Effect Measure

We take for it granted that technology can be used to tag objects in various ways, useful and otherwise. The anti-theft devices used on retail clothing stores are a familiar example. Radio Frequency Identification Devices (RFID) are used for this purpose as well as for security access. I have a device like that on my windshield for automatic highway tolls on the turnpike. Hospitals also have a strong interest in keeping track of lots of items like pharmaceuticals, equipment or even ordinary sponges used in surgery. Counting and keeping track of sponges is routine so none are inadvertently left in the patient. Tagging sponges with little RFIDs is just one obvious use. But a new study published in the Journal of the American Medical Association (JAMA. 2008;299(24):2884-2890) by a group of Dutch researchers raises some serious questions about indiscriminate RFID use in the high tech hospital setting:

Remko van der Togt, M.Sc., of Vrije University, Amsterdam, the Netherlands, and colleagues conducted a study in a controlled, non-clinical setting to assess and classify incidents of electromagnetic interference by RFID on critical care equipment. The tests were performed in a one-bed patient room in an intensive care unit (ICU) and with no patients present. Electromagnetic interference by two RFID systems (active [with batteries and ability to transmit information] and passive [without batteries, information retrieved by RFID reader] was assessed in the proximity of 41 medical devices (in 17 categories, 22 different manufacturers). The devices included items such as external pacemakers, mechanical ventilators, infusion/syringe pumps, dialysis devices, defibrillators, monitors and anesthesia devices. Incidents of EMI were classified according to a critical care adverse events scale as hazardous, significant, or light.

All 41 medical devices were submitted to 3 EMI tests resulting in 123 EMI tests. A total of 34 EMI incidents were found; 22 were classified as hazardous, 2 as significant, and 10 as light. The passive signal induced a higher number of incidents (26 in 41 EMI tests; 63 percent), and hazardous incidents (17), compared with the active signal.

Hazardous incidents included: total switch-off and change in set ventilation rate of mechanical ventilators; complete stoppage of syringe pumps; malfunction of external pacemakers; complete stoppage of renal replacement devices, and interference in the atrial and ventricular electrogram curve read by the pacemaker programmer. (AMA Press Release, hat tip Medgadget)

We now know this can happen. What makes the think it will happen is how all this gets done in a hospital. The business types are keeping track of the equipment and inventories, the lawyer types are trying to reduce the risk of errors and malpractice, the police types are trying to prevent drugs from being diverted. And what type is thinking about how all this is related to everything else? The median distance for a major interference was just under 10 inches.

There is an accompanying editorial in JAMA (2008;299(24):2898-289) by Don Berwick goes directly to the point:

From the particular case of RFID and EMI, therefore, emerge 2 important lessons. First, design in isolation is risky; even the most seductive technology will interact in the tightly coupled health care world in ways physicians and other members of the health care team had better understand, or they and their patients may pay a dear price. Second, no matter how good the design, in the end the battle for high safety and reliability in health care is never won. Safety is not a condition, it is a process. It can only emerge continually in a culture that is alert, cooperative, transparent, and resilient when the unexpected happens, as it always will. (JAMA)

As health care dollars become scarcer it becomes more and more likely there will be investment in keeping track of money rather than keeping track of bad things that happen to patients. It’s clear where the incentives and disincentives are in this. Certainly not in things that make the hospital culture more “alert, cooperative, transparent, and resilient when the unexpected happens.”

Comments

#1 caia
June 29, 2008

Scary. Any ideas as to why the passive devices were more likely to cause interference than the active ones? (I’m assuming the interference occurred when they were read by scanner.)
#2 Edmund
June 29, 2008

I wonder what implications this has for humans who have elected to get RFID implants. If such a chip, which is supposed to store life-saving medical information, actually interfered with life-saving procedures in the hospital, then I guess the conspiracy theorists wouldn’t be too far off the mark with their talk of ‘the mark of the beast’.
#3 Kevin
June 29, 2008

Hazardous incidents included: total switch-off and change in set ventilation rate of mechanical ventilators; complete stoppage of syringe pumps; malfunction of external pacemakers; complete stoppage of renal replacement devices, and interference in the atrial and ventricular electrogram curve read by the pacemaker programmer.

I’m sorry, can someone explain WTF is going on with these medical devices. I don’t know terribly much about them, but I do know a bit about RFID, radio & electronics in general, and computers in particular. And this all sounds really woo to me.

Pacemaker: Okay, this interfaces with very sensitive electrical systems (the heart), and so EMI might be a problem here. I’d expect some shielding on the external parts to keep down interference, but okay fine, maybe RFID is a problem for them. But if it is, then won’t also cell phones, cordless phones, and pretty much every other piece of electronics around.

mechanical ventilators; syringe pumps: Are these things ultra sensitive or something? Google seems to think these things are just glorified miniature pumps with a computer control. You should be able to bury these things in RFID and have no problems whatsoever.

-Kevin
#4 DBH
June 29, 2008

I’m with you Kevin, this is a huge liability suit waiting to happen. Even my toaster goes through EMI compliance testing; you would think things like heart-lung machines would among other things be pretty much immune to any influence that wasn’t physical. RFID is a broad-spectrum device, but very low power. I hate to think of cellphone left on, wireless email, etc. if RFID can do this.
#5 revere
June 29, 2008

Kevin: All I know is what is in this (peer reviewed) paper and editorial in JAMA. Like anything else it could be wrong, but it certainly isn’t “woo.” Lots of things that knowledgeable people think “can’t” happen do happen. We figure out the explanation eventually. In this case I hope you are right. Incidentally, in all ICUs I am familiar with, you are told not to use cell phones because of possible EMI. Some of these effects are also likely frequency dependent.
#6 Kevin
June 29, 2008

Revere,

Oh, yes peer review, I know. I’m assuming by people in medicine, not engineering, though. Not that I have anything against medicine (my dad is an MD). But my engineer brain is having a real hard time with this, because EMI is just not as strong as all that. A lot of RFID tags can’t even be reliably read at more than 1 – 3 feet, and that is by a device specifically designed to pick up the right frequency. And many of these medical devices don’t appear to (need to be) all that sensitive. If they are really this needlessly sensitive to EMI, then every engineer building them needs to have their license revoked.

