Second Take: Must patients really be warned of radiation risk?

2013 07 30 12 53 25 294 Second Take Am 200

Second Take is a series of polemics edited by Drs. Saurabh Jha and Jason Itri that discusses controversial issues in medical imaging, offering both sides of an argument by deliberately taking a one-sided view. It is hoped that the audience will appreciate both the duality of the argument and the difficulty in resolving the issue. The views expressed by the authors are not necessarily the views held by them.

In their Second Take column, Drs. Ian Amber and Po-Hao Chen make the case that informed consent should be procured from patients before they undergo medical imaging procedures that involve ionizing radiation, such as CT. For a variety of reasons, we believe this scenario represents not only a burden to radiology, but it also does not live up to the standard of quality care that radiology professionals should provide.

History of informed consent

The idea of "informed consent" was proposed by a lawyer in 1957 during a medical malpractice case at Stanford University in which a surgeon inadequately warned a patient of surgical risk. According to the American Medical Association (AMA), informed consent should involve open communication, allow for patient autonomy, and invoke patient-centered decision-making.

Specifically, according to the AMA, informed consent should discuss the following:1

  • Diagnosis (if known)
  • Nature/purpose of treatment/procedure
  • Risks/benefits of treatment/procedure
  • Alternatives and associated risks/benefits of alternative treatment/procedure
  • Prognosis if the discussed treatment/procedure is not performed

Unfortunately, over the past 60 years, informed consent is no longer a platform for open communication between patient and physician. Much too often it becomes degraded into one-sided rhetoric typically reserved for the "aura" of legal protection. In fact, several prominent lawsuits involving informed consent have been upheld in the plaintiff's favor despite satisfying the above requirements established by the AMA.2,3,4

It's become clear that informed consent has lost its communication value and doesn't necessarily protect us legally. So therefore why are we contemplating required informed consent from patients before performing CT scans?

Difficulty quantifying radiation risk

For a proper informed consent, the benefits and risks should be clearly stated and unequivocally and accurately measured, such that the patient can make a clear, level-headed decision.

If a patient with pancreatic cancer is a candidate for a Whipple procedure, the department of surgical oncology at Johns Hopkins University publicly advertises a 2% risk of death.5

When an oncologist proposes chemotherapy for patients, quantitative survival curves are discussed with and without chemotherapy.

In these examples, the patient can then make a calculated, informed, and educated decision, weighing the benefits of any procedure or treatment against the risks.

Unfortunately, establishing a precise radiation risk from a single CT scan is exceedingly complex and inaccurate. The calculation of the increased risk of cancer from a CT scan is a two-step process.

First, one has to know the radiation dose to which the patient was exposed. Second, this dose is compared to large population data to estimate the added risk of future malignancy. This two-step process, however, is highly inexact.

Although every CT scan performed results in a dose-length product that can be converted into a radiation dose (typically in mSv), it is important to note that that is not "truly" the dose that your patient received. These dose measurements are calculated from a plastic cylinder phantom, with which radiation doses are calibrated on each CT scanner.

Unfortunately, given the heterogeneity of our patients and widely varying body habitus, studies quote as much as a twofold variation in actual dose delivered from the reported dose-length product provided by the CT scanner.6

Second, the conversion of radiation dose to the increased relative risk of malignancy from a single CT scan is also imprecise. Our current knowledge on the risk of malignancy after radiation exposure is largely based on large population studies of atomic bomb survivors during and after World War II who were exposed to exceedingly high radiation levels, orders of magnitude higher than our current diagnostic radiation levels.

Mathematicians have therefore used linear regression models to theorize the population-based malignancy risk with the low levels of radiation commonly experienced in diagnostic imaging. Use of these radiation risk models is theory-based and, importantly, is not meant for use on an individual patient.

In summary, we cannot optimally consent patients if we cannot predict radiation risk.

Skipping medically necessary diagnostic imaging

The relatively low levels of radiation exposure with diagnostic imaging result in predominantly stochastic effects such as induction of cancer, rather than deterministic effects such as skin burns or cataracts. The topic of cancer in particular elicits an incredibly emotional response, as nearly everyone has experienced, directly or indirectly, the wrath of cancer treatments or death of loved ones from malignancy.

One of our concerns is that forcing patients to sign and authorize their exposure to radiation may cause them to overestimate the true risk of cancer (which we cannot adequately quantify in the first place!). This could lead them to forego diagnostic examinations that have clear and measurable benefits.


The logistics of implementing a plan to routinely consent patients for CT scans is daunting. Using the Hospital of the University of Pennsylvania as an example, for the one-year period from June 2011 to July 2012, 74,835 CT scans of all body parts were performed at this large academic center, according to the Pennsylvania Department of Health.7

If each informed consent requires five minutes of a radiologist's time, this would have added 6,200 hours of new work annually. This amounts to several additional full-time radiologists, purely for the time spent consenting patients.

Of course, keep in mind that Medicare and private insurers would not reimburse for this added work, resulting in an immediate fiscal shortfall for radiology departments around the U.S.

Alternatively, the added responsibility could be distributed to more financially affordable employees, such as radiology residents, where available. In a program the size of University of Pennsylvania's, with roughly 40 residents, this would amount to 150+ hours annually of additional consents (in addition to routine MRI/implantable device consents). This may require a new month-long rotation of "consents" to the radiology residency requirements, and keep in mind this would be a 24-hour, around-the-clock responsibility.


While we certainly understand the rationale for the Second Take argument by Drs. Amber and Chen, unfortunately, it is grossly impractical and, as discussed, quantifying radiation risk is highly inaccurate.

But while we disagree with implementing informed consent for all CT imaging, we are not advocating that the radiology community take no action on this issue. We should embrace this as an opportunity to be leaders with the equipment and technology that we employ.

Rather than informed consent, we propose that the American College of Radiology (ACR) and local radiology societies should more effectively educate our physician colleagues about the large-scale, community-wide implications of exposure to medical radiation.

Additionally, we should continue to educate the public through educational resources that are already well-established, although perhaps not well-advertised. Programs from the ACR and RSNA, such as Image Gently and Image Wisely, are fantastic educational tools to explain radiation risk to our patients.

Do we truly need more paperwork and "yellow tape" obstructing the delivery of quality patient care?

Drs. Freed, Benjamin, and Rodriguez are radiology residents at the University of Pittsburgh.


  1. American Medical Association. Patient/physician relationship topics: Informed consent. Accessed September 24, 2013.
  2. Murray, Bryan. Informed consent: What must a physician disclose to a patient? American Medical Association Journal of Ethics. 2012;14(7):563-566.
  3. Johnson v. Kokemoor, 545 N.W.2d 495, 199 Wis. 2d 615 (Wi 1996).
  4. Nixdorf v. Hicken, 612 P.2d 348 (Utah 1980).
  5. Johns Hopkins Medicine, Sidney Kimmel Comprehensive Cancer Center. The Whipple procedure and other pancreas surgeries. Accessed September 24, 2013.
  6. McCollough CH, Leng S, Yu L, Cody DD, Boone JM, McNitt-Gray MF. CT dose index and patient dose: They are not the same thing. Radiology. 2011;259(2):311-316.
  7. Pennsylvania Department of Health. 2011-2012 hospital reports: CT scanner and MRI data. Accessed September 26, 2013.
Page 1 of 656
Next Page