3

Workshop Summary

For increased readability, the chapter is organized by theme rather than chronologically based on the workshop agenda (see Appendix B). An integrated summary of the presentations and discussions are reported in this chapter. This summary should not be construed as reflecting consensus or endorsement by the workshop committee members (see Appendix C for committee roster), the invited workshop presenters (see Appendix D) and other participants, or the National Academy of Sciences.

3.1 OPENING COMMENTS

The organizing committee invited two speakers to provide opening remarks to help establish the context for the workshop discussions: Charles Miller (chief, Radiation Studies Branch, Division of Environmental Hazards and Health Effects, National Center for Environmental Health, CDC) and Donald (Don) Miller (acting chief, Diagnostic Devices Branch, Division of Mammography Quality and Radiation Programs, Center for Devices and Radiological Health, FDA). Both the CDC and FDA have been active in the discussions of tracking radiation exposures from medical diagnostic procedures.

The CDC initiated studies in 2004 on the feasibility of a tracking system for medical diagnostic procedures involving ionizing radiation (CDC 2004a,b, 2006). The specific question explored was: “How could the procedure code in patient medical records be used to derive a radiation dose?”



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 17
3 Workshop Summary F or increased readability, the chapter is organized by theme rather than chronologically based on the workshop agenda (see Appendix B). An integrated summary of the presentations and discussions are reported in this chapter. This summary should not be construed as reflecting consen- sus or endorsement by the workshop committee members (see Appendix C for committee roster), the invited workshop presenters (see Appendix D) and other participants, or the National Academy of Sciences. 3.1 OPENING COMMENTS The organizing committee invited two speakers to provide opening remarks to help establish the context for the workshop discussions: Charles Miller (chief, Radiation Studies Branch, Division of Environmental Hazards and Health Effects, National Center for Environmental Health, CDC) and Donald (Don) Miller (acting chief, Diagnostic Devices Branch, Division of Mammography Quality and Radiation Programs, Center for Devices and Radiological Health, FDA). Both the CDC and FDA have been active in the discussions of tracking radiation exposures from medical diagnostic procedures. The CDC initiated studies in 2004 on the feasibility of a tracking sys- tem for medical diagnostic procedures involving ionizing radiation (CDC 2004a,b, 2006). The specific question explored was: “How could the pro- cedure code in patient medical records be used to derive a radiation dose?” 17

OCR for page 17
18 TRACKING RADIATION EXPOSURE This effort culminated in a 2006 CDC workshop that concluded that it would be extremely difficult to monitor actual doses received by patients. The FDA recently published the Initiative to Reduce Unnecessary Radi- ation Exposure from Medical Imaging (http://www.fda.gov/downloads/ Radiation-EmittingProducts/RadiationSafety/RadiationDoseReduction/ UCM200087.pdf), aiming to promote safe use of medical imaging devices, support informed clinical decision making, and increase patient awareness. Charles Miller proposed that the workshop participants consider the following question: Is now the appropriate time to reconsider the impact of radiation doses from medical procedures? Specifically: 1. Can we measure and record real doses that patients receive? 2. Can we track individual doses, and should we? 3. How can we potentially use such data to inform decisions by patients and health care providers without interfering in the use of potentially life-saving medical procedures? He emphasized that information about patient doses from medical diagnostic procedures today is based on estimates and not actual measure- ments. He provided an overview of the efforts that have been initiated during the past five years to raise awareness about radiation exposure in the United States, which include the Image Gently and Image Wisely campaigns. Without endorsing them, he mentioned the many web-based applications that encourage patients to keep records of their imaging exams and share the information with their doctor. Patients can easily enter the type of imaging they received, their age when they had the procedure and, assuming some standard effective dose for a procedure (e.g., 8 mSv for an abdominal CT exam) the applications calculate the induced risk. Don Miller stated that any discussion on “whether,” “what,” and “how” to track exposure regarding CT, fluoroscopy, radiography, and nuclear medicine should be initiated with a clear understanding of the tracking’s purpose. Table 3.1 (adapted from Don Miller’s presentation) summarizes the information that in his view needs or does not need to be tracked to achieve the goal(s) of a tracking system. The four goals discussed were: • Justification • Optimization • Individual risk assessment • Research purposes As an example, Don Miller explained that if the purpose of tracking is to help the physician, dentist, or other health care provider to decide

OCR for page 17
19 WORKSHOP SUMMARY TABLE 3.1 Reason to Track Radiation Exposure from Medical Diagnostic Procedures and Information Needed to Achieve the Goal Facility Identifiersa Dose Datab Purpose PHI Yes No Justification Yes Optimization No Yes Yes Risk Assessment Yes No Yes Research Yes Yes Yes PHI=Personal Health Information aFacility identifiers may include name, location, and type of facility (e.g., hospital or indepen- dent imaging center). bDose data may include dose indices and dose estimates. In this content, the term does not refer to information on frequency and type of imaging exam or body part irradiated, which would be needed for all tracking purposes. SOURCE: Presentation by Don Miller (acting chief, Diagnostic Devices Branch, Division of Mammography Quality and Radiation Programs, Center for Devices and Radiological Health, FDA). whether an imaging exam is necessary (justification), the registry should contain information that can answer the following questions: Are there previous exams that could answer the clinical question? What were the find- ings? Where are the images? Having this information in a registry (which would likely be an electronic medical record rather than a “dose registry”) could avoid repetition of an exam that has already taken place. In such a case, the registry/record must contain personal health information (PHI, to identify the individual patient) and facility identifiers (to be able to retrieve the results and images of the past imaging exam) but not dose data, in order to serve the purpose of justification as described. In a later presentation (see Section 3.6.1), justification was also discussed in terms of “known clinical benefit” of an exam; in that case, tracking dose data through clinical trials that would provide the answers regarding the clinical benefits of the exams ordered may be necessary. In contrast, to optimize radiation delivery from medical imaging and establish reference levels, a registry would need to contain facility informa- tion and dose information for examinations from a number of patients but would not need patient-specific information (i.e., PHI). A registry that fits this purpose is the ACR Dose Index Registry discussed in Sections 2.5.3 and 3.2.4. Information needed to achieve the goals of risk assessment and research was also described.

