Both of the studies described in this chapter reviewed reports of AMDEs received by the FDA to identify and characterize EMI-related adverse events. The earlier study focused on medical devices in the cardiovascular panel, relied on text searches to identify EMI events, and for the reports selected rated the likelihood that the event was EMI-related. The more recent study focused on reports that met the definition for suspected EMI cases, used problem codes to identify suspected EMI cases, excluded reports in which the information needed to determine whether the event was EMI-related was insufficient or ambiguous, and included reports sent by manufacturers, distributors, user facilities, and voluntary sources. Although the methodologies used to identify EMI events differed, the results of the two studies were similar. In both studies, pacemakers were the medical devices most frequently reported to be affected and deaths were rarely reported.
In the earlier FDA study, an activated electrosurgical unit or an external defibrillator in the vicinity of the pacemaker or ICD were the most frequent potential EMD sources. In the more recent study, the most frequent non-medical EMD sources were radiated electromagnetic energy emitted from products that included wireless communication devices and security systems in public places. The different study periods and the recent proliferation of new wireless technology may explain the difference in EMD sources identified by each study.
The adverse events reported in the MAUDE database mirrored the reports in the medical literature. Cellular or mobile phones were reported as potential causes of interference with a variety of medical devices. Cellular phones have been reported to interfere with implanted devices, such as a bone-anchored hearing aid , monitoring devices, such as ionizing radiation dose monitoring equipment , and therapeutic devices, such as mechanical ventilators [24,25] and infusion pumps . Appropriate medical device design, testing, and separation distance appear to be adequate in some cases, resulting in recommendations for patients with implanted electronic devices, such as pacemakers, implantable cardioverter-defibrillators, and implantable deep brain stimulators, to keep a separation distance between their cellular phone and their implanted device [27,28]. Cellular phone and personal digital assistant (PDA) use in the hospital or clinic has been controversial, with some asserting that medical device filtering, shielding and separation distance from the radio-wave emitter are adequate with modern medical devices and modern cellular phones, and others unable to agree [29-32].
A case of acute epinephrine poisoning caused by cellular phone interference reported in a medical journal and to FDA illustrates the serious health consequences of EMI with a medical device . An 18 year-old with septic shock was admitted to the pediatric intensive care unit of a US hospital. An epinephrine infusion was initiated, and after 9 hours the patient abruptly complained of headache and abdominal and chest pain. The patient's blood pressure increased, and he developed pulmonary edema, electrocardiogram (ECG) changes and myocardial damage, as evidenced by elevated cardiac enzymes. After noticing that the epinephrine infusion bag contained 24 ml less than it shouldhave, the hospital staff noted that a relative of the patient had received a phone call on a personal cellular telephone just before the patient's acute episode. The hospital's engineering department removed the infusion pump from service, reset its rate and observed the device as several cell phones were held at different distances from the pump while receiving a call. The pump spontaneously delivered a higher rate of epinephrine during several test calls. Although the engineering department was unable to consistently reproduce the infusion malfunction, it noted that the rate change was more frequent when the cell phone was closer to the infusion pump. A review of the history log of the pump showed no record of the event. It was concluded that EMD from the cell phone caused the infusion pump to malfunction and deliver an overdose of epinephrine, which led to epinephrine poisoning. This case also illustrates how the problem was recognized, investigated by the clinical and engineering staff, and reported to FDA.
Patients with pacemakers or implanted cardioverter defibrillators are made aware of the potential for EMI and are typically warned to avoid contact with potential sources of EMD, including instructions not to put a cellular phone in their shirt pocket or near their implanted device. While informing patients is important, this is not enough to mitigate the problem. Adequate EMCs designed into devices and policies to prevent EMI and the adverse patient outcomes that may result are also needed.
As with reports to the FDA, medical device interference caused by other medical devices was identified or considered in a number of situations. Examples include infusion pumps failing during radiotherapy sessions , interference with pacemakers by endoscopic video capsules , potential interference with cochlear implants by dental devices , the interference of a thalamic stimulator with an electrocardiogram monitor , and the interference with implantable cardioverter-defibrillators by surgical devices or interventional procedures [38,39].
Interference with implanted devices has also been reported or considered from common household electronic products, including convection ovens  and washing machines . The workplace may also present potential problems and consideration of the workplace environment may be necessary for patients' safety when returning to work after surgery [42,43]. Reports of airport metal detectors and electronic article surveillance (antitheft) device interference with implanted devices are also considered and dismissed [44.45]. However, such conclusions would seem to ignore the concerns raised by the significant number of reports identified by FDA as attributable to medical device EMI . Also, some types of medical devices are more susceptible to these EMD sources than others or their failure has a more significant adverse impact on the patient.
The prevalence and incidence of EMI events with medical devices and their significance are difficult to gauge. Underreporting of EMI appears to be common, and a number of factors contribute to it . The healthcare staff may not be aware of the causes of EMI or how EMI is manifested. The intermittent nature of the effects of EMI on device function often makes it difficult to recognize EMI events. The device may function normally after the incident or the user may find a 'work around' that allows him/her to use the device. Fear of liability or litigation may also inhibit a potential reporter.
Typically, events that result in a patient injury are more likely to be noticed and reported. While some events may be distinctly recalled by patients, such as the patient with a spinal cord stimulator for pain relief who reported being thrown to the floor from the interference caused by a library antitheft system (Table 20.2), others, such as the silent resetting of a patient's ICD, may go unnoticed until noted during routine follow-up with the patient. One study reported the incidence of antitachycardia therapy suspension due to possible EMI in implantable cardioverter defibrillators as 11% per patient year . In this unique study, medical records for 46 patients with ICDs capable of storing information on magnet reversions were retrospectively reviewed for such events. In none of these patients were tachyarrhythmias present at the time of the possible EMI-related ICD inactivation.
