The difficulty of regulating x-ray radiation in the ER

A recent article in the New England Journal of Medicine touched off another salvo about how non-clinicians have no problems judging the abilities of clinicians in the world of medicine.

The article begins by presenting the case of a woman who awoke with facial paralysis and then went to the emergency department. On arrival, she received CT scans and MRI scans of her brain. When those were normal, she was diagnosed as having Bell’s Palsy. Two weeks later, she developed hair loss and other symptoms and it was found that during her first ED visit, the radiology department mistakenly exposed her to 100 times the normal dose of radiation for a brain CT scan.

She now has a federal class action suit pending against the CT scanner manufacturer and a medical malpractice lawsuit pending against the treating physicians.

The author of this article then uses her own calculations to conclude that the “risk of cancer from a single CT scan could be as high as 1 in 80 — unacceptably high.” The study she cites shows that radiation doses for the same tests vary – as they should. Giving the same dose of radiation to a 90 pound grandma and a 500 pound grandson would result in at least one uninterpretable study. Based on the 4 hospitals they studied, they disputed the risk of cancer being 1 in 2000 from a CT scan and stated that the risk to a 20 year old woman from a single chest CT or single multiphase abdomen and pelvis CT could be as high as 1 in 80.

The paper advocates lessening and standardizing radiation doses for examinations, noting that the improved image quality obtained with higher radiation doses often has no change in clinical outcomes. Diagnostic accuracy would not be affected if the radiation dose were reduced by 50%.

The paper also suggests tracking a patient’s dose of radiation over time and including that measurement in the medical records. Great idea, but how is a radiation dose from someone living in California going to help me when the patient has a potential cervical spine injury in my town while she’s on vacation?

Finally, the author suggests that we need to reduce the number of CT scans being performed. Each year 10% of the population receives a CT scan and 75 million scans are conducted each year – with the rate growing more than 10% annually. At the heart of the increased number of scans is “increasing ownership of machines by nonradiologists” and the resulting “self-referral” which increases the use of the scanners. Those bastard non-radiologists. Only radiologists should be able to self-refer and get away with it.

In general, I think that Dr. Smith-Bindman is on point with her suggestions. It would be great if patients’ total radiation doses could be tracked throughout their lives. However, assuming that could happen, would a high dose of radiation make any difference in determining whether an 80 year old lady with abdominal pain got a CT scan? How about in determining whether a hypotensive unconscious 50 year old trauma victim should undergo CT scanning? What about deciding whether the obese 30 year old complaining of severe difficulty breathing should get a chest CT to rule out a pulmonary embolism?

Can we reduce radiation dose at the sacrifice of less clear scans? That’s a radiologist’s call. Is Dr. Smith-Bindman following her own suggestions? If we missed a small nodule that later became metastatic cancer, would the defense that “at least the patient didn’t get as much radiation” be a sufficient defense in a medical malpractice trial?

The suggestions are good, but they don’t apply to clinical practice.

In addition, while the FDA does regulate “radiation-emitting electronic products” including diagnostic x-ray equipment, telling patients how many diagnostic radiographic studies they may “safely” obtain is likely an area of mission creep for the agency – akin to regulating how many hours of television people may watch in a day (yes, television receivers emit radiation) or how many hours George Hamilton may spend under his radiation-emitting sunlamp. I’m not so sure that having the FDA limit the number of scans a patient can receive is a good thing.

I also take issue with Dr. Smith-Bindman’s statistics that demonstrate greater than a 1% incidence of developing cancer from a single CT scan. If 1 in 80 patients can get cancer from a single CT scan and almost 80 million CT scans are performed every year, each year we are causing close to 1 million cases of cancer in US citizens. According to the American Cancer Society, it is estimated that 1.5 million total cases of cancer will be diagnosed in the US in 2010.

Is our use of CT scans really causing more than half the cases of cancer in the US each year? Even if we cut the incidence in half, causing 500,000 new cases of cancer each year is a hard allegation to substantiate.