OCR for page 17
20 TRACKING RADIATION EXPOSURE 3.2 POPULATION UTILIZATION OF IMAGING Several workshop participants affirmed that comprehensive and detailed data concerning diagnostic imaging utilization and associated radiation doses would help to evaluate whether concern over the dramatic increase in the population’s exposure to radiation is warranted. Dr. Mythreyi Chatfield (director of data registries, American College of Radiology) separated the issue of the measurement of population utilization of medical diagnostic procedures into two challenges: a) counting the number of imaging pro- cedures performed on the population and b) grouping these procedures into meaningful categories that represent a single imaging procedure with comparable radiation dose levels across patients and facilities. A number of data sources that cover patient populations in the United States are available, and several were discussed during the workshop. Dr. Chatfield categorized the sources of existing information on population utilization of imaging as surveys of patients or providers, administrative claims, and registries. However, the information from these sources exists only fragmentally and not in the detail required for assessment of the asso- ciated risks and benefits. 3.2.1 FDA Surveys David Spelic, physicist with the FDA, provided an overview of the FDA’s past and present efforts to characterize U.S. population doses from diagnostic x-ray imaging. The predominant means by which FDA has gath- ered such data is by nationwide surveys. Covering a period of roughly five decades, these surveys document the state of practice for a broad scope of diagnostic x-ray procedures, capturing indicators of patient dose, image quality, and an array of related technical parameters that characterize sur- veyed exams. The U.S. Public Health Service (USPHS) conducted the first national, large-scale surveys, the X-ray Exposure Studies (XES), in 1964 and 1970. These surveys captured comprehensive data regarding the state of practice in diagnostic radiography. Dr. Spelic said that each survey consisted of two components: a household interview of selected members of the U.S. popula- tion and the capture of technical information from clinical sites regarding x-ray equipment and radiologic practices for selected exams. Data regard- ing x-ray exam history were collected for 31,289 persons representing 9,653 households in 1964 and 67,000 persons or 22,500 households in 1970. Major outcomes from these surveys included publications providing comprehensive statistical summaries of findings as well as detailed dosim- etry for the exams covered by the surveys (USPHS, 1966, 1969, 1973). Their scope was large and included dental, medical x-ray, fluoroscopy,

OCR for page 17
21 WORKSHOP SUMMARY and x-ray therapy. Film packs were sent to clinical sites to capture beam size and dosimetry (USPHS, 1973); separate film packs were used for each modality. Because dosimetry was an important endpoint for these surveys, the Bureau of Radiological Health developed models to compute patient exposure based on reported x-ray technique, collimation, and film packet measurement. Doses were computed using phantoms; exposure ratios and scatter were measured for dose calculations. Dr. Spelic then discussed the Breast Exposure Nationwide Trends (BENT) project that begun in the late 1970s. It was a joint effort by the FDA and the National Cancer Institute (NCI) to study the current practice of mammography with the aid of state radiological programs. Among the survey findings was a broad variability of patient exposures ranging from 2.2 mGy to 140.0 mGy. Direct exposure film provided the highest exposures, while screen film the lowest. The Dental Exposure Normaliza- tion Technique (DENT) program followed a similar pattern to the BENT program. The Radiation Experience Data (RED) study was conducted in 1980 by the FDA’s Center for Devices and Radiological Health (CDRH) to estimate numbers and types of diagnostic imaging procedures performed in hospitals in the United States; no dosimetry data were collected. Data were collected on all types of imaging procedures including CT, ultrasound, and nuclear medicine from 81 sites, which is a small population compared to the XES surveys. Among the findings was that 130.2 million x-ray procedures were performed annually in short-stay hospitals, a 59 percent increase from the number of procedures performed in 1970 (81.7 million). There were 2.2 mil- lion CT exams performed, and 73 percent of these exams were of the head.1 Dr. Spelic also discussed the current FDA program. The Nationwide Evaluation of X-ray Trends (NEXT) program was conceived in the early 1970s to address the lack of a program to collect comprehensive popula- tion exposure data representing the state of practice in diagnostic x-ray imaging. A committee of federal and state radiation control representatives was formed to develop such a program, and within a few years NEXT was annually collecting data on 12 commonly performed diagnostic x-ray exams. State radiation control personnel conducted site visits to randomly identify clinical facilities and captured data regarding patient exposure, clinical technique factors, and exam workloads. By the early 1980s, NEXT abandoned the annual collection of data for multiple exams in favor of focusing on a single procedure. The surveys became more comprehensive, and patient-equivalent phantoms were developed to invoke radiation out- put representative of a typical patient. Film processing quality and the integrity of the darkroom were evaluated. 1 Primarily because at that point body CT exams were in their infancy.

