Accuracy in radiology can be difficult to gauge; even interpretation, as the term for a radiologist’s work on an imaging study, suggests shades of gray. To quantify accuracy better, 12 years ago, the radiology department at Monmouth Medical Center (Long Branch, New Jersey) launched a project in which it cataloged its attending radiologists’ interpretations based on RADPEER™ scores. It also archived all resident on-call interpretations and attending over-reads.
Richard Ruchman, MD, chair and program director of the department, says, “I was interested in coming up with a way to measure how well residents do when they are on call. We came up with a framework for analyzing discrepancies between two interpretations of a study, and that got me thinking about how we do quality control in radiology.”
Making the most of RADPEER, the department began accumulating data on the accuracy of work done by radiologists and radiology residents. The RADPEER system rates discrepancies between interpretations on a four-point scale, with A and B modifiers based on clinical implications; a grade of 1 means no discrepancy, while a grade of 4B means a major discrepancy with the potential for an adverse clinical outcome. “You accumulate all the data and compare to see where every radiologist falls,” Ruchman says, stressing that a certain discrepancy rate is expected. “Then, you take a deeper dive,” he adds.
The findings are discussed in a monthly meeting, where major discrepancies are analyzed so that the entire department can learn from them. Ruchman, whose background includes medical education, notes, “All physicians are mandated to measure their performance and outcomes. We keep data on all of our misses and look at the trends, and if there is a finding that a lot of people have been missing, we pay extra attention to it.”
Certain findings are missed more often than others, Ruchman says. “There are definite patterns, and that’s one of the frustrating things. Very often, the things that are missed don’t change that much from month to month. People tend to miss the same things: We are all programmed, for whatever evolutionary reason, to look in certain areas, and forcing yourself to look in others breaks a natural perceptual ability. That’s the tide we are swimming against,” he says.
In a new quality project, Ruchman and colleagues are looking at the quality of the imaging information that is sent with a patient who is transferred to another facility. “We have made the results of our research available to the transferring hospital for it to put in a process change,” he says. “Then, we plan to resurvey it to see if this has had an impact.”
Communicating Improved Quality
Ruchman is of the opinion that radiologists will need to communicate their quality proposition more strongly in the future, particularly with the advent of new payment and delivery models that risk commoditizing the profession. “In the past, the radiologist was someone sitting in a dark room reading films, with very little patient and physician contact,” he says. “All of that has changed, for a number of reasons. In the current environment, the radiologist who is anonymous is a radiologist who can be easily replaced.”
For that reason, Ruchman participates in surgery morbidity/mortality meetings at Monmouth Medical Center weekly, and he recommends that radiologists find similar ways to integrate themselves more deeply into the continuum of care. “You may go in and speak to patients and discuss their results with them, or you may go into departmental meetings,” he says.
The result of his participation in meetings is that the surgical staff’s members, he notes, “know that I am the consultant for them, for every one of their cases. They know they can give me a call to discuss anything, and that I will contribute whatever expertise I have.” He is considering sending his department’s radiologists on hospital rounds with other physicians, meeting with patients to discuss their care.
“The whole concept of the radiologist has really changed,” Ruchman concludes, “and referrers absolutely see when we make those changes. They are going to be much more likely to send a case to someone they can discuss it with—someone who is not just an anonymous person, sitting in a silo, reading films.”
Cat Vasko is editor of RadAnalytics and associate editor of Radiology Business Journal.