N Brown BSc, MBBS, MPhil, GCSpMed; L Jones BSc (Hons).
Knowledge of medical imaging radiation dose and risk among doctors
Article first published online: 28 DEC 2012
© 2012 The Authors. Journal of Medical Imaging and Radiation Oncology © 2012 The Royal Australian and New Zealand College of Radiologists
Journal of Medical Imaging and Radiation Oncology
Volume 57, Issue 1, pages 8–14, February 2013
How to Cite
Brown, N. and Jones, L. (2013), Knowledge of medical imaging radiation dose and risk among doctors. Journal of Medical Imaging and Radiation Oncology, 57: 8–14. doi: 10.1111/j.1754-9485.2012.02469.x
Conflict of interest: There are no conflicts of interest for either author.
- Issue published online: 4 FEB 2013
- Article first published online: 28 DEC 2012
- Manuscript Accepted: 11 JUN 2012
- Manuscript Received: 22 MAR 2012
- computed tomography;
- diagnostic imaging;
The growth of computed tomography (CT) and nuclear medicine (NM) scans has revolutionised healthcare but also greatly increased population radiation doses. Overuse of diagnostic radiation is becoming a feature of medical practice, leading to possible unnecessary radiation exposures and lifetime-risks of developing cancer. Doctors across all medical specialties and experience levels were surveyed to determine their knowledge of radiation doses and potential risks associated with some diagnostic imaging.
A survey relating to knowledge and understanding of medical imaging radiation was distributed to doctors at 14 major Queensland public hospitals, as well as fellows and trainees in radiology, emergency medicine and general practice.
From 608 valid responses, only 17.3% correctly estimated the radiation dose from CT scans and almost 1 in 10 incorrectly believed that CT radiation is not associated with any increased lifetime risk of developing cancer. There is a strong inverse relationship between a clinician's experience and their knowledge of CT radiation dose and risks, even among radiologists. More than a third (35.7%) of doctors incorrectly believed that typical NM imaging either does not use ionising radiation or emits doses equal to or less than a standard chest radiograph.
Knowledge of CT and NM radiation doses is poor across all specialties, and there is a significant inverse relationship between experience and awareness of CT dose and risk. Despite having a poor understanding of these concepts, most doctors claim to consider them prior to requesting scans and when discussing potential risks with patients.
The critical role that radiology plays in modern medicine is expanding, and so too is the population radiation dose from medical imaging sources. Computed tomography (CT) and nuclear medicine (NM) imaging scans have greatly improved the diagnosis and management of a wide range of illnesses and their use has increased enormously over the last 30 years. When employed appropriately, the potential benefits of CT and NM greatly outweigh potential risks; however, injudicious use of these modalities when not clinically indicated is becoming a feature of modern medical practice.[1, 2] This results in unnecessary radiation exposures and possible increased stochastic effects for patients. Although radiation from a single CT or NM scan is relatively small and the individual risks may seem trivial, within the worldwide context of 3.6 billion diagnostic scans performed each year, even slight risks from inappropriate scanning can result in large numbers of adverse outcomes.[3, 4]
Young patients and certain tissue types have a greater susceptibility to radiation damage, highlighting the importance of ensuring appropriate use of X-rays and gamma rays in some patient groups.[5, 6] The average risk of inducing cancer from a CT chest is approximately 1 in 2500, but ranges from 1 in 1900 for a 5-year-old female to 1 in 9000 for a 65-year-old male; the lifetime cancer risk attributed to some NM positron emission tomography scans is much higher. The overall lifetime mortality from diagnostic radiation induced cancers is estimated to be between 0.6% and 3% above the normal baseline,[9, 10] and up to 29,000 future cancers may be caused annually by CT radiation exposure in the United States. In 2001, it was estimated that 500 cancer deaths arose from 600,000 CT exposures in children under the age of 15 years.
