Recommendations

The American Association of Physicists in Medicine (AAPM), which represents more than 6,000 medical physicists throughout the United States and other countries, is a Member of the American Institute of Physics. The AAPM promotes the application of physics to medicine and biology and encourages interest and training in medical physics and related fields. AAPM appreciates the opportunity to offer its views on the draft recommendations of the International Commission on Radiological Protection (ICRP), and we commend the openness and responsiveness of the ICRP in making drafts available for public comment. The most recent draft, posted on the Internet for public comment on June 7, 2006, represents a considerable evolution from the earlier draft on which the AAPM commented in 2004. AAPM looks forward to interacting with the ICRP as these recommendations continue to evolve. General Comments: 1. In general, the 2006 draft is an improvement over the earlier version of the document, and addresses many of the suggested additions and clarifications. AAPM agrees with the suggestion made at the recent meeting in Rockville, MD that it is essential that the 2006 recommendations contain a section (or perhaps an Annex) that summarizes the changes in recommendations relative to ICRP Publication 60. This summary should consist of both a concise narrative description of the changes, and a table that lists them. 2. AAPM remains skeptical of the need to publish revised ICRP recommendations. In the document and during public presentations, ICRP stated that these recommendations are intended to “consolidate, simplify, and elaborate on the previous set of recommendations published in 1991 as ICRP Publication 60.” However, the current draft does not appear to make a convincing argument that there is a need to make changes in the recommendations since there has not been any significant change in radiation risks, there appears to be no compelling public health and safety argument to make any changes to the recommendations, or to national regulations that implement those recommendations. 3. AAPM recommends that ICRP include a statement in the front of the document that the assumption that the risk of detriment is proportional to dose at low doses is used only to make judgments related to the control of radiation exposures. In addition, ICRP should also include a statement that the risk values quoted and used for radiation protection purposes are not appropriate for determining the risk to individuals or specific populations for specific exposure situations. 4. Concept of Dose Constraint. The ICRP’s attempt to clarify the meaning and use of dose constraint is an improvement over the previous draft, but further clarification is needed. The use of the phrase “provides a fundamental level of protection” clouds the relationship of constraints and dose limits. ICRP should further clarify how constraints function within a radiation protection program and the optimization of protection for a source to ensure that adequate protection for an individual is achieved. The rewrite of the NCRP 49 Report encountered the problem of whether to use 1 mSv or 0.25 mSv for public protection limits in shielding design of diagnostic radiology facilities. After months of struggle, the NCRP published Statement No. 10, which clarified: “After a review of the application of the guidance in NCRP (1993) to medical radiation facilities, NCRP has concluded that a suitable source control for shielding individuals in uncontrolled areas in or near medical radiation facilities is an effective dose of 1 mSv in any year. " Yet, many countries (the UK for instance) have adopted 0.3 mSv as a shielding constraint. If the new ICRP keeps this recommendation, many more countries will adopt this value. Since the money available for health care is limited, radiological equipment maintenance and/or replacement as well as staff training are sacrificed in order to comply with regulatory requirements for shielding. The net result is a significant detriment to patient management, especially in developing countries. The problem may lie in the definition of single source. How can “the xray equipment in a hospital” be a single source? What are we going to do for shielding calculations? Take the “geometrical center of all the xray units as an “effective point source” or the edge of the closest one to the point of measurement? Attachment 1 (below) contains a paper titled, “Radiation Protection Standards: Their Evolution From Science To Philosophy” by R. L. Dixon, Joel E. Gray, B. R. Archer and D. J. Simpkin; Radiation Protection Dosimetry (2005), Vol. 115, No. 1–4, pp. 16–22 which elaborates on this point. 5. The ICRP proposes changing the radiation weighting factors, tissue weighting factors, and nominal risk coefficients for cancer and hereditary disease. Of all of the material in the draft recommendations, these changes have the greatest potential for a major impact on regulations promulgated by national authorities. Yet, some of these changes may be premature. The cancer incidence data used by the Biology Working Group is largely based on data published in the early 1990s (Thompson et al., 1994; Preston et al., 1994) using Japanese Abomb data and the DS86 dosimetry. A new dosimetry system has since been developed, but the “new analyses of the latest Abomb cancer incidence data are expected soon (Preston et al., in preparation)” [see Annex A, lines 1647–1648]. Promulgating this series of tissue weighting factors and nominal risk coefficients may result in a system of radiological protection that is overly conservative. AAPM believes that recommendations of the ICRP should be based upon published, peerreviewed scientific information that reflects the current state of knowledge. Thus, AAPM recommends that the ICRP not adopt a new set of tissue weighting factors and nominal risk coefficients until the assessment of the Abomb data is completed and published in a peerreviewed journal for public scrutiny. 