OCR for page 17
22 TRACKING RADIATION EXPOSURE TABLE 3.2 Summary of NEXT Surveys and Survey Years Examination Survey Years Chest radiography 1984, 1986, 1994, 2001 Mammography 1985, 1988, 1992 Abdomen and lumbo-sacral (LS) spine radiography 1987, 1989, 1995, 2002 Fluoroscopy 1991, 1996, 2003, 2008 Computed tomography 1990, 2000, 2005 Dental radiography 1993, 1999, 2013 Pediatric chest 1998 SOURCE: Presentation by David Spelic, FDA. Surveys now routinely collect data regarding patient exposure, indi- cators of image quality, facility exam/procedure workloads, and staffing levels, as well as features of quality-control and quality-assurance practices, Dr. Spelic said. Surveys of particular exams are repeated periodically to capture trends in the state of practice. Statistical summaries of past NEXT surveys are available from the Conference of Radiation Control Program Directors.2 Approximately 40-43 states participate in each survey. A sum- mary of the NEXT surveys and survey years is presented in Table 3.2. The 2005 NEXT CT survey is an excellent example of the mutual benefits gained from collaboration with representatives from the manufac- turing sector, Dr. Spelic said. Representatives from the National Electrical Manufacturers’ Association (NEMA) supported the survey planning efforts with insight into the state-of-art CT technology. CDRH also has active representation on a number of International Electrotechnical Commission (IEC) committees, with standards activities directed at various sectors of diagnostic imaging from CT to digital-based imaging. He said that the wide acceptance of such standards by the international community underscores the need for continued presence at the federal agency level. NEXT supports these efforts with population data for exam frequencies, patient exposure, image quality indicators, and trends in the practice. Finally, IAEA has recognized the NEXT chest and abdomen/spine phantoms and associated protocols as scientifically established methodologies for conducting dosim- etry for these exams. Dr. Spelic identified several challenges that NEXT faces, including lim- ited human and financial resources.3 Moreover, the technology is changing faster than the ability to develop, execute, and publish surveys. NEXT aims to continue to complement and coordinate with newer efforts to capture 2 See: http://www.crcpd.org. 3 For example, an analysis of CT survey data from 2005 has not been completed because of insufficient resources.

OCR for page 17
23 WORKSHOP SUMMARY complex data via dose registries and to focus on surveys of exams and modalities that are presently outside the scope of current efforts to auto- mate dose data collection. 3.2.2 IMV Surveys IMV is a market research and database provider that uses a variety of survey methods to track diagnostic medical procedures, including those that expose patients to radiation. Although IMV covers a large number of imag- ing facilities, it does not provide a detailed categorization of procedures. Instead it provides estimates of the number of procedures overall or of the numbers by broad categories such as CT or MRI. Mr. Shah (general manager, IMV) and Ms. Prochaska (vice president, IMV) provided an overview of the data collected by IMV and perspectives on large-scale data collection. IMV classifies its studies into two categories: census databases and mar- ket reports. Census databases cover about 65 percent of the universe and include both hospitals and independent imaging centers. Time required for data collection depends on the modality. For example, it may take one year to complete the data collection process for PET, whose universe is about 2,000 sites, while it might take two years for CT, whose universe is 8,000 sites. Collection costs increase significantly after about 30-40 percent of the sample has been interviewed. Because the census database information is quite detailed and covers a large population, it can be segmented and drilled down depending on the question to be answered. For instance, by facility type or geography, summary information can be obtained on: • Availability of services (i.e., CT, PET, nuclear medicine) • Adoption of new technology • Number and/or age of systems in use A powerful tool that IMV uses to achieve its satisfactory participa- tion rates for the census databases (and a motivator for facilities) is that it donates a free benchmark report to participating institutions (price of report: $750), which contains information that they can use to compare their status and performance to the national average. Variables of interest to institutions may include number of clinical patient studies, variation of patient studies per site, productivity, radiopharmaceutical utilization and budgets, and inpatient versus outpatient composition. In contrast, IMV’s market reports cover approximately 300-400 sites (participation rates >10 percent), interviews are conducted by telephone and online, and collection of information takes about 8-12 weeks. Mr. Shah explained that recent changes in the health care environment

OCR for page 17
24 TRACKING RADIATION EXPOSURE have increased the workload of potential respondents who now have less time to participate in surveys. Additional challenges include an increasing number of providers to survey; a larger number of entities who survey the health care community (for example many manufacturers now survey their customers); facilities that do not allow their employees to participate in surveys; and the almost inevitable routing of calls to voice mail. Regardless of survey method, a trade-off exists between the level of detail requested and the response rate achieved. Workshop committee member Fred Mettler (New Mexico VA Health Care System) acknowledged IMV’s great contribution as a source of infor- mation on the utilization of medical diagnostic procedures for the NCRP report 160 (NCRP, 2009). 3.2.3 Medicare Administrative Claims Although there was no workshop presentation dedicated to admin- istrative claims as a source of information about population utilization of medical imaging, Dr. Chatfield briefly discussed this source. She said that detailed data on counts of procedures by current procedural termi- nology (CPT) code (or equivalent) for large populations have historically been available from administrative claims such as Medicare claims. She explained that CPT codes offer an advantage over the broad categories often used in surveys but still may not be granular enough to capture the full range of appropriate variation in radiation doses, protocols used to image patients for a broad range of indications, or amongst practices. Infor- mation is automatically collected using claim submissions from Medicare beneficiaries and is publicly available. However, it is limited to patients aged 65 or over who use this social insurance program. When available, data from private payers only cover each plan’s participants. 3.2.4 ACR Dose Index Registry The Dose Index Registry could serve as a source for both procedure counts and dose index measurements. Workshop committee member Rich- ard (Rick) Morin (chair, ACR Dose Index Registry) provided an overview of the registry (see Section 2.5.3 for additional information). By sending information to the Dose Index Registry, facilities can opti- mize protocols, implement standards, and contribute to the development of reference levels with the ultimate goal to improve imaging performance over time. Using a report (currently generated semi-annually) with descrip- tive statistics (mean, median, 25th and 75th percentiles) of the reported dose indices of participating facilities broken down by location, region, and type, the sample facilities can compare where they rank in these categories