Best medical practices and ‘prima non nocere’ require doctors to have an awareness of the potential risks from CT and NM imaging to allow for valid risk/benefit evaluations and ensure that ionising radiation is used appropriately. The emergence of medicolegal proceedings in the United States concerning radiation exposure from CT scans, as well as escalating health care costs, add further relevance for all doctors. Furthermore, radiation doses from medical imaging sources have been the focus of recent attention from the public media and within medical literature.[13-15]
This paper quantifies doctors' knowledge and attitudes regarding radiation doses used in common medical imaging scans. While awareness within some sections of the medical community about the doses and potential stochastic (cell mutating) effects of diagnostic radiation has been shown to be poor, this is the largest survey of its kind and the first to assess knowledge across multiple specialties and experience levels.
In 2010 and 2011, an online survey was distributed to all doctors at 14 major Queensland public hospitals, as well as Queensland fellows and trainees in radiology, emergency medicine and general practice. The survey consisted of 12 questions, and respondents were asked to estimate the individual radiation doses from a non-contrast chest CT scan, a typical diagnostic nuclear medicine scan, a magnetic resonance imaging (MRI) scan and ultrasound scan (USS) compared to a standard chest radiograph (CXR). Respondents' knowledge of, and attitudes towards, potential health risks and medicolegal considerations of ionising radiation scans were also assessed.
Surveys that were more than 50% completed were considered valid. Overall and sub-group analyses by years of experience and area of medical practice were performed. Categorical variables were summarised using frequencies and percentages and analysed with Chi-squared tests. When sample size requirements were violated, Fisher's exact test was used. Binary logistic regression was also used to analyse data. Data were analysed using SPSS (version 19, IBM SPSS Statistics Version 184.108.40.206 (2010); SPSS, Inc., Chicago, IL, USA), using α = 0.05 (two-sided) to determine significance.
The average doses of a CXR, a non-contrast chest CT and the range of common nuclear imaging scans for adult and paediatric patients were obtained from data released by the Australian Radiation Protection and Nuclear Safety Agency (ARPANSA), Radiation Protection Division of the Health Protection Agency (HPA, UK), the National Council on Radiation Protection and Measurements (NCRPM, USA) and the Health Physics Society. Paediatric values were obtained from optimised dose readings from The Royal Children's Hospital, Brisbane.
The average radiation dose of an adult frontal CXR is 0.02 mSv; a single phase CT chest is 4–10 mSv; and the most common NM studies emit 1.7 mSv–17 mSv. Optimised paediatric, non-contrast CT chest mean effective doses range from 0.95 mSv (1-year-old phantom) to 1.72 mSv (5-year-old phantom). A single optimised paediatric CXR dose is 0.0038 mSv (infant), 0.0042 mSv (5-year-old) and 0.0113 mSv (15-year-old). Therefore, a CT chest scan is equivalent to approximately 200–500 CXRs for adults, 250 CXRs for infants and 400 CXRs for 5-year-olds. NM scans involve effective radiation doses equivalent to up to 830 CXRs. MRI and USS do not utilise ionising radiation.
Overall survey results
Responses were received from 624 doctors, of which 608 responses were valid. Specialty areas represented were radiology, emergency medicine, general practice, medicine, surgery, paediatrics, resident medical officers and other doctors (Table 1). Experience levels ranged from 1 year to >20 years of medical practice (Table 2), and all levels of experience were represented in the eight specialist groupings.
|Area of specialty||n (%)|
|Emergency adult||119 (19.6)|
|General practice||28 (4.6)|
|Years in practice||n (%)|
Only 17.3% of all doctors correctly estimated CT chest radiation dose compared to a CXR, and 78% underestimated the doses involved. Moreover, 54.9% of doctors underestimated by more than half, while almost 1 in 6 doctors (15.3%) underestimated CT dose by a factor of between 20× and >500× (Fig. 1). Radiation dosage of NM imaging was also poorly understood, with 35.7% of all doctors believing that NM scans involved radiation dose levels equal to or less than a chest radiograph. MRI and ultrasound scans were well known to not utilise ionising radiation by 97.7% and 99.3% respectively.
Ionising radiation was acknowledged to be potentially harmful by 94% of doctors, although almost 10% wrongly felt that CT radiation is in no way associated with any potential increased lifetime risk of cancer.