6. AAPM is concerned about the inclusion of the discussions regarding the termination of pregnancy and believe that they are beyond the scope of the ICRP’s mission. Such discussions should be held on casebycase bases between competent medical practitioners and their patients, and it is therefore inappropriate for the ICRP to propose any numerical value that could be interpreted as the basis for terminating a pregnancy. AAPM recommends that this discussion be deleted from the ICRP recommendations. 7. AAPM is concerned that the ICRP has not clearly explained its rationale for the decision to not recommend genderspecific data for the purposes of radiological protection and how it accounts for gender differences in radiation sensitivity. This difference in radiation sensitivity observed in females has been described in publications of the U.S. National Academy of Sciences in 1990 (BEIR V) and 2005 (BEIR VII) and by the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR 2000). AAPM is concerned that without such a statement or basis, regulatory agencies, such as the U.S. Nuclear Regulatory Commission may choose to or be directed to adopt genderspecific data in subsequent revisions to national regulations for radiation protection. 8. In the introduction to the draft recommendations, the ICRP states that Section 10, “Protection of the Environment,” describes a policy approach for radiological protection of nonhuman species. Section 10 does not, in fact, state a policy but only provides a brief description of ongoing work of the ICRP. AAPM recommends that this section be deleted until stakeholders have the opportunity to provide input and comment as the ICRP develops a policy and the framework for assessment. 9. AAPM recommends that a thorough editorial review be performed before publishing the final document. There are numerous instances of incorrect spelling, incorrect usage of terms, references to publications that are not included in the reference lists, text and table numbers that do not agree, and references to documents that are yet to be drafted, being drafted, under review, or in press. In addition, there were many comments made during the public meeting in Rockville, MD that indicate the intent of the ICRP may not be exactly as stated by the printed draft text. All referenced material should be publicly available at the time of publication or reference to that material should be deleted. Specific Comments: 1. Paragraph 89. The term, “equivalent dose” was going to be replaced by “radiation weighted dose”, which solved two problems: the confusion in English speakers remembering when to use equivalent dose and when effective dose, and the problem for Spanish speakers, for whom both equivalent dose and effective dose translated to the same term. It would be good to make the change in terminology. 2. Chapter 7 – Much of the chapter on natural sources seems unnecessary. A simple distinction making manmade accumulation of radioactive materials no longer natural would place it under the previous chapters. 3. Relating to the radon issue, the report seems to misrepresent the conclusions of Lubin et al. 2004, and ignore the comments of Heidi et al, 2006 (Heid IM, Kuchenhoff H, Rosario AS, Kreienbrock L, Wichmann HE. Impact of measurement error in exposures in German radon studies. J Toxicol Environ Health A. 2006;69:70121) on the sensitivity of the studies to the poorly defined variables. Also missing was any reference to the significant work by Bernard Cohen. 4. Paragraph 317. The Commission asserts that the simplest way of dealing with potential exposure is through probability of radiationrelated death rather than effective dose. That seems inconsistent with considering potential exposure on the par with actual exposure. If potential exposure is seen in the context of death, than a simple prevention approach is the only that makes sense. Only if the potential exposure can be seen in a context of effective dose does its inclusion in this framework form coherence. 5. Paragraph 361. Most of the discussion in 11.1 assumes a Western, European style culture, particularly regarding public input into regulations. The tone might be modified to be less alienating to societies where the policy making is less open. Minor Comments: 1. Paragraph 37, delete single parenthesis in after “situations”. 2. Paragraph 49, line18, rewrite the paragraph to remove the conglomeration of four prepositions together. Attachment: Dixon et al paper, Radiation Protection Dosimetry (2005), Vol. 115, No. 1–4, pp. 16–22 RADIATION PROTECTION STANDARDS: THEIR EVOLUTION FROM SCIENCE TO PHILOSOPHY R. L. Dixon1,