OCR for page 17
25 WORKSHOP SUMMARY and against the Dose Index Registry. The indices are expected to become less variable and more aligned with the benchmarks. Dr. Morin suggested that each facility should task a qualified “safety committee” with reviewing the report and evaluating whether the facility’s dose indices are too high or too low compared to the ACR benchmarks. The committee could be com- prised of diagnostic radiologists, physicists, technologists, and diagnostic imaging experts. Participation of imaging experts was deemed essential by many workshop participants who stressed that monitoring the dose indices detached from image quality does not provide the required overall quality assurance. It is not surprising that “when somebody is watching, behavior changes,” Dr. Morin said. Using the Advanced Cardiovascular Imaging Consortium in Michigan as an example, he stated that voluntary, collab- orative quality improvement programs have proven to be successful in the past. The consortium achieved a marked reduction in estimated radiation doses following implementation of a radiation dose-reduction program, with no impairment of image quality. The one-year program used educa- tional intervention to disseminate to participating sites the best-practice recommendations for radiation dose reduction followed by a two-month monitoring stage (Raff et al., 2009). At the time the workshop took place, about 300 facilities were in the process of participating in the ACR Dose Index Registry and more than 100 had initiated data submission.4 These facilities are of different types (academic, community hospital, multi-specialty clinic, freestanding center) and are distributed around the country. Data from more than 350,000 CT exams were recorded. Drs. Morin and Chatfield described several challenges associated with the Dose Index Registry, which reflect general outstanding issues in the radiology community. For example, in the early pilot phase of the registry, naming conventions were largely inconsistent. Even if they used the same machine, different facilities may have named the procedure referred to as “CT head” differently. The issue also existed within a facility if differ- ent machines or different software were used. Now all exam names are standardized and mapped to RadLex5 terms. As a result, procedures can be grouped into standard categories. However, even though the names are standardized, the protocols between facilities may differ. Therefore, what 4 The number of institutions participating in the ACR Dose Index Registry increased to 400 between the times the workshop took place and the report was completed (communication with Rick Morin, chair, ACR Dose Index Registry). 5 RadLex is the lexicon for uniform indexing of radiology terminology implemented by the Radiological Society of North America (RSNA).

OCR for page 17
26 TRACKING RADIATION EXPOSURE is really needed is a standard name for acquisition protocol, Dr. Chatfield said. The second challenge relates to the variability in dose indices due to patient size. This issue was discussed many times throughout the workshop (see, in particular, Section 3.4) and highlighted by Dr. James Brink, profes- sor and Chair of the Department of Diagnostic Radiology at Yale University School of Medicine. Dr. Brink, together with colleagues, recently published findings that, for body CT examinations performed with automatic expo- sure control, the radiation used to examine a 100-kg patient is approxi- mately three times that for a 60-kg patient and results in organ doses that are generally twice as high as those in a 60-kg patient (Israel et al., 2010). The ACR Dose Index Registry currently does not cover all imaging modalities. It includes only CT but plans to also include computed radiog- raphy and digital radiography and fluoroscopy within the next year or two. Despite rapid growth, the registry currently includes only a small fraction of the CT facilities in the country. Participation is voluntary and therefore unlikely to be nationwide any time soon, Dr. Chatfield said. The fee to participate is modest ($500 one-time registration and additional charges scaled to the size of the practice) but may prevent participation by some facilities. Data transmission to the registry is completely automated, with high accuracy and minimal effort by the facilities, but some facilities may still hesitate to participate because of a reluctance to undertake a new and “unknown” effort. The registry uses industry standard practices for data protection and signed Business Associate Agreements (BAAs)6 to protect patient privacy. Facility information is shared only with the facility, and facility permission is sought before the facility’s name is included on the list of participants. Dr. Chatfield addressed the question of whether the Dose Index Reg- istry could be potentially used for population exposure monitoring. She responded that before that could happen there must be expert consensus and AAPM guidance on how to measure organ doses. If a system is imple- mented without expert consensus or without having adequate scientific justification for its value, then facilities will be reluctant to participate in the registry. This would hinder the main goal of developing better diagnos- tic reference levels for dose indices and would deprive facilities of a much needed tool for protocol review and radiation dose optimization. Because patient data are currently anonymized, multiple exams on the same patient cannot be identified, and patients cannot be followed as they move from facility to facility. 6 A BAA is a standard contract for the purpose of providing services involving the use of protected health information.

OCR for page 17
27 WORKSHOP SUMMARY 3.3 NATIONAL AND INTERNATIONAL EFFORTS IN DOSE TRACKING Many workshop participants emphasized that there is momentum for archiving of radiation exposure but multiple substantial barriers remain. These barriers include the questions of how to translate the various dose indices into a single quantity, whether the dose should be organ dose or effective dose, how to automate the collection process, how to account for individual variation in patient size, shape, and age, how to manage patient privacy and security issues, and how to control the multiple disparate pur- poses for which the data might be used. 3.3.1 Veterans Health Administration (VHA) The Veterans Health Administration (VHA) health system is the larg- est integrated health system in the United States, treating a specific patient population and only few young patients. VHA has developed an open source electronic medical records system, which facilitates communica- tion of the medical history of the patient, including access to the patient’s imaging exams. VA practitioners are protected from personal malpractice liability and their salary is not dependant on procedure volume. Despite the absence of these potential motivational parameters,7 diagnostic imaging and especially body CT usage in the VA is increasing at a rate similar to the private sector, noted Charles Anderson, chief consultant for diagnostic services at VHA. Many of these CT exams are performed as part of cancer screening, diagnosis, treatment, and surveillance protocols. In light of recent FDA notices regarding potential CT overexposures, the VHA surveyed all its hospitals for dose parameters associated with brain perfusion studies and found that none had exceeded the threshold for deterministic effects. VA has taken several steps to minimize the radiation dose received by patients, including the requirement for a national dose registry, although it is not clear whether funds will be available to develop the software, Dr. Anderson said. The plan is to send Digital Imaging and Communications in Medicine (DICOM)8 dose structured reports from CT scanners and fluo- roscopes to the VA image storage system (VistA Imaging). From there, dose parameters (volume CTDI and DLP for CT, cumulative air kerma and dose area product for fluoroscopy) will be extracted and placed in the procedure 7 These parameters have been discussed as few of the many reasons of increased diagnostic imaging utilization (Baker et al., 2008). 8 DICOM is an information technology standard designed to automatically capture and electronically report machine settings from various imaging procedures. DICOM is managed by the Medical Imaging and Technology Alliance, a division of NEMA.