Although they possess poor understanding of the doses involved, and are therefore unable to accurately estimate potential risks, a majority of doctors claim that they still routinely factor the potential risks from ionising radiation into their decision to request CT scans. Just over one-third of doctors (35.5%) consider this only when they feel it is relevant, which contrasts against 4.3% of respondents who rarely or never think about the risks. Despite inabilities in quantifying doses and risks from CT radiation, 12.7% of doctors claim to discuss these risks with patients every time they request a CT study, 56.9% only when they feel it is appropriate and 8.4% only when the patient enquires about it. More than 1 in 5 (22%) never discuss these issues with patients.
Results by years of experience
A total of 431 doctors with ≤15 years of clinical experience responded, compared to 177 with greater than 15 years of medical practice. There is a statistically significant inverse relationship between clinicians' years of experience and awareness of doses and risks from CT radiation (Fig. 2). Doctors with >15 years experience have 2.1 the odds (confidence interval (CI): 1.2, 3.5) of incorrectly identifying the radiation dose of CT scans (P = 0.007), and have 3.2 times higher odds (CI: 1.8, 5.5) of not attributing medical imaging radiation with any increased risk of cancer (P < 0.001) compared to less experienced doctors (Table 3). Older doctors also have seven times the odds (CI:2.7,18.3) of believing that CT scans utilise the same or lower dose than a CXR (P < 0.001), thus underestimating the true dose of CT by a factor of >200. Even among radiologists, those with more than 15 years experience correctly estimated radiation doses of a CT only 17% of the time, compared to 43% for less experienced radiologists (Fisher's exact, P = 0.036).
|Years in practice||‘No increased lifetime risk of cancer from CT scans’ (%)|
Inaccuracy of NM radiation doses showed no relationship to years worked as a doctor 1.0 (0.7, 1.1, P = 0.887). The percentage of doctors who believed that typical NM scans either did not utilise ionising radiation, or involved doses that were equal to or less than a CXR was consistent across all experience levels, ranging from 34.9% to 36.5%. There was also good recognition across all experience levels that USS and MRI do not utilise ionising radiation (Table 4).
|Technique||Years experience||Utilisation of ionising radiation (%)|
|MRI (magnetic resonance imaging)†||0–15||2.8|
|USS (ultrasound scanning)‡||0–15||0.2|
|NM (nuclear medicine)§||0–15||78.4|
Results by area of practice
Analysis was performed after grouping respondents into specialty or areas of practice, including radiology, medicine (physician), surgery, intern/resident, emergency (adult), paediatric, general practice and other (Table 2). Specialties grouped in the ‘other’ category included psychiatry, pathology, anaesthetics, intensive care and medical administration. Orthopaedic responses were included in the surgical category.
Knowledge of CT radiation exposure was inaccurate across all medical specialties. Correct CT dose was recorded by only 32.9% of radiologists, 17.2% of interns/residents, 22.7% of emergency doctors, 12.9% of physicians, 10.7% of general practitioners, 18.8% of surgeons and 3.0% of paediatricians (Table 5). Interns/residents had the highest recognition that CT radiation was associated with increased lifetime risk of cancer (96.8%) followed by radiologists (93.4%), while surgeons were least likely to acknowledge the potential cancer risks (79.7%) (Fisher's exact P = 0.003) (Table 6).
|Area of specialty||Estimated dose of CT chest compared to CXR|
|Correct response: 200–500× (%)||Underestimate (%)||Overestimate (%)|
|Specialty||CT radiation associated with increased lifetime risk of cancer (%)†||NM utilises ionising radiation (%)‡|
Radiologists demonstrated the highest awareness that typical NM imaging doses are higher than CXR doses (89.5%), while paediatricians were least accurate (42.4%) (χ2 (7, n = 608) 31.9, P < 0.001). More than one third (34.4%) of surgeons who responded believed that NM does not utilise ionising radiation (Table 6).