, Joel E. Gray2, B. R. Archer3 and D. J. Simpkin4 1Department of Radiology, Wake Forest University School of Medicine, Winston-Salem, NC 27157-1088, USA 2Landauer, Inc., Glenwood, IL 60425, USA 3Baylor College of Medicine, Houston, TX 77030, USA 4St Luke’s Medical Center, Milwaukee, WI 53201-2901, USA The concept of applying constraints on individual sources to a small fraction of the public dose limit has been deemed inappropriate when shielding the medical X-ray sources. This represents a broad-based consensus of medical physics and radiological societies in the United States, and the report series on the shielding design for medical X-ray sources (including dental, X-ray imaging and therapeutic X ray) from the National Council on Radiation Protection and Measurements (NCRP) utilises 1 mSv y1 as a source control limit. In the present study, the rationale for such a conclusion is discussed, and a somewhat critical look at the current model of radiation protection of the public is made. INTRODUCTION In this study, we take a critical look at the current state of radiation protection guidelines for the protection of the public, and the consequences if our present course ensuing is not altered. For the shielding design of medical X-ray sources, the concept of applying constraints on the individual sources to a small fraction of the public dose limit has been deemed inappropriate, as described below, and the recommended limit for shielding design is 1 mSv y1. The recently published report series on the shielding design for medical X-ray sources from the National Council on Radiation Protection and Measurements (NCRP) contains the following wordings: Based on ICRP (1991) and NCRP (1993) recommendations for the annual limit of effective dose to a member of the general public, shielding designs shall limit exposure of all individuals in uncontrolled areas to an effective dose that does not exceed 1 mSv y1. After a review of the application of the guidance in NCRP (1993) to medical radiation facilities, NCRP has concluded that a suitable source control for shielding individuals in uncontrolled areas in or near medical radiation facilities is an effective dose of 1 mSv in any year. Additionally, in December 2004, NCRP published Statement No. 10(1) ‘Recent applications of the NCRP public dose limit recommendation for ionizing radiation’ that reinforces the intent of the preceding paragraph. Statement No. 10 refers to the NCRP Report 145 on dental facilities(2), Report 147 on medical X-ray imaging facilities(3), Report 148 on veterinary facilities(4), and on-going work in the Scientific Committee 46-13 on megavoltage radiation therapy facilities; and supports the fact that these facilities ‘shielded in accordance with the recommendations contained in NCRP Reports 145, 147 and 148 (i.e. designed not to exceed 1 mSv y1 to the maximally exposed individual in an uncontrolled area), will provide adequate protection to the employees and the members of the public that access the uncontrolled areas. The association of a risk with our current public dose limit of 1 mSv y1 (a dose less than natural background) using LNT extrapolation, as if it were a confirmed scientific fact, is clearly not justified as can be seen from the statement by the NCRP(5) in Report 136 ‘ . . . the probability of effects at very low doses such as are received from natural background is so small that it may never be possible to prove or disprove the validity of the LNT assumption’. Acceptance of LNT is more a matter of conviction than of science; however, it is not our intention to enter into that rather sterile debate, since a public dose limit of 1 mSv y1 has been adopted into regulation in the United States. It is simply a ‘rule’—no more, no less. The concept of source constraints states that any source (or group of sources) under one administrative control should be limited to a small fraction of the public dose limit, using the rationale that a given member of the public may be exposed to several such sources. This sounds eminently plausible at first reading, and this belief has been widely promoted by the ICRP, which has typically suggested a fraction of 1/3 (and is also supported by certain wordings in
Corresponding author: rdixon@wfubmc.edu the NCRP Report 116(6) that suggest a factor of 1/4). Yet, this concept cannot withstand closer scrutiny, based on simple logic, as will be described below. Those schooled in scientific principles would never seriously consider setting a limit on a quantity, which is lower than its ambient noise (natural fluctuation), yet the proposed value of 0.25 (or 0.3) mSv y1 is well below the fluctuation of natural (‘immitigable’) terrestrial and cosmic ray background radiation from one locale to another (excluding the even larger variation due to radon), i.e. the signal is ‘down in the noise’. This leads to strange inconsistencies as will be described here. Fortunately, this suggested ‘quartering’ of the public dose limit(6) has until now been ignored in the United States while setting public dose limits for medical radiation sources, and its application would have significant negative impact on healthcare in the United States; not only owing to the increased costs involved, but also since it directly affects the patient, as we will point out. The authors, as members of the NCRP committee, were charged with the development of a new report on structural shielding design for medical imaging facilities(3) (NCRP Report 147), and were well aware of the significant and unpleasant consequences of reducing the shielding design limit from 1 to 1/4 mSv y1, as well as the obvious fact that it could not possibly have any significant positive impact on protection of the public; however, there was an initial pressure to adopt 1/4 mSv y1. There was no doubt that if we had adopted and utilised a