OCR for page 17
42 TRACKING RADIATION EXPOSURE nomenclature that would facilitate any method or purpose of tracking. However, the manner in which radiation output, radiation doses, and any corresponding radiation risks are to be presented on medical imaging equipment must be based on a consensus by the medical imaging scientific community. Walter Huda commented that it is unrealistic and impractical to expect manufacturers to play a leading role in any such endeavor. Dominic Siewko (Philips Healthcare) noted that now more than ever it will take a coordinated effort of transparent communication between researchers, manufacturers, regulators, and care providers to ensure that the industry moves forward in lock-step. Tracking the Physician’s History of Ordering A member of the audience asked whether tracking the physician’s ordering history could reduce unnecessary imaging. Dr. Hricak (chair, Department of Radiology, MSKCC and workshop committee vice-chair) responded that given the fact that the end result of this group effort is to improve patient care, it is essential to track the physician’s history. Dr. Sodickson (section chief, Emergency Radiology, Brigham and Women’s Hospital) clarified that comparisons between clinical practices to assess physicians’ ordering histories should account for justified differences due to the patient populations being cared for. Dr. Morin shared examples of how a system that tracks the physician’s history of ordering improved practices at the Massachusetts General Hospi- tal. The hospital adapted and modified the ACR appropriateness criteria for exam ordering, and ordering physicians with low scores were consulted and subsequently received feedback regarding their ordering behavior (personal communication with Dr. Keith Dreyer, Harvard University, 2005). Physi- cians do not try to do something inappropriate, Dr. Morin said, “they just do not know.” Fred Mettler noted that easy and relatively quick checks on the ordering habits of the physicians can happen by routine review of the billing databases. Dr. Denison added that tracking a physician’s history with respect to dose (rather than number of procedures) even in a somewhat anonymized way is particularly important in interventional radiology. It is important to share the values with all physicians and alert them when important steps should be taken to lower doses. 3.6.3 Dose Monitoring and Individual Risk Assessment Although they voiced no arguments against tracking radiation expo- sures or doses and dose indices for the purposes of justification and opti- mization, some workshop participants disagreed about the desirability of tracking for the purposes of individual dose monitoring and risk assessment.

OCR for page 17
43 WORKSHOP SUMMARY A member of the audience asked Dr. Huda for his opinion as to whether patients should be given their dose periodically. In response, Dr. Huda rephrased the question to “Do I—as a patient—want to know my cumula- tive risk?” and categorically responded, “No! What an individual needs to know is whether he or she will benefit from the exam that is about to be ordered.” Dr. Mettler added that focusing on either dose or risks may become a problem in the future if patients refuse to have or physicians refuse to give an exam that the patient needs because of anxiety over the risks rather than appreciation of the benefits. Although for stochastic effects such as cancer risk dose tracking for individual risk assessment may not be needed, for deterministic effects, it may be good to know when those limits have been reached. Although not arguing with the points made, Dr. Brink reminded the workshop participants that if the medical community does not monitor individual doses responsibly and with control, then somebody else will provide (in fact, already has started to provide) cumulative dose and risk to the patient, potentially in a poor and inconsistent manner. The question remains, however, about what one does with the tracked information. Dr. Brink’s statement that the medical community should take the lead in track- ing individual doses was supported by others. What to Track and Communicate Although patients rely on their physician to guide them through clini- cal decisions, many workshop participants identified a trend in health care worldwide whereby patients want to know and understand more about the procedure they are about to have with the ultimate goal to improve their health care. This trend was compared to that of the implementation of nutritional information facts at the back of the products two decades ago. Although initially consumers were unsure about how to use the informa- tion, today many look at it for different reasons and want to know how to use the information effectively to make good choices. Dr. Hricak emphasized that because the physician still plays a funda- mental role in informing the patient, the only way to provide the necessary reassurance to the patient or help the patient understand the risks and benefits of a procedure is by helping the physician understand the poten- tial risks and benefits of the procedures ordered. It is important that the physician is able to provide the answers by being familiar with the current status and limitations of radiation dose estimations and risks. Today, many physicians are not adequately familiar with the radiation exposure effects, and training is crucial. Dr. Donald Frush (chief, Division of Pediatric Radiology at Duke Uni-

OCR for page 17
44 TRACKING RADIATION EXPOSURE versity Medical Center) identified elements of reassurance related to radia- tion exposure that come from knowing that the patient is about to get a good quality exam based on standard practice; the treating facility has active programs of optimizing exams; the providers are mindful of safety; and the scanners are accredited by an organization. However, some patients seek more than reassurance and ask about the dose received or the risks associated with a particular exam. Therefore the question remains: What is the most effective way to communicate doses or risks with those patients? Although there was no obvious approach as to how the dose or risk would best be communicated to the patient, or which of the two parameters should be tracked and communicated, many workshop participants said that little would be gained by communicating dose index metrics with the patient, especially because dose index metrics vary by modality and therefore do not provide a uniform recording system. Seemingly more meaningful is translating the dose index metrics to doses or risks or communicating in some generic way the increase of risk per exam (for example a 0.3 percent increase on top of the 42 percent baseline cancer risk) based on current knowledge. An alternative is to translate the risk to something more familiar to the patients, for example, the exposure to radiation during a flight from East to West coast, exposure to background radiation when living in Denver versus living in New York. (This approach was found to be too simplistic by some workshop participants including Drs. Brink and Sodickson who spoke against it.) Dr. Kevin Crowley (director, Nuclear and Radiation Studies Board, National Academy of Sciences) commented that there is substantial litera- ture on risk communication, including some National Academy of Sciences reports. It is known that different risks mean different things to different people based on subjective factors, and although numbers and statistical evidence may mean less to the general public, comparisons of risks from other sources, when risks are sufficiently similar, may prove useful to put risks into perspective. Regardless of the preferred method, a number of workshop participants commented that communication must be catered to both the patient’s inter- est to know and prior knowledge on the topic and should be done in an appropriate language and in a way to avoid causing panic. Uncertainties in the dose and risk estimations also should be clearly communicated. Related to the need for effective communication is an increasing debate regarding informed consent for ionizing radiation in diagnostic imaging (Nievelstein and Frush, 2012).