The utilisation of rapidly evolving medical imaging technology has outstripped doctors’ understanding of the doses and risks associated with these technologies. These findings, from the largest published survey of doctors across all medical specialties, confirm prior studies assessing knowledge of radiation dose and risk within only specific groups of specialists. A 2010 study of Australian emergency doctors found that 78% underestimate the lifetime risk of malignancy from CT scans and a similar survey of paediatricians in Germany found that 94% underestimated CT dose. Both emergency physicians and radiologists underestimated the risks from CT scans in a smaller 2004 prospective study, and only 22% of patients reported being fully informed about the doses and risks of CT scans prior to their examinations. A more recent survey of 158 radiologists and non-radiologists revealed accuracies of 40% and 16%, respectively, when estimating doses of common radiological procedures.
In this context, the enduring growth of population radiation doses from medical sources and their associated risks, is concerning. Greater frequency of scanning, easier access to medical imaging, novel applications of CT technology, defensive medicine and patient expectations have all contributed to this increase. In Australia, the average population radiation dose from CT increased 50% between 2002 and 2008, and it has doubled worldwide in the last 10–15 years. In the United States, CT use rose more than 20-fold between 1980 and 2007, and the per capita effective dose from medical imaging sources rose 600%. A recent 5-year review of data at a paediatric hospital in North Carolina found that while emergency department presentations increased 13%, the number of CT chest scans increased 226%. In the same study, the use of CT increased between 51% and 463% for scanning various anatomical locations, and the authors concluded that it was uncertain whether the increased use of CT had resulted in improved patient outcomes. A 2009 retrospective study of 31,462 patients in Massachusetts found that 5% of patients underwent between 22 and 132 CT examinations each, and 4% were exposed to cumulative doses that increased their theoretical lifetime risk of a fatal cancer between 1.25% and 6.8%. A similar study of nearly 1 million patients in the United States found that approximately 2% are exposed to radiation doses many times higher than the recommended safe maximum doses.
Of particular concern is the inaccuracy of radiologists. As specialists who receive training in radiation dose, risk and safety, it was expected that radiologists would demonstrate the greatest accuracy when estimating CT scan dose. However, more than two thirds of radiologists still answered incorrectly, and greater experience was associated with a statistically significant decline in knowledge. Although the current convention within published literature and radiology training is to use the dose of a single frontal CXR when calculating relative doses, it has been suggested that older conventions (used by some more experienced radiologists) calculated doses based on frontal and lateral CXR views. Even if such an allowance is made, and the correct answer was extended to include doses of ‘100–500 CXRs’, still only 41.3% of radiologists with more than 15 years experience would answer correctly, and the odds of them being incorrect would remain 3.3 times greater than for younger radiologists. It must be noted that this survey included only 76 radiologists and a larger study is required to confirm the results in the wider profession.
Dose minimisation strategies are of greatest relevance to paediatric patients because of their increased sensitivity to radiation, and awareness of the stochastic effects of radiation is generally considered high among paediatricians. It was therefore surprising that paediatricians were the least accurate specialty in estimating CT doses (3%). A potential reason for this unexpected result could be that paediatricians inadvertently compare paediatric CT doses to the higher and more widely published adult CXR doses. However, almost 94% of paediatricians acknowledged that CT radiation is associated with an increased lifetime risk of developing cancer, indicating great awareness of the potential concerns. Only surgeons (79.7%) and general practitioners (89.3%) had an acknowledgement of CT's potential to cause cancer of below 90%.
