Jenifer Siegelman, MD, a radiologist with Norwich Diagnostic Imaging (NDI), had noticed something disturbing about the CT images coming out of the hospital her practice serves. Siegelman has been the director of CT at William W. Backus Hospital in Norwich, Connecticut, for eight years, and, she says, “What I have seen is that there is variable image quality coming from various CT systems the hospital has purchased through the years, and there are variable radiation levels being used. There was no real way to keep track of what was happening or figure out whether there were any trends.”
With the backing of the community hospital’s administration—Siegelman notes that “as new innovations come through, the hospital is very interested in doing what is best for patients at every juncture”—she launched an initiative to look into the variances in image quality and radiation dose. “In some cases, the quality issues were leading to patients receiving repeated scans,” she recalls. “I felt a personal responsibility to the patients, and to the clinicians taking care of them, to come up with more certain answers. It’s not desirable to have a radiologist who isn’t sure of what he or she is seeing.”
Variability and Quality
Siegelman, who holds a master’s in public health, observes that “variability is a marker of a place where quality can be improved—particularly when the variability isn’t explainable by medical necessity.” The team at Backus set out to reduce variability in image quality and dose, but first, baseline data were needed—and there was no ready source for it or automated solution with which to aggregate it.
Mahadevappa Mahesh, PhD, chief physicist at Johns Hopkins Hospital, recommended that Siegelman and her team begin by identifying the most commonly performed CT exams, as focusing their efforts on those exams would have the biggest impact. NDI’s billing company supplied the requisite information, and Siegelman compared it against ACR national benchmark figures utilized in its accreditation program. “The basic premise of quality improvement is comparing to benchmark and seeing how you’re doing,” she says. “For us, that process had to be completed manually and required significant time.”
With the hospital’s support, technologists working the overnight shift began manually pulling data points, including CT dose index volume and dose length product, during their downtime. “We evaluated variability in a very systematic way,” Siegelman says. “We also had the technologists keep track of how much time it took them—we wanted to know how much work was involved.” When it became clear that the significant time investment was paying off, the hospital was willing to invest in new technology, including iterative reconstruction techniques and software from Radimetrics that aggregates dose information automatically.
Empowered with these tools, Siegelman and colleagues have begun to quantify the changes in Backus’ CT dose protocols to see if they have had an impact on image quality and radiation dose consistency. Most importantly, she says, the implementation of the Radimetrics tool has enabled faster and more robust improvement. “We know how much time we put into improving these protocols previously, and we’re measuring to see how much time we’re putting into it now,” she says. “We expect to see a significant change in the improvement—it will be broader, easier, and more substantive.”
As a result, she says, staff members at every level have been engaged in the dose reduction initiative. “Having the Radimetrics tool in our organization allows us to have continuous teaching, even outside of the dose committee,” she notes. “The dose committee meets in a conference room, but now it has tentacles reaching into every aspect of our department—now, our medical physicist interacts with our technologists, and they are educated beyond a level we never could have dreamed of before.”
One of the most striking changes has been to the level of engagement from hospital leadership, Siegelman adds. “Organizational leaders are used to dealing with compliance-related matters when it comes to quality and safety, and there are hard rules there,” she says. “With radiology, it’s not quite as clear-cut. It was previously assumed that everything was okay as long as nothing bad happened. Radimetrics has given me a very facile way of showing what the issues are, so that I can communicate to our administration which processes