specific value of 0.25 mSv y1 in our report, which (unlike NCRP Report 116) was a utilitarian working document that could not be ignored and would be utilised for the medical shielding design in the United States for decades, then it would become a de facto dose limit. We chose to logically argue the case for design to 1 mSv y1; and, to its credit, the NCRP listened and approved our report, using the value of 1 mSv y1 (or 1 mGy y1 air kerma as a practical design target). The report utilises a new approach that has already been described in the literature(7–10), and some of the methodology has been adapted for a report on diagnostic X-ray shielding design recently published by the British Institute of Radiology(11). PUTTING THINGS IN PERSPECTIVE The best way to set the stage, is, perhaps, to look at some data extracted from a study by Tengs et al. (12) on the costs of various life-saving interventions listed in Table 1. Table 1. Cost per life-year saved(12). Radiation emission standards for nuclear power plants: $100 million per life-year Radionuclide emission control at NRC-licensed facilities: $2.6 x 10 9 per life-year Widen lanes on rural roads by 2 ft: $120,000 per life-year What sort of thinking could result in this inequity and egregious waste of resources? Those of us working in radiation protection often think in terms of a ‘universe’ in which radiation represents the only risk, whereas in reality our ’slice of the risk pie’ represents a very tiny percentage (<1%) of the total probability of death from all causes. This myopic world-view is a natural tendency in many fields. The coupling of this imaginary universe in which ‘lower is always better’ with the real world is often not considered adequately. Lower is not always better, since lowering a dose limit can have negative consequences, and can produce unforeseen new risks that far outweigh the reduction in radiation risk. The annual public dose limit that the shielding designers must utilise has been ‘ratcheted down’ from 5 to 1 mSv, and now what appears to be an attempt under the guise of ‘source constraints’ to create an even lower de facto public dose limit of 0.25–0.3 mSv y1—there can be no doubt about this. Although some still try to cover these machinations with the veil of ‘science’, it is clearly out of that realm. THE MYTH OF EXPOSURE TO MULTIPLE SOURCES Assume a ‘universe’ without source constraints, such that each man-made source is shielded to 1 mSv y1. The stated rationale for source constraints that a member of the public might be exposed to three or four such man-made sources to an annual cumulative dose of any significance compared with 1 mSv y1 can be readily refuted by the application of simple logic based primarily on the fact that a given person cannot be in more than one place at a time, and the fact that the various ‘other sources’ to which they might be exposed do not typically converge to a given location. The sources being shielded are highly localised in space and time (not a factor if you are >20 ft away) and most are inactive (switched off) after normal working hours. The chances of working or otherwise spending any significant time within this radius of an X-ray machine is less than one in a million for a member of the public. Only those members of the general public who regularly occupy areas immediately adjacent to the X-ray rooms or sources have any significant potential for exposure. This is a very small subset of the general population—probably not significantly larger than the number of occupationally exposed(13) medical radiation workers due to the highly localised radiation field. In addition, it is highly unlikely that any member of the general public who works in, or regularly occupies, uncontrolled areas immediately adjacent to medical X-ray rooms will encounter any other sources of man-made radiation on a regular basis. Furthermore, it is doubtful that random source encounters would provide any additional annual dose of any significance relative to the annual dose limit of 1 mSv. It would be an extremely rare case in which a person who occupies an area next to an X-ray room during the normal working day goes to, and occupies on a regular basis, another area immediately adjacent to yet another X-ray room (or any other man-made source of radiation) that produces any significant radiation after normal working hours. Living spaces, such as houses or apartments, rarely share adjacency with the X-ray facilities, and even in those rare cases where they do, the sources are typically ‘switched off ’ at night. Even if the person has a night job in another medical facility, the probability that their ‘night job’ office or station is beside yet another X-ray room or source is small (Why should it be? They are not radiation workers). Even in the unlikely event that it were, the workload would be much reduced after regular working hours. In addition, the likelihood that a given member of the public, who is currently working in the environs of an X-ray source, will spend a significant fraction of his ‘working life’ beside that X-ray unit (or any other source) is small, since either he/she or the source are likely to move during that time. The median tenure in a job in the United States is 3.7 y(14). Consider the following case study. An accountant, Jane Doe, works full-time in an office immediately adjacent to an X-ray room, which has been shielded to a design level of 1 mSv y1 in that office, and assume that she actually receives 1 mSv y1 from