Donald M. Berwick, M.D., president and CEO, Institute for Healthcare Improvement, Cambridge, Mass., also is a clinical professor of pediatrics and health care policy at the Harvard Medical School, Boston. An elected member of the Institute of Medicine (IOM), Berwick now serves on the IOM's governing Council. He served on President Clinton’s Advisory Commission on Consumer Protection and Quality in the Health Care Industry.
When a team of researchers in the Netherlands tested the effects of radio frequency identification (RFID) tags in hospital settings, they uncovered some surprising—and potentially alarming—results: The radio waves the tags emit for tracking patients, devices and medications in some cases altered ventilator settings and turned off IV pumps. The June 25 issue of the Journal of the American Medical Association (JAMA) presented the study’s findings. Based on the results of 123 EMI tests, the authors found 34 cases of electromagnetic interference caused by RFID devices. Twenty-four of the incidents were serious enough to be classified as “hazardous” or “significant,” the JAMA report said.
Although the study took place in a nonclinical setting, the researchers concluded that RFID can potentially create hazardous incidents in medical devices and that RFID implementations in critical-care environments “should require on-site [electromagnetic interference] tests and updates of international standards.” In the same issue, JAMA published a related editorial about health care safety practices by Donald Berwick, M.D., president and chief executive officer of the Institute for Healthcare Improvement, a not-for-profit organization headquartered in Cambridge, Mass. In an interview conducted after the JAMA article first appeared, Berwick stopped short of advocating a moratorium or outright ban on RFID devices in clinical settings. But he did call for device makers, the FDA, materials managers and others to push for testing that not only ensures that new products work as advertised, but investigate how they interact with other technologies in real-world environments.
Q In your view, how serious is the problem the Dutch research team uncovered? Are patients at risk because of the presence of RFID devices in hospitals?
A There are two levels of concern. One is the specific concern of RFID. [The authors] revealed some well documented forms of technical interference between RFID devices and [medical] devices, and some of the unanticipated impacts on other equipment could be of consequence to patients such as changing settings on respirators, turning IV pumps on and off, and so on. In the study, there were no patients; it was all done in vitro, as it were, and in intensive care units with machines that were not connected to patients. We don’t know whether any patient would have directly been harmed, but there’s reason to believe they might. So this needs urgent attention from the FDA, from the manufacturers and from intensive care leaders to determine the potential for harm. It would be overreacting, in my opinion, to declare an instant moratorium or to change processes immediately, but I do think understanding interference and its consequences is urgently needed. Of course, there’s also a technical, research-and-development piece, because presumably, some of this interference can be mitigated with better designs. And the responsibility for that lies with the manufacturers. Manufacturers have a duty based on this single study to look for mitigating design changes that would reduce interference.
Q So why not call for a moratorium on RFID in health care settings until everything is sorted out?
A I think there is probably as much of a risk of overreacting as there is of underreacting. It doesn’t seem wise to me that everyone shuts machines off right now just because there’s this initial study showing there might be some trouble.
Q You’ve also said that hospitals have some immediate work to do to ensure patient safety regarding RFID in clinical settings. Which departments in particular should be looking at this issue and what sorts of steps should they be taking?
A Clinicians who deal with very ill patients and the materials managers who manage supplies and equipment in hospitals should ask a number of questions: “Where is RFID in common use in our hospital?” and especially “In which areas are patients being helped by other types of devices?” “Can we locally study interference and interactions to see if it’s happening here?” I suspect those areas are going to be intensive care, operating rooms and units where patients are sickest.
Q Before further studies can take place, is there anything materials managers can do as they go about purchasing technology to mitigate any risks that might be out there? How can they navigate in these waters?
A I guess there has to be some form of local testing going on as part of the materials management process. One thing that is obvious to me is that this single study should be the object of some conversation at national materials management professional meetings. And this should be raised to the plenary podium because we have a report that should cause us concern. How does this relate to patient safety? What do we know about this?
Q You mentioned at the beginning that this study uncovers two areas of concern—what does it say beyond just RFID?