I looked through the article briefly, and a few things stood out. First, they are testing both the RFID chips *and* the RFID reader simultaneously. Even if hospitals went hog wild with using RFID, they would not necessarily be putting the reader in close proximity to sensitive devices.

Second, it wasn’t clear to me that it was blind (there was some mention of blinding of the incident classification, but I didn’t see anything about blinding of the actual data collection). It would have been easy, for example, to take an RFID tag and reader and rip out the electronics. (It also wasn’t clear how they actually did this data collection. I can imagine some guy holding the RFID thing and moving it in a circle about the device, while someone else holds up a ruler to get the distance right. But even so I don’t see how this matters — they should have been able to whack the pump over the head with the RFID reader with little effect).

I guess my main gripe is just the lack of explanation. It is like if an engineering journal published an article saying “we made cell phone calls in close proximity to various houses, and in 34 of 143 tests the occupants were cured of cancer.” Without some explanation of the mechanism, or more specific data (direct measurements of the amount of radiation, etc.), it just isn’t convincing to this skeptic.
#7 ebohlman
June 30, 2008

caia: “Passive” RFID tags actually derive their power from the reader’s signal, which means that a reader for them has to put out a more powerful signal. To everyone who seems increduluous, remember that it was the reader that caused the interference in all the cases. The mere presence of tags didn’t cause any harm.
#8 madder
June 30, 2008

@Kevin:

Yes, I too would like to know exactly what is happening as these devices interact. But there is a very important distinction between these observations and woo. If this were woo, we’d hear lots of stuff about how RFID tags “could” do X, and “have been shown” (by unnamed others) to do Y, so therefore Z automatically occurs. Every time. And it either cures or causes cancer.

But this is not woo. This is observation of actual events. When RFID tags were placed near hospital apparatus, observable events occurred. Someone else could replicate or modify this experiment, and (gasp!) even falsify the authors’ conclusions. That makes this the opposite of woo. It is definitely a preliminary study in that it documented effects instead of systematically evaluating possible causes, but this is science.
#9 Calli Arcale
June 30, 2008

One issue that has bothered me for a while is that certain devices are exempted from various regulations for various reasons. These are primarily medical and military devices, for humanitarian and national security reasons, respectively. You’d think that medical devices would be held to a higher standard, since they have to pass FDA approval, but the FDA isn’t necessarily going to insist they be tested the same way the FCC would. The FDA is concerned with safety and efficacy — not electromagnetic interference. I would not be surprised if EMI testing has been widely overlooked. Some manufacturers will have thought of it; others will not.
#10 Luna_the_cat
July 1, 2008

Speaking as someone with no engineering qualifications whatsoever, but a fair amount of experience of computers, my guess about this would be that the issue is not so much with the pumps/syringes themselves, as with the chips or printed circuit boards which run them. Almost everything these days has one or more dedicated microchips attached, generally with a limited embedded function. And usually, these chips are made as physically small and inexpensive as is absolutely possible — and this means that the silicon channels are very small and very close together, and extremely sensitive to channel-to-channel interference, AND outside EMI which creates channel noise. As far as the chip is concerned, this equals confused/uninterpretable instructions, and results in very flaky behavior. An EMI field in close proximity would (and demonstrably does, testing has been part of the development of the IEC 60601-1 and specific ISO 14971 standards) bump up interference and channel bleed a good few levels.

Calli Arcale, as far as I am aware medical equipment is supposed to comply with the ISO 14971 standard I mentioned above, or the CISPR 22/EMC Directive 89/336/EEC standards in the UK; however, the problem lies in testing the devices in a real-world situation where many different pieces of equipment operate in close proximity, as opposed to testing one or two pieces in isolation.
#11 Luna_the_cat
July 1, 2008

Oh, as a side note, Kevin — start here, this describes the basis of the problem:
http://www.edn.com/index.asp?layout=article&articleid=CA6515348
#12 Bender
July 1, 2008

Well I know almost nothing about RFID chips, however I do know a lot about medical devices. Specificly pacemakers, defibrillators, and electocardiograms. Each of these three devices need to be extra sensitive so that they can accurately track the electrical activity of the heart. There has been one occasion on specific that I can think of where a cell phone disrupted my view of the heart with the EKG/ECG. As for the pacemaker and defibrillators these two are implanted into the body and these patients can use cell phones all they want without fear of interference. This makes me skeptical as to why they said this could effect pacemakers, because I am sure many patients with pacemakers use RFID’s.
#13 Luna_the_cat
July 1, 2008

Bender, note that the paper mentioned only external pacemakers. I would think they mean products like this: http://www.apc-cardiovascular.co.uk/miniature.htm . These are common for temporary pacing.
#14 AnnieRN
July 1, 2008

I wonder if the relatively small size of the average ICU patient room could contribute to the “overlap” in signals. Amazing findings. I know in my own experience that my cell phone has interfered with neurofeedback readings. Thanks for passing this on. AnnieRN