While the number of CT scans is allegedly increasing at 10% per year, the number of new cancer cases in the US was 1.22 million in 2000 and is 1.53 million in 2010 – hardly an increase of 10% each year over 10 years. Those increases in new cancer cases also paralleled an increase in the population size – from 281 million in 2000 to 305 million in 2009. On a per capita basis, the incidence of newly diagnosed cancer went from 4.3 per thousand to 5 per thousand during those nine years.

Dr. Bruce Hillman wrote an accompanying article citing how “an unknown but substantial fraction of imaging examinations are unnecessary and do not positively contribute to patient care.” As some causes of the unnecessary use of diagnostic imaging he cites patients who “pressure their physicians to refer them for imaging studies even when imaging is unlikely to provide any value.” He also cites defensive medicine, self-referral, medical training programs that ingrain “shotgun” diagnostic testing to confirm diagnoses with the “greatest possible certainty.” He also acknowledges that radiologists also share in the blame for fueling the explosion in diagnostic imaging.

His ideas for changing the system are much more realistic and include tort reform, better physician and medical student education, engagement of radiologists as consultants, and “a change in mindset among physicians.”

I agree with Dr. Hillman on every point except the last one. Physicians have a “mindset” that is created by attorneys and by the public. In most cases, physicians are expected to be perfect or to exhaust all possible testing in finding a diagnosis (I still haven’t had one person who disagrees with me on this point present me with a diagnosis that it is okay to miss). If a diagnosis is missed, the lack of “appropriate” testing that would surely have made the diagnosis is a central theme in the physician’s malpractice trial.

We don’t need to change the physician’s mindset. We need to change the public’s mindset. If less testing is performed, more people will have diseases that won’t be diagnosed … or that won’t be diagnosed early enough. That is an inevitable result of reducing the use of diagnostic radiologic testing.

Will the public and the juries sitting in medical malpractice trials accept this fact? Can we say that a 95% possibility you don’t have a deadly disease or severe injury is “good enough” and non-actionable? Until society makes the commitment to lower the bars over which clinical physicians must jump, the incidence of diagnostic imaging – and all the radiation that accompanies it – will go up and not down.

There were many news articles published about this study, including USA Today and Forbes.com.

I didn’t see any clinical physicians interviewed in these articles – only radiologists. That should be the first clue that something is amiss — news articles with non clinicians commenting on how clinicians should do their jobs. How confident would we be if USA Today encouraged readers to pick up the paper next week when USA Today will have chemical engineers commenting on how mechanical engineers should stress test products more thoroughly? Hey – they’re both engineers, right?

The money quote from Dr. Hillman in a Reuters article really irked me, though: “We need to convince physicians that a quest for certainty is impossible, costly and is harmful because of indirect diagnoses.”

If radiologists are so certain that diagnostic imaging doesn’t need to be done, then cancel the test. That’s right. You think a test is a prospective waste of radiation? Refuse to perform it. Right now, you’re talking the talk, but you’re doing a face plant on the concrete when you try to walk the walk.

How about this: When the dumb ER doc orders the next total body scan, walk over to the emergency department, examine the patient, and come up with your own diagnosis without using your deadly CT scanner. Get rid of your hindsight bias and make a prospective diagnosis without having the benefit of a “normal” diagnostic test sitting on the computer screen in front of you.

Isn’t as easy as your little news sound bites make it seem, is it?

Want to regulate something to really stop the flow of radiation into patients? How about making the American College of Radiology start regulating the number of diagnostic radiology reports from their members that contain phrases such as “cannot rule out underlying lesion, recommend CT scan for comparison,” and then “CT scan non-diagnostic, recommend bone scan for further clarification”. We could cut down on costs if radiologists would stop recommending “MRI for clinical correlation,” also.

I’m betting that we won’t see too many sound bites about the implications from this radiology report lingo hitting the headlines any time soon.

WhiteCoat is an emergency physician who blogs at WhiteCoat’s Call Room at Emergency Physicians Monthly.

Submit a guest post and be heard.

View 19 Comments >