Without a solid understanding of the radiation utilised by CT and NM scans, doctors are unable to weigh the risks against potential benefits, and may find it difficult to justify exposing patients to medical radiation and its associated risks. Ignorance of such issues may also prevent referrers from being able to fully inform their patients about the potential risks to which they may subsequently attach significance, regardless of likelihood. Such practices have medicolegal implications, given the precedent set by Rogers v. Whittaker in Australia, and the need to inform patients about all risks to which they may attach significance. It is acknowledged, however, that in cases where medical imaging use would be justified, a detailed discussion of possible radiation risks should clearly be weighed against a thorough explanation of the likely benefits. Indeed, appropriate use of CT and NM technology in-line with evidence-based research may render risks from individual scans negligible, particularly if potentially life-threatening conditions are suspected. Nevertheless, the small but increasing frequency of requests for diagnostic radiation scans when not clinically indicated has potential medicolegal and patient care implications for all referrers. In an ironic twist, legal firms in the United States that were previously involved in prosecuting cases of ‘under-investigation’ – which drove the evolution of ‘defensive medicine’ – are now seeking claimants who feel they may have been ‘over-investigated’ and ‘harmed’ by ‘unnecessary’ or over-exposed CT radiation. Although a causal link between cancer and a distant history of diagnostic radiation exposure would be difficult to prove, this survey suggests that a majority of Australian doctors may be exposing themselves to potential claims of negligence and medicolegal uncertainty because they cannot accurately quantify radiation doses and associated risks arising from the diagnostic scans they request.
The value of CT and NM technologies is undisputed; however, the enormous growth in their use has not been paralleled by adequate education about the associated doses and risks of diagnostic ionising radiation. The results of this survey confirm that knowledge of diagnostic radiation and its associated cancer-causing risks is inadequate across the medical profession, particularly among more experienced doctors. In certain clinical settings, doctors with inadequate awareness of these issues may be unable to perform required risk–benefit analyses, and therefore be incapable of fully informing their patients about these issues. Injudicious use of diagnostic radiation is a small but concerning feature of modern medicine, and such practices may result in unnecessary exposures, avoidable stochastic effects, medicolegal uncertainty and, in some cases, an abandonment of evidence based medicine. Improved education about radiation doses and potential risks from CT and NM imaging is necessary across the medical profession to ensure optimal use of these important diagnostic tools and the preservation of best medical practices.
Dr Richard Slaughter, Director of Medical Imaging, The Prince Charles Hospital.
- 1Australian Government. Professional Services Review. Report to the Professions. Australian Government, 2008/09. Page 2.
- 2US Government Accountability Office. Medicare Part B Imaging Services: Rapid Spending Growth and Shift to Physician Offices Indicate Need for CMS to Consider Additional Management Practices. US Government Accountability Office. 2008 June 21–22. Publication GAO-08-452. [Cited 2 Jan 2012.] Available from URL: http://www.gao.gov/products/gao-08-452.
- 6Editorial: cancer risks and radiation exposure from computed tomographic scans: how can we be sure that the benefits outweigh the risks? Arch Intern Med 2009; 169: 2049–2050..
- 7Insideradiology.com.au. RANZCR. Radiation risk of medical imaging for adults and children. 2009 May. [Cited 2 Jan 2012.] Available from URL: http://www.insideradiology.com.au/pages/view.php?T_id=57.
- 11Projected cancer risks from computed tomographic scans performed in the United States in 2007. Arch Intern Med 2009; 169: 2071–2077., , et al.
- 13Australian Broadcasting Corporation. CT scan scare putting patients at risk. ABC News online. 2010 April 12. [Updated 2010, April 14. Cited 2 Jan 2012.] Available from URL: http://www.abc.net.au/news/2010-04-12/ct-scan-scare-putting-patients-at-risk/393014.
- 14CT Scan use ‘risk’ to children. The Australian. 2010 March 17. [Cited 1 Jan 2012.] Available from URL: http://www.theaustralian.com.au/news/nation/ct-scan-use-risk-to-children/story-e6frg6nf-1225841575738..
- 15Alarm about computed tomography is unjustified. Letters to the editor, Med J Aust 2010; 193: 246., .
- 17Doctors' knowledge of patient radiation exposure from diagnostic imaging requested in the emergency department. Med J Aust 2010; 193: 450–453., .
- 21ARPANSA. Monitoring radiation doses from diagnostic radiology. 2012. [Cited 2 Jan 2012.] Available from URL: http://www.arpansa.gov.au/science/medical/index.cfm.
- 25Rogers v. Whitaker  HCA 58. (1992) 175 CLR 479 F.C. 92/045.