that source (which is also highly unlikely considering the conservatism usually inherent in shielding design). Where then can she go after work, on a regular enough basis, to acquire any additional dose that is significant compared to 1 mSv y1? As discussed above, the probability is very small that her living quarters are beside yet another source, much less one that is operated after normal working hours; or, if she has a second job, that her second job as a non-radiation worker also places her beside a source operating after normal hours. Even considering her exposure at her current job, how long will she work in the same office, and how long will the X-ray room be operating beside it? If we assume a period of 5 y, and reasonable probabilities for her future work and present and future living quarters being beside another source, then her average exposure over her 35 y working lifetime is <0.15 mSv y1. What other sources might she encounter that we have not considered? For this information, we consulted the NCRP Report 95 ‘Radiation exposure of the US population from consumer products and miscellaneous sources’(15). Of the sources listed, Table 2 lists the largest contributors. Table 2. Radiation exposure of the US population from consumer products and miscellaneous sources per annum (from NCRP Report 95). Source Exposed population (mSv) Average to US population (mSv) Building materials 0.07 0.035 Cooking with gas 0.004 0.002 Highway construction 0.04 0.0008 materials Domestic water supply 0.01–0.06 0.01–0.06 Electronic products <0.01 <0.01 Tungsten welding rods 0.16 0.0002 None of these will contribute any significant dose to Jane Doe (unless she perhaps works in the evenings as a welder). Additional random encounters with sources, such as nuclear medicine patients, will not be regular enough to contribute any significant annual dose. Even if you live at the fence of a nuclear power facility, the dose is <0.01 mSv y1. Even the ICRP in ICRP Publication 60(16) states that it does not believe that rationale, as quoted: (188) Concerning the possibility of cumulative public exposures from multiple ‘other sources’: the ICRP ‘does not believe that this occurs to a significant extent’. Clearly, a cost-effective, national policy for radiation shielding design should not be based on the most unlikely occurrences affecting perhaps one or two people out of the entire population, but rather on reasonable expectations; and certainly not on an arbitrary factor of 1/4! A MYOPIC WORLD-VIEW Those of us working in radiation protection often think in terms of a ‘universe’ where radiation is the only risk, so it is useful to remind ourselves exactly how big a slice of the public ‘risk pie’ we actually own. At the public dose limit of 1 mSv y1, and assuming the current risk coefficient of 5 105 mSv1 (based on LNT), our slice of the pie represents only 0.3% of the total lifetime probability of death, compared to its other causes. Table 2 also provides examples of this limited world-view, where attempts are made to capture into our sphere of influence, not only man-made sources, but ‘man-enhanced’ natural sources. For example, Table 2 implies that the domestic water supply presents a radiation risk; however, this risk is insignificant compared to the risk associated with having no domestic water supply. Similarly, the radiation risk from highway construction materials is totally negligible compared to the risks of driving on the highway. In addition, we should remind ourselves that natural sources of radiation contribute much more radiation to the public than stray radiation from man-made sources: an average of 3 mSv y1 as presented in Table 3. (If one takes into account the recent doubling of the radon riskestimate by the US Environmental Protection Agency (EPA), the radiation from our sources seems even more trivial by comparison). Table 3. Average natural background radiation levels in the United States. Natural radiation source Average annual dose (mSv) Cosmic radiation 0.3 Terrestrial 0.3 Internal (40K, etc.) 0.4 Radon 2.0 Total average background 3.0 Although radon exposure can be reduced (mitigated) to a certain extent, the others cannot be easily reduced . The practice of setting dose limits based on radiation risk calculations alone, without due regard to their cost-effectiveness, cost vs. benefit, or their other (often unintended) negative effects on society, is fundamentally flawed. Since such broader analyses are admittedly more difficult, the simplified approach being advocated for our future seems to be: better just to divide by 3 or 4, and not have to think about it—just make a rule. One size fits all. Simplified public dose recommendations Rule 1: 1 mSv / y. Rule 2: Divide Rule 1 by 4. Can this be science? The 0.25 or 0.3 mSv ‘constraint’ now becomes a de facto dose limit. If all shielding designers in the United States have to use the same design value for medical X-ray shielding (and one cannot imagine otherwise), then it is a ‘limit’ and we might as well call it a ‘shielding design limit’. In addition, there appears to have been a subtle, and disturbing, change in interpretation which is being applied with respect to that rather arbitrarily set public dose limit of 1 mSv y1. Both ICRP Publication 60 and NCRP Report 116 contain words to the effect that it is not a matter of concern if a member of the public receives up to 5 mSv in some years, if their average remains <1 mSv. The interpretation now seems to be that we must avoid, at