A If you take this as a case study of technology and its side effects, there are a couple more general points that come out of this study, which I think are a heads-up for everybody, regardless of whether they are using RFID in intensive care units. One is that any innovation—and that includes any technical innovation—brings with it both benefits and risks. That’s always true. So if you add cruise control to a car, it has benefits, but it also puts a new demand on the driver and probably creates new forms of accidents. We need to understand that technology is double-edged. So management and leadership should go into using new technologies with an alert sense of risks and with what Carl Weick, the great scientist of risk in organizations, calls a “preoccupation with failure.” It means that when you make a change, even a well intended one, if you become preoccupied with a possibility of failure you have a much higher chance of safe change. But remember that technology has a benefit. So to overreact and say, “We are going to be Luddites about this and just end using RFID” or not have technology would be foolish. It’s the safe introduction of technology that we are talking about, not the nonintroduction of technology.
Q In that context, were you surprised to learn about the questions raised by the study and perhaps a lack of pre-testing done on RFID devices?
A I don’t know if pretesting was done. I think an interesting case study could be around what was and wasn’t tested in advance with these devices. I presume, for example, that it’s possible that manufacturers were testing interference against their own equipment but not against the equipment of others. That might be an explanation of why this wasn’t detected earlier. What surprised me was the severity of the interference. I would have assumed there would have been fuzzy screens and minor glitches. But a very high proportion of the instances of interference were rated by intensive care physicians as potentially quite serious. And that to me is a surprise.
Q Do you think it’s appropriate or necessary for the Federal Communications Commission (FCC) or the Food and Drug Administration (FDA) to get involved with this?
A I assume the FDA is involved with this. I don’t know about the FCC. But since the FDA does have, I believe, jurisdiction over device changes as well, I assume they are already in this game.
Q Would you like to see well defined regulatory guidelines surrounding RFID?
A Yes I would, both for premarket introductions and, more importantly in this case, post-market surveillance. Just as with drugs. We need a robust and healthy FDA that can gather information and know even after something’s been approved whether there’s a hazard that was not detected during the design and early introduction phases.
Q What do you think are some of the biggest takeaways for manufacturers from studies such as this one?
A Some questions for manufacturers are: “Did you know this in advance?” and if not, “what changes in your testing, design, and development are advised by this study?” “Have you now learned something about your approach to design?” Or maybe they did do the testing and maybe this was already known, in which case my questions are, “What have you done to both mitigate the interference and to make sure that the clinicians who are responsible for the patients are aware of this as a potential problem?” In other words, were you caught sleeping or not?
Q Do you think manufacturers do enough to work with the health care industry to ensure that standards are being met and to identify potential problems such as these?
A I’m sure that manufacturers on the whole are responsible citizens regarding the safety of their products. Their children and their parents may well become the patients who could be affected by this. So I have a rather trusting view of this. But I think there are probably some serious challenges here, which is true with any innovation and certainly ones with a commercial interest: When [manufacturers] introduce a change, they may be unconsciously biased toward noticing the benefits without really being aware of the side effects.
So there is always a need for diligence on the producers’ side to maintain skepticism, especially when it might go against their commercial interest. More than that, and I don’t know this, but I wonder if maybe this issue of cross-manufacturer testing is part of the problem here. So if your company makes ventilator pumps and my company makes RFID devices, do I get your equipment and test [the two together]? I don’t know the answer to that. We have a very important example of that in the drug side, which has to do with look-alike labels and names. We need a much stronger [push] in this country to ensure that when a new drug is made or a new label is printed that they aren’t easily confused with a different drug and a different label. And we have innumerable examples of this kind of confusion, which leads to hazards. Mike Cohen from the Institute for Safe Medication has for decades spoken about and documented this. But here we have a failure of cooperation. The industry on the pharmaceutical side is not committed to coordinating and cooperating to avoid the possibility of confusion. We can approach that either with self-guided industry standards that insist on cooperation, or within a regulatory framework by saying, “If you’re not going to do it, we’re going to make you do it.” I don’t know which solution we have to turn to, but I do know that at the moment, we are not where we need to be.