OCR for page 17
45 WORKSHOP SUMMARY Where to Report the Information Dr. Mettler compared the ordering of an exam with the ordering of a prescription, which makes it obvious that documenting the amount of radiation used in an exam is a reasonable thing to do. The question of where to document the amount of radiation—in the radiology report or the medical record—was further discussed. A member of the public pointed out that documenting doses in the radiology report, which is a legal document, may lead to future problems because doses are based on nonaccurate models and, in pediatrics in par- ticular, may differ from the actual doses received. Dr. Frush responded that these doses do not necessarily need to be recorded in the report but can be archived in some fashion in the medical record, and whether or not they go in the report will depend on state requirements while being mindful of the issues mentioned. Because the methods to estimate patient doses are not yet fully devel- oped and it is uncertain which are the most relevant, it might be necessary to record all parameters and dose indices provided by a scanner. This is in agreement with the prototype used in the California’s legislation (see Sec- tion 2.5.6), which notes that metrics such as CTDIvol and DLP or a dose unit as recommended by AAPM should be recorded. Dr. McNitt-Gray noted that such a recommendation is not likely to be implemented before the July 2012 deadline, but the AAPM can make a future recommendation for a more meaningful metric to be reported. For patients that want more information about their exposure and the possible health effects, different levels of information covering the different levels of interest could be incorporated into the report. Dr. Frush suggested that the report could include links directing the patient to the appropri- ate organization (such as ACR, RSNA, IAEA) or federal agency (such as FDA and NIH) for general dose information, specific dose information per modality, or risk estimations with an option to contact the institution’s radiology program if more patient-specific dose information is desired. The debate of whether the reported amount of radiation should be converted to dose or risk continued, with Dr. Brenner asserting that risk is what the patient most cares about. Fred Mettler pointed out that the idea of report- ing risk in relation to radiation amount in the medical record does not agree with current practices, for example in radiotherapy, chemotherapy, and other treatment options, which may also carry some risks. Assuming that a dose-recording system is needed for individual pur- poses, it has to be portable and cross boundaries to facilitate information tracking for all, including the mobile populations, Dr. Hricak said. Several possible approaches were discussed, including Smartphone technologies or

OCR for page 17
46 TRACKING RADIATION EXPOSURE centralized technologies that enable patients to enter treatment information not matter where they receive the treatment. Past Exposure Informing Decision Making Section 3.1 of this report discussed that insufficient information about a patient’s history of exams (e.g., whether a procedure was done in a different facility and its outcome) may lead to unjustified ordering of an exam. The opinions discussed in this section regarding past exposure informing deci- sion making are not related to the issue of insufficient information resulting in duplicate or questionable ordering. Instead, they relate to whether and how the history of exposure to medical radiation (e.g., too many CT exams in a patient’s record) fits into the clinical decision of ordering the next exam that utilizes ionizing radiation. A member of the audience stated that, even within the walls of a hospi- tal, a database that is easily accessed from a workstation and provides the ordering history for a patient can affect practices and reduce the number of exams ordered. In such a database, collecting information on the number of procedures rather than the doses may provide a sufficient wake-up call for the chief technologist. Dr. Sodickson agreed with the comment and added that real-time support rather than dose registry type of implementations can be factored into clinical decision making. Furthermore, the clinical model of taking clinical decision today based solely on clinical presentation today is changing; practice must move from episodic decisions to more long-term care of the patient by looking at the entire medical history and exposure. This may be more evident at the primary care level, when deciding whether physicians are doing the right thing over the course of a patient’s treatment. Ms. Gwen Darien (at the time of the workshop, executive director, Samuel Waxman Cancer Research Foundation; currently, director, The Pathways Project), representing the cancer survivors’ views, shared the idea that ordering history information should fit with clinical decision mak- ing. It is critical to have knowledge of the previous exposures to radiation and potential harms from past, present, and future exposure as well as how those exposures might interact with medical treatment of the specific patient. To truly improve patient outcomes, it is vitally important to con- sider the long term when making decisions, she said. Other participants, including Drs. Huda, Brenner, and Frush disagreed and stated that if an exam is clinically justified, then it is justified regardless of the past history; however, knowing the exposure history of a patient may serve as an important reminder that other means to diagnose a problem are possible. Dr. Brenner described screening as an area where a risk versus benefit justification discussion is valid. Dr. Sodickson responded that in the emer-

