass images from catheterization laboratories, CT, MRI, echocardiography, and general ultrasound.
The report uses the 2005 edition of the Nuclear Medicine Census Database to analyze changes in utilization. The database—profiling 5,500 of the 7,200 identified nuclear medicine facilities in the United States—includes more than 3,000 nonhospital locations.
For more information, visit www.imvlimited.com or call (847) 297-1404.
NYU Medical Center Rolls Out Intuitive Voice Recognition System
The Department of Radiology at New York University (NYU) Medical School began beta-testing the RadWhere voice recognition system from Commissure Inc, New York, more than 1 year ago. Now, a scheduled rollout of the system is in progress, beginning with the neuroradiology section, which has been using RadWhere for about 8 months, and continuing across the department to eventually supplant the legacy program, which is nearing the end of its life cycle.
“The program fills in the blanks and generates a formal report. So, all you’d have to say is ‘five centimeter frontal lobe mass’ and boom, you’ll get an entire report out of that.” —Edmond Knopp, MD New York University Medical School |
RadWhere features real-time speech reporting, mining clinical data as the radiologist speaks to enable access to relevant data during diagnosis; the program also automatically tags reports with the proper ICD-9 codes. “On the speech end, we’ve seen increased accuracy,” says Edmond Knopp, MD, neuroradiology section chief at NYU. “And when interacting with and using it, various things make it a little more user-friendly. We’ve dictated accession numbers into it. Also, when using macros, the macro is in one text color and dictation is in another, so when proofing, you look at it in only one color, which shortens things.”
Knopp says Commissure’s claim that its programs “embrace, not replace” existing IT infrastructure has been proven in his department. “It completely interfaces with our RIS, and it’s also interfacing with [our old voice recognition system],” Knopp explains. “We’re doing this staged rollout, so the neuroradiology section is using it; the other sections are not. If residents read cases on call at night, they’re dictating into [the old system]. When I open the case in RadWhere, the preliminary is there. It interfaces seamlessly.”
Knopp also looks for exciting developments in the future as the radiology department adapts to the new system. “We’re headed to the findings-only mode, where you come up with smart templates and basically only dictate what’s abnormal. The program fills in the blanks and generates a formal report,” he explains. “So, all you’d have to say is ‘five centimeter frontal lobe mass’ and boom, you’ll get an entire report out of that.”
The development of the smart templates initially will be a challenge for department members, but the potential for long-term time savings cannot be ignored. “It’s going to be a pain trying to generate all of these smart templates,” he says. “And it’s work that the radiologists are going to have to do, because only they know how they want their reports to be. But once it’s done, it’s done. It’ll be a lot easier once we retrain our minds to dictate that way.”
Knopp hints that smart templates could simplify electronic medical records. “In all honesty, if generics were used across the board, it would simplify the entire electronic medical record,” he says. “Everything would always look the same, in one template, to the referring physician.”
But he is unsure whether radiologists are ready for the template revolution: “I would imagine that people are going to want it their own way.”
—C. Vasko