Clinically, medical devices do amazing things–they monitor vital signs (which, as the term indicates, can have life-or-death implications), deliver care, and measure health in the form of fitness devices. But technologically, medical devices fall way short–particularly in areas of interference, interoperability, and security.
The weaknesses of devices, their networks, and the settings where they reside came up over and over again in a joint workshop held by the FCC and FDA on March 31. I had a chance to hear most of it via live broadcast, a modern miracle of networking in itself.
Officially, the topic of the gathering was test beds for medical devices. Test beds are physical centers set up to mimic real-life environments in which devices are used, hosting large numbers of devices from different manufacturers running the popular software and protocols that they would employ out in the field. The workshop may have been an outgrowth of a 2012 report from an FCC mHealth Task Force which recommended “FCC should encourage and lend its expertise for the creation and implementation of wireless test beds.” (Goal 4.4, page 13) I thought the workshop had little new to offer on test beds, however, as the panelists concentrated on gaps between clinical needs and the current crop of devices and networks.
Medical settings are notoriously difficult places to employ technology. One panelist even referred to them as “hostile environments,” which I think is going a bit far. After all, other industries employ devices outdoors where temperatures drop below zero or rise precariously, or underwater, or even on battlefields (which actually are also medical settings).
I don’t dispute that medical networks present their own particular challenges. Hospitals crowd many devices into small spaces (one picture displayed at the workshop showed 15 wireless devices in a hospital room). Some last for decades, churning away while networks, environments, and requirements change around them. Walls and equipment may contain lead, blocking signals. Meanwhile, patient safety requires correct operation, resilience, and iron-clad security. Meanwhile, patients and their families expect access to a WiFi networks just like they get in the cafe down the street.
And yet Shawn Jackman, Director of Strategic Planning at Kaiser Permanente IT, said that problems are usually not in the infrastructure but in the devices. Let’s look at the main issue, interference (on which the panelists spent much more time than interoperability or security) and then at the ideas emerging from the workshop.
All the devices we associate with everyday network use (the IEEE 802.11 devices called WiFi) are all squeezed into two bands of the radio spectrum at 2.4 Gigahertz and 5 Gigahertz. When the inventors of WiFi told the world’s regulators that they had a new technology requiring a bandwidth in which to operate, freeing up existing bandwidth was hard to do, and the inventors were mere engineers, not powerful institutions such as the military or television broadcasters. So they resigned themselves to the use of the 2.4 and 5 Gigahertz bands, which were known as “junk spectrum” because all sorts of other equipment were allowed to emit radio-frequency noise in those bands.
Thus, because the bands are relatively narrow and are crowded with all sorts of radio emissions, interference is hard to avoid. But you don’t want to enter a patient’s room and find her comatose while a key monitor was unable to send out its signal.
Ironically, at the request of health IT companies, the FCC set aside two sets of spectrum for medical use, the Medical Device Radiocommunications Service (MedRadio) established in 1999 and the Wireless Medical Telemetry Service (WMTS) established in 2002. But these are almost completely ignored.
According to Shahid Shah, a medical device and software development expert, technologies that are dedicated to narrow markets such as health are crippled from the outset. They can’t benefit from the economies of scale enjoyed by mass market technologies, so they tend to be expensive, poorly designed, and locked in to their vendors. Just witness the market for electronic health records. So the medical profession found devices designed for the medical bands unsatisfactory and turned to devices that used the WiFi spectrum.
In 2010, by the way, the FCC relaxed its rules and permitted new devices to enter the little-used spectrum at the edges of television channels, known as white spaces, but commercial exploitation of the new spectrum is still in its infancy.
Furthermore, the FCC has freed up the enormous bandwidth used for decades to broadcast TV networks, by kicking off the stubborn users (known with respect as the “last grandmas”) who didn’t want to pay more for cable. An enormous stretch of deliciously long-range spectrum is theoretically available for public use–but the FCC won’t release it that way. Instead, they will sell it to other large corporations.
Networks are unreliable across the field. How often do you notice the wireless Internet go down at a conference? (It happened to me at a conference I attended the next day after the FCC/FDA workshop. At one conference, somebody even stole the hubs!) Further problems include network equipment of different ages that use slightly different protocols, which prove particularly troublesome when devices have to change location. (Think of wheeling a patient down the hall.) And you can’t just make sure everything is working the first time a device is deployed. Changes in the environment and surrounding equipment can lead to a communications failure that never turned up before, or that turned up and you thought you had fixed.
Medical device and wireless expert David Höglund claims that WLAN can work in a healthcare environment for medical devices. He lays out three overarching tasks that administrators must do for success:
They have to understand how each application works and its communications patterns: real-time delivery of small packets, batch delivery of large volumes of data, etc.
They have to provide the coverage required for each device or application. Is it used in the hallways, the patient rooms, the labs? How about the elevators on which patients are transported?
They need to obey the application’s quality service requirements. For instance, how long is a failure tolerable? For a device monitoring a patient’s heart in the ICU, a five-second interruption may be too long.
Medical devices and hospital networks need to be more robust and more secure than the average WiFi network. This calls for redundant equipment, separate networks for different purposes, and lots of testing. Hence the need for test beds, which many hospitals and conglomerates set up for themselves. Should the FCC create a national test bed? We’ll look at that in the next installment of this article.