OCR for page 17
47 WORKSHOP SUMMARY gency room setting many exams border on being screening exams, and the yield is low. For example, the positive rate for a study to rule out dissection in a patient with chest or back pain is only 2 percent; however, the impor- tance of making a life-or-death diagnosis in these few patients is critical. He also emphasized that justification of an exam is often a grey area, and the right thing to do is not always obvious. Although past exposures might be part of the decision-making process for ordering the next exam, there is no threshold above which you cut off some patient from further imaging, Dr. Sodickson clarified. Dr. Berrington de González noted that the United States could possibly make use of the justification systems that are in place in other countries such as the United Kingdom, where CT use is seven-fold lower than in the United States. All requests for diagnostic imaging procedures that involve ionizing radiation have to be approved by a radiologist, and the process requires justification of the need for the test.22 The Patients’ Perspective Gwen Darien discussed the patients’ perspectives on the risks and ben- efits of radiation exposure from medical diagnostic procedures. As a cancer survivor herself, her perspective was that of the cancer survivor rather than the typical symptomatic or asymptomatic patient. Ms. Darien explained that the health goals and concerns of cancer survivors may be different from those of other patients, and therefore their perspectives and expectations may also differ. In order to pose a question to their health care provider, they need to know that there is a question to ask. It is not clear whether most cancer survivors know that there is a question to ask regarding risks associated with medical imaging. She noted that overall there is little discus- sion between patients and health care providers on the risks and benefits of radiation exposure from medical diagnostics. Although the benefits are often assumed even when not explicitly dis- cussed, the risks are rarely mentioned. In the absence of risk/benefit discus- sions between patient and health care providers, there is often a tendency for patients to request more procedures, Ms. Darien said. For many cancer survivors there is fear and anxiety of “not knowing” and of disease recur- rence. From the perspective of a cancer survivor, it is critical to understand how radiation exposure might interact with medical treatment. With the ultimate goal of informed decision making, patient-provider interactions must include a discussion of the need for test, what knowledge will be gained from the test, and how that knowledge will be used to benefit 22 Royal College of Radiology—A guide to justification for clinical radiologists. See: http:// www.rcr.ac.uk/publications.aspx?PageID=310&PublicationID=2).

OCR for page 17
48 TRACKING RADIATION EXPOSURE the patient. Ms. Darien noted that cancer survivors would be primarily concerned with not being able to get the tests they perceive they need if a tracking system for radiation dose were in place. Are We Ready to Track? On the topic of whether—if desirable—we are ready to track indi- vidual doses, Dr. Sodickson described an effective system of individual dose tracking, which is based on electronic medical records as likely the most efficient way to store information about the patient’s history of radiation exposure. This system would include all care sites, and as a patient moves from one state to another his/her doses would be recorded via a unique patient identifier. The system would be able to track all modalities and sources of exposure and modality-specific exposure metrics or technique parameters together with accurate patient-centric dosimetry. The database format would be standardized, and all systems would be connected without firewalls or barriers. This system differs from the current reality, which involves some inde- pendent modality-specific efforts, he said. This is because exposure metrics and platforms are different, and most of the captured modalities are for CT, because of its high doses and public attention, and for fluoroscopy, because it is more regulated in terms of the deterministic effects. Impor- tant data elements have been missing, such as exposure metrics, dose, and parameters related to the patient’s size. Moreover, data access is limited, and data collection processes are not often automated. The format is largely inaccessible such as screen captures and text reports and is buried in discon- nected systems. For these reasons, many workshop participants asserted that—even if desirable—tracking cumulative dose estimates from a single or multiple modalities of a patient at the national level cannot happen today. At an institutional level, dose indices or dose estimates of a patient associated with a single procedure could be tracked, and possibly some institutions could track cumulative dose estimates from a single but not multiple modalities. In support of isolated institutional efforts to track cumulative dose estimates from a single modality of a patient, Dr. Sodickson presented efforts developed at the Brigham and Women’s Hospital. The team devel- oped an open source informatics toolkit named GROK (General Radia- tion Observation Kit), which can locate and retrieve CT exposure metrics CTDIvol and DLP from existing digital image archives and convert them to text. Moreover, anatomy assignment algorithms use the combined dose report screen text and DICOM attributes data to determine the anatomic regions irradiated, creating large repositories of historical anatomy-specific radiation exposure metrics information from CT (Sodickson et al., 2012).

OCR for page 17
49 WORKSHOP SUMMARY Using similar logic, an open source toolkit, PARSE (Perl Automation for Radiopharmaceutical Selection and Extraction), was created to extract exam- and patient-specific dosimetry from the medical records of patients, which contain unstructured text including the administered activity and the radiopharmaceutical name (Ikuta et al., 2012). Both methods proved satisfactory validation yields in data retrieval (97-99 percent) and anatomic assignment precision (94-99 percent) and may prove to be promising tools for estimating patient-specific radiation dose and cumulative risk. 3.6.4 Research The gaps in current knowledge and the need to explore and refine models of biological effects at low doses were demonstrated by the presen- tations of Drs. Brenner and Mabuchi. Dr. McNitt-Gray argued that as the natural experiment of the effects of medical imaging procedures that use radiation is happening, it would be wise to collect good dosimetry data. This effort could answer the epidemiologists’ questions and improve current knowledge of the biological effects of low-level radiation without the need to extrapolate from other population sources, which introduces uncertainty into measurement and interpretation. Dr. Berrington de González responded to this idea by saying that, although using a tracking system would be beneficial for epidemiologic studies, capturing the study end point, such as cancer occurrence or death from cancer, and linking it with the exposure information is necessary for an epidemiologic investigation. In the absence of a centralized cancer registry in the United States that could provide the cancer ascertainment information, this is a difficult task. Dr. Lauer added that attempting to find the potential association of imaging and cancer risks is important, but other risks not related to cancer also need to be tracked to assess the appropriateness of an imaging exam. All medical procedures contain an element of danger, and a potential to discover incidental findings that require subsequent medical evaluations. These evaluations may not only not improve outcomes but are likely to induce harm (Lauer, 2009). 3.7 LESSONS LEARNED FROM PEDIATRICS Dr. Frush was invited to discuss the exposure reduction efforts and lessons learned from pediatrics. The pediatric radiology subspecialty is not typically on the horizon-defining medical trajectory, he said. Exceptions to this are efforts related to medical radiation, specifically radiation dose and potential risk from CT imaging, and education and advocacy for radiation