To complicate matters, it is not certain that all of these types of events can be attributed to EMI. It is possible that device malfunction, exclusive of EMI, may contribute to deactivation of ICDs or inappropriate behavior of implanted electronic devices. Ruling out device malfunctions, particularly when they are idiopathic, is not always possible. Suspected EMI is difficult to confirm and few attempt to do so. A review of the potential for mobile phone interference with medical equipment found considerable heterogeneity in conduct of tests for EMI with specific medical devices . A clear understanding of when EMI becomes clinically relevant is generally lacking, which leaves the importance of EMI with medical devices difficult to interpret.
A major limitation of our analyses is that not all EMI-related AMDEs are reported to FDA, so the reports we identified do not represent the universe of cases that occur. Among the other limitations of our analyses was the high proportion of missing values for some of the study variables, which limited our ability to characterize the EMI-related events. Data on many study variables are not systematically collected. The search strategies used in the studies we reported retrieved many EMI-related AMDEs, but it is likely that we may not have found all the relevant reports. We relied on problem codes that might be assigned by the manufacturer when they make the report or assigned by a contractor who enters the data into the FDA database. We found that 10% of the reports coded as EMI-related were in fact not related to EMI.
The problem codes in the MAUDE database could be improved. For example, modifying problem codes to differentiate between electromagnetic interference and interference related to the reactions with reagents would focus the search results on relevant reports. Similarly, developing more specific problem codes might make it easier to distinguish EMI events from events that resulted from electrical overstress. Information in the event report may be inadequate to distinguish between these two types of events. We recommend using text searches to complement the search by problem codes; text searches could help to identify additional relevant reports not identified by problem codes and may also be useful to validate the results of searches conducted with problem codes.
Despite the limitations noted, the analyses we reported were based on a nationwide database that provided over 10 years of data on EMI events. The earlier study helped to define problem codes for the MAUDE database developed when user facility regulations went into effect. The information gathered from the reports also contributed to the development of test methods and improved EMC requirements for national and international consensus standards. For the more recent study, the MAUDE database provided problem codes designed specifically to identify EMI events. Furthermore, these studies illustrate the methodologies that can be used to evaluate other problem codes designed to identify other postmarket problems.
The tip-of-the-iceberg paradigm diagrammed in Figure 20.5a and b describes the data we currently have and the data we would ideally have to conduct postmarket surveillance of EMI-related adverse events. The base of the pyramid shows the universe of adverse events, which must be identified, reported, properly coded, and entered in the MAUDE database in a timely manner. We currently only see the tip of the iceberg; efforts to increase public awareness of the injuries that result from EMI must continue. FDA established the CDRH EMC Work Group and, in collaboration with professional organizations, is disseminating information about EMI and approaches to assure EMC by posting public health advisories when necessary, publishing and presenting scientific papers, and participating in the development of guidelines and recommendations for EMC/EMI management and standards for medical device EMC. One of the EMC Work Group's goals is to educate the public and increase recognition of the problem, and in turn increase the percentage of events reported.
Improving the quality and completeness of the information reported about those events is a related goal. Postmarket surveillance of EMI-related ADMEs would be more informative if the reports systematically documented the patient's history and outcomes, the indications for the use of the device, the nature of the malfunction, the likelihood that the event is EMI-related, the duration of the EMI's effect on the device, and whether the device function is restored spontaneously, restored with the patient's intervention, restored with clinical intervention, or fails irreparably. As mentioned above, MedSun is an FDA program that is designed to improve postmarket surveillance for adverse events. MedSun is expected to provide more complete clinical and laboratory data and information about use errors and the conditions under which new devices are used after FDA approval. Online reporting is also being considered to solve the limitation of the time lag in entering reports into the system.
Finally, event reports must be evaluated, coded properly, and entered in the database. About 10% of the reports we retrieved as EMI-related were coded incorrectly, according to our criteria. Some non-EMI events were coded as EMI, and some EMI events were not coded as such. The proportion of EMI events incorrectly coded as non-EMI events remains unknown. Ideally, the number of EMI adverse events identified by the postmarket surveillance system, documented and entered in MAUDE, would be closer to the universe of actual EMI events (the bottom of the 'ocean floor' in Figure 20.5b). This is difficult to achieve.
As public awareness increases, and the recognition of events, reporting, coding and data management improves, more accurate estimates of the number of EMI-related adverse events can be obtained. Information provided by an improved surveillance system can be used to further increase public awareness about the medical devices that could be affected by EMI and the types of malfunctions that EMI can cause. Additionally, an improved postmarket surveillance system can help identify environments and conditions of use that may need to be mitigated or studied further; and it can help identify appropriate actions to be taken in response to new EMI-related adverse events.
With the rising dependence on medical devices of all kinds and particularly on implants and other devices with electronic components, ultimately design that protects medical devices from unwanted EMI will be necessary to protect patients from the wide variety of electromagnetic energy and sources increasingly found in the environment in which medical devices are used . In the meantime, medical device manufacturers and healthcare providers should provide patients with adequate information about the prevalence of EMI with medical devices, particularly with devices that are life-sustaining, inform them about the risks inherent in using a particular device, and report EMI problems to the manufacturer or to MedWatch.
Was this article helpful?