all costs, even the slightest possibility that any given member of public would receive >1 mSv in any given year, by dividing the public limit by 3 or 4 as discussed. The implication being made is that exceeding the public dose limit represents an ‘unacceptable’ risk (or danger, by insinuation). DOWN IN THE NOISE Table 4 lists the broad fluctuation of natural (and ‘immitigable’) terrestrial and cosmic radiation with geographic location in the United States (exclusive of radon that exhibits a much larger variation). This leads to the curious result that the owner of a medical linear accelerator in Coastal America will have to spend an extra $250,000 USD to ‘upgrade’ the shielding of his/her machine to meet the reduced dose limit of 1/4 mSv y1 in order to protect a few members of the general public who work in its immediate environs; when every pizza parlor, ice cream shop, or any other business or home in the Denver region will have a higher radiation level than that beside his/her accelerator, affecting every one of more than two million people who live in that region (even excluding radon, which is also much higher in the Denver region). Not only that, but even those few members of the public the owner exposes will receive 10 times more radiation from natural background than from his accelerator (and even slightly more from the 40K inside their own bodies). Perhaps someone would care to explain the logic behind this to the accelerator owner, or to his patients who are also affected? Table 4. Variation of natural terrestrial and cosmic background radiation in the United States—exclusive of radon and internal (NCRP Reports 45 and 94(17)). Atlantic and Gulf coastal plain 1.05 mSv / y Middle America 1.25 mSv / y Rocky Mountain Plateau 1.45 mSv / y Denver Colorado 1.65 mSv /y Population-weighted average 1.09 mSv / y Population-weighted average + 0.25 mSv 1.34 mSv / y Perhaps it would be more fair to correct this inequity, and to shield everyone to the same ‘risk’, i.e. all sources to an annual dose limit equal to the population average background ţ0.25 mSv, irrespective of location. In this case, referring to Table 4, those in the Denver region would be faced with a negative constraint, thus obviating the possibility of obtaining medical or dental X-ray units. CONSEQUENCES OF REDUCING THE SHIELDING DESIGN LIMIT IN THE UNITED STATES TO ONE-FOURTH OF THE PUBLIC DOSE LIMIT Are there any significant increased costs, or other negative effects that would result from adopting this lower dose limit? This recommendation is not just discretionary application of the as low as reasonably achievable (ALARA) principle, but rather would have mandated a design dose limit of 0.25 mSv y1. Popular misconceptions Those unfamiliar with the details of medical shielding design often present the following arguments. Misconception 1 These new recommendations would only apply to new facilities and practices, and the existing X-ray rooms or linear accelerator vaults would not be required to meet this lower dose requirement. The fallacy in this is that when the equipment in a room is replaced, the regulators will require a new shielding evaluation (‘plan review’), and they will certainly require you to use the new (lower) dose limit. The use of two different design limits for new and old X-ray rooms would be an anathema to the regulators, not to mention a logical anomaly. This creates a significant problem, as will unfold. More than one-half of the diagnostic X-ray shielding projects that we do are for equipment upgrades (e.g. CT scanners with ever-increasing capability), and most of the radiation therapy equipment sold in the United States is for replacement in existing vaults.Adding shielding to existing X-ray or radiation therapy rooms is a significant problem, requiring considerable expense and ‘downtime’. Adding the shielding inside the room, requires removal and re-installation of all the fixtures, such as cabinets and sinks, whereas outside of the room, one has bathrooms with fixtures, multiple walls at right angles to the X-ray room wall, and the other side of the wall might even belong to some other facility. In many cases, as described below, imposition of a 1/4 mSv annual limit will require lead in the floor and ceiling of X-ray rooms, and extending the wall lead up to the ceiling, in which case lighting fixtures and a myriad of ductwork, plumbing and electrical wiring must be removed and re-installed. What seemed to the administrator to be a simple equipment replacement has, to his surprise, not only turned into a major expense, but one requiring weeks of downtime. Suppose it were the only CT scanner, cardiac catheterisation room, vascular laboratory or linear accelerator in the hospital (or in the region). This represents a significant and real health risk to patients needing such facilities for diagnosis and treatment which far outweighs the hypothetical risk associated with a 3/4 mSv exposure reduction. Misconception 2 It does not cost much extra to shield an X-ray room to reduce the ambient dose by extra factor four— <1 mm of lead should do it. This can easily be discounted for radiation therapy facilities, where an extra 25 cm of concrete or an extra 3.8 cm of lead would be required. Even for diagnostic X-ray rooms (<140 kV), this is clearly not true for ‘upgrading’ existing X-ray rooms, as described above in Misconception 1; however, even in the case of new construction, it applies only to the wall shielding (and even then, not in all