AuntMinnie.com is pleased to present the next installment of Leaders in Imaging, a series of interviews with individuals who are shaping the radiology landscape. We spoke with Dr. Rebecca Smith-Bindman, professor of radiology and biomedical imaging; epidemiology and biostatistics; and obstetrics, gynecology, and reproductive medicine at the University of California, San Francisco. Smith-Bindman discussed the need to standardize radiation doses, particularly for advanced imaging procedures such as CT.
AuntMinnie: You've published studies addressing the issue of high radiation doses in imaging, especially CT. How would you describe the problem?
Dr. Rebecca Smith-Bindman from the University of California, San Francisco.When we began reviewing the radiation dose used for CT, it became clear that the use of multiphase studies was increasing. Thus, the greater use of imaging and higher dose per examination was driving up patients' exposure. Our numbers were a bit surprising -- no one was paying attention to how often we use these more sophisticated protocols that lead to higher doses.
My research is not only focused on radiation dose amounts, but also on trying to figure out how we can use imaging more effectively. What actually happens to patients being imaged? Are they helped or harmed? It's ironic that with the improved capacity to see smaller and smaller lesions on imaging exams, we're finding more and more, but we don't know if the findings really matter. We then often order follow-up imaging because the lesion is so tiny we don't know what we're looking at or if it has relevance to the patient. As a clinical radiologist myself, I find it striking how often we do studies without really questioning whether they're necessary.
What prevents radiologists from talking frankly about radiation and its potential for causing harm?
I think radiologists fear that if we describe radiation risk to our patients, it will lead to a reduction of imaging -- and patients won't get necessary scans; for example, scans performed in acute trauma settings. But patients are smart, and if we educate them appropriately, I believe they will make good choices. Patients are asked to make intelligent choices throughout medicine, to balance risks and benefits, and there's no reason they can't do that with imaging, too. This may lead some patients to forgo some tests, and this is appropriate, given the high number of tests currently ordered that are simply not necessary.
Radiologists are concerned about other specialists moving in on our turf. There are clearly legitimate concerns regarding self-referral issues in other specialties driving up unnecessary imaging, but it's also self-referral in my mind when radiologists recommend a follow-up exam that they truly believe is unnecessary, such as for liver hemangiomas or renal cysts.
But taking the lead in dose reduction and safety will very much boost our position [and] give us something to brag about. I think there's an opportunity here: Rather than seeing the dose reduction issue as an attack, we can see it as a way to come out ahead.
How does "defensive medicine" contribute to high radiation doses?
The practice of defensive medicine to avoid malpractice lawsuits exposes patients to more imaging and radiation than necessary. It's a commonly reported practice, and we find higher imaging rates in areas with greater malpractice concerns and a more hostile malpractice environment. More work needs to be done to address concerns about malpractice, and we need a broader approach than state-by-state tort reform. But I don't believe malpractice law reform is where the ultimate answer lies in terms of reducing imaging use or dose.
What can radiologists do to reduce unnecessary dose?
It would behoove us as radiologists to consider possible harms from imaging, whether it's from radiation dose, overdiagnosis, or overutilization and the associated costs. We need to ask ourselves whether the patient is going to benefit from studies we do.
This is not a gatekeeping role, but an educational one, and we need to educate our referring physicians: Radiologists have not been proactive with referring doctors. Obviously, we will not convince all doctors to avoid unnecessary tests, but we should try to reduce those exams where we see no value. We currently do about 75 million CTs per year, and it's widely believed that as many as 25 million of those are unnecessary. So the risk-benefit discussion is key.
But then there's how we perform CT exams. Currently, we use an enormous number of different protocols. Every radiologist at every institution may use a different exam protocol. If you take, for example, a patient with chest pain, suspicious for a pulmonary embolism, depending on where that patient is sent for imaging the radiation can vary 10-, 20-, 30- to 100-fold based on how individual radiologists choose to image the patient.
This makes no sense to me. Facilities and physicians have to take responsibility for how these examinations are done to eliminate situations where we use higher doses than necessary, and reduce variations. There's no excuse, in my mind, not to be doing this immediately.
California recently passed a law that requires reporting of radiation dose delivered during CT scans. Does this kind of legislation help?
I believe it's a huge step forward. The California law will take effect in 2012, and it requires that the radiation dose used for CT exams be included in a patient's medical record. From this, researchers could start to collect data across institutions that would help us get our doses down. We really don't know what doses we tend to deliver to our patients and how we quote the dose for CT: 10 mSv for an abdominal scan and 7 mSv for a chest scan is grossly inaccurate and bears little resemblance to the doses patients receive in actual practice. Talking about the dose patients should have gotten if they were plastic phantoms, or if they had a limited-region scan (when they actually had their entire body scanned), isn't terribly relevant to quality or safety.
Right now I'm working on trying to standardize CT dose protocols across the University of California system. It involves a close collaboration between doctors, physicists, and technologists, and my contribution is trying to get a lot of extremely smart people, with differing expertise, around one table.
Part of the approach is to get everyone in the room and show them the doses they use and how these vary by protocol choices that are made. Then they get to argue about what is the best way to answer a particular clinical question, and if a doctor feels one protocol that has a high dose is particularly useful, she can argue why it's "worth" three times the radiation dose as the same question imaged another way.
You submitted a radiation dose measuring metric to the National Quality Forum (NQF). What was the response?
Yes, I submitted a metric to NQF that calls for facilities to measure the radiation dose they use in actual patients by collecting clinical data on consecutive patients, rather than a handful of specifically chosen cases. The steering committee initially passed the measure easily, but then there was concern expressed that the measure was not accurate or precise enough, and that reporting dose information could discourage imaging.
I don't accept these criticisms at all. We have so much evidence right now that people are getting doses much higher than necessary, and the best way to improve this is to have facilities monitor dose. If the quality measure passes, and if it's adopted as part of groups concerned with quality, such as the Centers for Medicare and Medicaid Services or the Physician Quality Reporting Initiative, radiologists could even be financially rewarded for studying their doses.
Should scanner manufacturers have a role in reducing radiation dose?
There are two issues with suboptimum radiation dosing for CT. One is severe radiation overdosing, as patients have received in brain perfusion scans -- which are actually a very small number of exams, by the way. Then there are routine examinations that deliver higher doses than necessary.
Some people say that the manufacturers should address the first issue by making the machines incapable of delivering incorrect doses. This is one area the [U.S. Food and Drug Administration (FDA)] is exploring, putting fail-safe mechanisms in place so it would be nearly impossible to deliver a dose akin to radiation treatment for brain cancer when a diagnostic CT is done. But the issue of ensuring that day-to-day doses patients receive are as low as possible is not being addressed by manufacturers. There are software products to lower dose, but they are priced so high that many facilities can't purchase them.
Manufacturers have one constituent, and that is the people who own the company. One of the ways they have sold CT devices is to develop the capability to produce more beautiful, higher-dose images, and it's harder to move toward noisier, lower-dose images now. Again, there are new ways the manufacturers can help us lower doses immediately, but I don't see this happening fast enough, or comprehensively enough. Radiologists need to be in the forefront of standardizing dose, and that requires acting as a group to come up with appropriate standards.