OCR for page 17
50 TRACKING RADIATION EXPOSURE awareness and protection. Dr. Frush described some of the accomplish- ments resulting from efforts through pediatric imaging. First, he said, we can lower doses and read noisier examinations, and we ought to do this. A number of reports in the pediatric population as well as a growing body of literature in the adult population promote dose reduc- tion with maintenance of image quality. There is systematic work in dose reduction in adult and pediatric renal calculus evaluation (Karmazyn et al., 2009; Paulson et al., 2008). A study that compared the diagnostic capabili- ties of standard- and reduced-dose CT in the detection of nephroureteroli- thiasis in children showed that use of the 80 mA setting for all children and 40 mA for children weighing 50 kg or less does not significantly affect the diagnosis of pediatric renal stones (Karmazyn et al., 2009). Second, multiphase (repeated scanning before and after contrast injec- tion) examinations should be justified in adults, as has been promoted in the pediatric population, although multiphase examinations in children are generally not protocol driven, Dr. Frush said. When necessary, they could be based on a more case-by-case approach. Overall, multiphase examina- tions constitute fewer than 5 percent of all pediatric body CT examinations. This philosophy is less pervasive in adult imaging. One recent investigation noted that the frequency of “double scans” of the abdomen was highly variable, and the mean effective dose could have been reduced by about one-third overall with adherence to ACR appropriateness criteria (Guite et al., 2011). He recognized the success of the Image Gently campaign, whose main pillars stand upon the foundations of a respected organization and leader- ship with independence and integrity, and consensus involvement. The mes- sages are simple, important, and promoted in a positive and constructive (rather than alarmist) manner, with carefully controlled delivery of content, timing of releases, and schooled spokespersons to assure consistency and maximized penetration and impact. He noted, however, that one must still be mindful of the various remain- ing needs in both pediatric and adult imaging, which include helpful dose alerts and notifications, improved dose estimations that account for the patient’s characteristics (age, weight, size, gender), and establishment of reference values. Although templates exist on new scanners, without rea- sonable guidance for their use, these capabilities may be underutilized or incorrectly utilized, and therefore ineffective. Dose estimations for CT in children are often inaccurate. The avail- ability of improved dose estimations, such as through the AAPM task group 204 (AAPM, 2011), is a clear improvement. Current work in many labora- tories focuses on patient-specific (i.e., age, weight, size, gender) organ dose estimations and resulting effective dose estimations. Dr. Frush discussed that a pilot registry for pediatric body CT (QuIRCC) is working in parallel

OCR for page 17
51 WORKSHOP SUMMARY with the ACR Dose Index Registry. Early results from this pediatric registry demonstrate that the body CT dose indices at six pediatric institutions are below those reported in the European community (personal communica- tion, Dr. Marilyn Goske, Cincinnati, Ohio). 3.8 SOME POSSIBLE NEXT STEPS SUGGESTED AT THE WORKSHOP This section summarizes the key points and suggestions on some pos- sible next steps discussed throughout the 1.5-day workshop and highlighted during the final panel session moderated by Dr. Barbara McNeil, profes- sor and head of the Department of Health Care Policy at Harvard Medi- cal School and workshop committee chair. The four panelists were Drs. Brenner, Frush, Hricak, and Mettler. Many workshop participants noted that a primary motivator for track- ing doses was to implement and maintain dose reduction strategies through optimization and justification with the ultimate goal to improve care. Sev- eral participants asserted that such strategies ought to be adopted by all facilities that perform diagnostic imaging, including hospitals and imaging centers, as well as free-standing private physician, dental, and chiropractor practices. Dr. Hricak emphasized that although it may be straightforward for major hospitals to adapt and adhere to practice guidelines, it may be challenging for free-standing imaging centers and small community hospi- tals to do so. Still, the goal of any imaging facility ought to be to improve radiologic services to the patient independent of the available resources. Some workshop participants stated that it would be desirable to have a national registry that tracks radiation exposures and/or doses from medical diagnostic procedures. However, such a national effort is not likely to be implemented in the near future for many reasons, including lack of shar- ing of medical information across different health care facilities, lack of a unique patient identifier and integrated medical records, non-automated dose information collection processes, and data protection and patient privacy issues. Similarly, the current health care delivery system and cancer registration system precludes a longitudinal study of dose for large popula- tions in the United States that are exposed to ionizing radiation from medi- cal diagnostic procedures. In view of the above mentioned barriers, the key points and sugges- tions on some possible next steps discussed by the panelists and workshop participants were to: • Continue to track and monitor overall trends and patterns of use of medical imaging.

OCR for page 17
52 TRACKING RADIATION EXPOSURE • Continue ACR dose index registry efforts and expand them to include additional modalities (e.g., nuclear medicine, computed radiography and digital radiography, interventional radiology) and other sites, particularly outpatient facilities. • Within each institution, routinely report dose metrics performance with benchmarks. • Create or use existing committees within institutions and outside facilities to ensure that imaging protocols are being followed; create routine reports for this purpose for technologists and radiologists. • Work with industry and information technology vendors to incor- porate dose metrics directly into medical records; ensure that dose metric information is attached to images. • Encourage the performance of national-level clinical trials that quantify the benefits of imaging exams. • Implement informed decision support systems at all stages of patient care to optimize procedure use and ensure that only appro- priate examinations are performed. • Have institutions and ambulatory settings implement or continue to implement comprehensive safety programs and educational tools promoting awareness of radiation doses. The above mentioned points do not represent a consensus of the workshop participants or the authoring committee. Instead, they represent some of the important points made by individual participants during the workshop.