cases). In many cases, as discussed below, application of a 1/4 mSv annual limit, would require addition of lead to the floor and the ceiling of an X-ray room—a difficult (and hence expensive) construction job due to the ductwork, and the like, above the false ceiling as well as problems with putting lead in the floor. In addition, even if there is no occupancy above, if the wall lead is installed to its usual height of 2.1 m above the floor, scatter from the ceiling structure over the wall into the next room (analogous to ‘skyshine’) may be comparable to 1/4 mSv, and require extending the lead to the ceiling(18). Predictions if designing shielding to 0.25 mSv / y Radiation oncology facilities The penetrating radiation used in radiation treatment, typically from linear accelerators operating at 6–18 MV, requires significant additional shielding to achieve a factor of four reductions in ambient dose. The following is adapted from an analysis provided by James Rodgers from Georgetown University in collaboration with architect Paul Torp(19). Most new accelerator purchases are replacements in existing vaults, in which case adding lead to the inside of the walls is usually the only viable option. Table 5 lists the additional shielding and estimated costs that would be required to ‘upgrade’ existing vaults from the current design limit in the United States of 1 mSv y1 down to 0.25 mSv y1. Table 5. Estimated costs to ‘upgrade’ the US base of radiation therapy vaults of 6–18 MV linear accelerators in order to meet a lower dose limit of 1/4 mSv / y. Additional shielding thickness 2.5 cm lead Weight of additional shielding per vault 2.7 x 10 5 kg Material cost of additional lead per vault $115,000 USD Construction costs per vault: $150,000 USD demolition and reconstruction Total additional cost per vault (1 mSv to 0.25 mSv) $265,000 USD Total per vault if built when public $385,000 USD dose limit was 5 mSv / y Total estimated cost nationwide $1,000,000,000 USD (3000 vaults in the US; 1999 survey) Total lead required nationwide 8 x 10 8 kg For new construction, the additional cost is nominal. An additional concrete shielding thickness of 25 cm would be required (2  105 kg) for a total increase in cost of only about $15,000 USD. Medical diagnostic X-ray facilities The predicted consequences for medical X-ray of reducing the shielding design limit from 1 to 0.25 mSv per annum are listed below:  CT scanner rooms and cardiac cath labs having occupancy above and below will require lead in the floor and ceiling—a seamless, lead box around the X-ray unit.  Many CT scanners, cardiac cath labs and vascular specials rooms will require wall lead to extend to the ceiling slab—even if not occupied above, and a lead thickness in excess of 1.5 mm.  Dental X-ray units: these presently require (nor do they have) any shielding beyond that provided by gypsum wallboard, which is currently sufficient; however, those walls which bound fully occupied areas will require adding lead shielding.  Nuclear medicine camera rooms, almost universally unshielded at present, will require lead shielfding if located beside fully occupied areas. In some cases even corridor walls would require lead.  Positron emission tomography (PET) or PET/ CT scanners will require an extra 10 mm of lead; and also lead in the floor and ceiling, if occupied above or below. This is likely to exceed the structural limits of the building.  Low usage radiographic rooms (<10 patients per week) for which gypsum board walls were previously sufficient will require lead. Rural medical clinics will face an additional economic burden, as will veterinarians and podiatrists.  Bone densitometer (DEXA) units will require lead shielding in many cases if located beside a fully occupied area. These units, which are important to women’s health, have become widely available in physicians offices in part because they can be put in existing examination rooms without the addition of structural shielding. Estimated costs for diagnostic X ray  Dental (1 wall/machine): $520 million USD total nationwide.  CT upgrades: $220–400 million USD nationwide.  Cardiac cath lab upgrades: $20,000 USD per room.  Low use X-ray rooms: $9,000 USD per room.  Additional lead required: 2  108 kg nationwide. OTHER FACTORS AND CONSEQUENCES Effect on the patient population We should note that these increased costs are not simply borne by the owner or operator of the facility, but are eventually passed on to the patient in terms of increased costs of healthcare, affecting the entire patient population. If a hospital or clinic wishes to offer a new service to the community, such as PET/CT, or simply replace the existing equipment with the state-ofthe-art equipment, the increased costs or the impossibility of altering the existing structure may kill the project; such that the possible life-saving service is not offered in the community, or the facility labourson with outdated and inferior equipment resulting in sub-optimal patient care. In addition, just replacing equipment turns into a major construction project with weeks of downtime, which can likewise negatively affect patient care, making vital diagnostic tools such as CT completely unavailable in some cases. The adverse effects would be felt more strongly by healthcare in rural areas. Thus, these considerations represent real health risks for the patient population—the same general public for whom we were attempting to provide increased protection (albeit only hypothetical increased protection, based on no firm evidence). Other (unforeseen) risks as a result of ‘quartering’ the public dose limit The total additional lead required for medical facilities (diagnostic and therapeutic) is 109 kg, which is more than half of the current annual production of lead. Although it is not clear how many more metal miners and smelters would be required, the actual risk for these workers is 56  105 per annum(20), which is 10 times larger than the (hypothetical) risk of 1 mSv y1 of radiation. Similarly, the risk to the construction workers required to perform these room modifications is 14  105 per annum(20), which does not include the risk of lead poisoning in handling the lead during construction, nor during the demolition of abandoned X-ray rooms. Also, what are the effects on the environment from increased emission due to the smelting of this lead (of both lead and the radioactive isotopes it contains)? CONCLUSION It has been shown that designing shielding for medical X-ray sources to 1 mSv y1 is not likely to result in members of the public receiving more than the public dose limit. It has also been illustrated that attempting to increase public protection by forcing the doses allowed from medical X-ray sources to unrealistically low levels (to a small fraction of the public dose limit), that the overall effect is likely to represent net harm to the patient population (the same public) in terms of both increased healthcare costs as well as increased health risk to individual patients. Much of the pressure to do so is in no doubt based on public misconceptions about the dangers of radiation (in many cases incited by the strident fringe elements of society), and the net effect is that the public is unknowingly hurting themselves. Table 1 aptly illustrates this point where public safety is clearly more directly served by spending money on widening roads than in pushing radiation protection limits to unrealistically low levels. Similarly, the public interest would not be served if healthcare should suffer the same fate as nuclear power. Even more troubling is the fact that developing nations, who have great difficulty in finding funds to purchase modern diagnostic or therapeutic X-ray equipment, will adopt such recommendations (1/3 mSv y1) as shielding criteria, thereby wasting precious resources. REFERENCES 1. National Council on Radiation Protection and Measurements. Recent applications of the NCRP public dose limit recommendation for ionizing radiation. NCRP Report 10 (Bethesda, MD: NCRP) (2004). 2. National Council on Radiation Protection and Measurements. Radiation protection in dentistry. NCRP Report 145 (Bethesda, MD: NCRP) (2003). 3. National Council on Radiation Protection and Measurements. Structural shielding design for medical X-ray imaging facilities. NCRP Report 147 (Bethesda, MD: NCRP) (2004). 4. National Council on Radiation Protection and Measurements. Radiation protection in veterinary medicine. NCRP Report 148 (Bethesda, MD: NCRP) (2004). 5. National Council on Radiation Protection and Measurements. Evaluation of the linear-nonthreshold dose–response model for ionizing radiation. NCRP Report 136 (Bethesda, MD: NCRP) (2001). 6. National Council on Radiation Protection and Measurements. Limitation of exposure to ionizing radiation. NCRP Report 116 (Bethesda, MD: NCRP) (1993). 7. Dixon, R. L. On the primary barrier in diagnostic x-ray shielding. Med. Phys. 21, 1785–1793 (1994). 8. Dixon, R. L. and Simpkin, D. J. Primary shielding barriers for diagnostic x-ray facilities: a new model. Health Phys. 74, 181–189 (1998). 9. Simpkin, D. J. and Dixon, R. L. Secondary shielding barriers for diagnostic x-ray facilities: scatter and leakage revisited. Health Phys. 74, 350–365 (1998). 10. Simpkin, D. J., Archer, B. E. and Dixon, R. L. Radiation protection design and shielding in diagnostic installations. In: Biomedical Uses of Radiation, Vol. 1. W. H. Hendee, Ed. (Weinheim, Germany: Wiley- VCH) (1998). 11. Sutton, D. G. and Williams, J. R., Eds. Radiation shielding for diagnostic X-rays. Report of a joint BIR/ IPEM working party (London: The British Institute of Radiology) (2000). 12. Tengs, T. O., Adams,M. E., Pliskin, J. S., Safran, D. G., Siegel, J. E., Weinstein, M. C. and Graham, J. D. Five hundred life saving interventions and their cost effectiveness. Risk Anal. 15, 369–390 (1995). 13. National Council on Radiation Protection and Measurements. Exposure of the US population from occupational radiation. NCRP Report 101 (Bethesda, MD: NCRP) (1989). 14. US Bureau of Labor Statistics (2003). 15. National Council on Radiation Protection and Measurements. Radiation exposure of the US population from consumer products and miscellaneous sources. NCRP Report 95 (Bethesda, MD: NCRP) (1987). 16. International Commission on Radiological Protection. 1990 Recommendations of the ICRP. ICRP Publication 60. Ann. ICRP 21(1–3) (Oxford: Pergamon Press) (1991). 17. National Council on Radiation Protection and Measurements. Exposure of the population in the United States and Canada from natural background radiation. NCRP Report 94 (Bethesda, MD: NCRP) (1987). 18. McRobbie, D. W. Radiation shielding for spiral CT scanners. Br. J. Radiol. 70, 226 (1997). 19. Rogers, J. E. and Torp, P. In: Proceedings of the American Association of Physicists in Medicine President’s Symposium, San Diego, CA (2003), October 2003, private communication. 20. Wilson, R. and Crouch, E. A. C. Risk-benefit Analysis (Boston, MA: Harvard University Press) (2001).