What IoMT Really Stands For

The basic goals of engineering include the achievement of a product’s purpose, safety, cost, manufacturability, and supportability, among other things. For internet of things (IoT) applications, much of the essential purpose relates to wireless communications that untether communication from wires and cables. This is especially true of the rapidly growing internet of medical things (IoMT), where connected medical devices enable applications from inexpensive wearables and sensors to remotely controlled surgical robots and imaging systems.

Although the IoMT is part of the IoT, testing medical devices often goes beyond the tests required for the IoT. The additional testing and key performance indicators (KPIs) required to validate and de-risk a connected medical device are so significant that IoMT could also mean “Internet of More Testing.” Let us consider a few examples of performance parameters that are measured for IoT devices and how they apply to connected medical devices.

Connect time is the amount of time it takes for an IoT device to connect to a central device, coordinator, access point, or another node in a mesh network. While connecting quickly is always good, in the IoMT, safety and life may depend on fast connections, especially in emergency situations. Therefore, connect time should be tested in challenging contexts, including weak signal strength, multiple devices connecting simultaneously, and interfering transmitters with different protocols operating in the same spectral band.

Latency is the time it takes for an IoT device to respond to a request, command, or sensed stimulus. Again, low latency is always desirable, but in surgical robots, every millisecond counts. Imagine, for example, the damage that a bone saw running even a tenth of a second too long could cause. For this reason, it is not enough to ensure that a medical device has a low average latency; the distribution of the latency measurements must be very tight. Be sure to include some measure of variability, such as standard deviation or variance, as one of your latency KPIs.

Throughput is the rate at which data goes from the device to its destination (or vice versa), usually specified in bits per second. For most IoT devices and even many connected medical devices, the amount of data transmitted is small, and the throughput is not a concern. But for real-time imaging, especially remote surgical applications, throughput is critical. Another concern is the device’s behavior when some unexpected situation or behavior limits throughput. For example, suppose a camera in a surgical operation can only stream data at half its normal rate. Do you accept that situation and allow your supposedly real-time system to get further and further behind in conveying information to the surgeon? Perhaps a better solution would be to implement some sort of compression algorithm, such as quantization, spatial downsampling, or another technique that compresses the data. Yes, the image quality suffers, which is not ideal, but it may be better than a perfect image that is several seconds behind reality.

Roaming behavior is a collection of performance indicators that characterize the device’s behavior when it switches between central devices or access points, analogous to how a cell phone switches between cell towers when the phone is in a car traveling down a highway. Some key aspects of roaming include the frequency at which handovers must occur, the time it takes for a successful handover to occur, and any data lost.

Many connected medical devices do not need to roam, as they are in a fixed position or connected to a cell phone that manages the roaming. However, with the increasing use of mobile, wearable, and implanted devices, the need for handovers from a home to an ambulance or an ambulance to a hospital gurney increases. Within a hospital, a device such as an infusion pump or signal monitor might experience multiple roaming handovers as the patient moves between various wings or floors of the hospital.

Dropouts are brief interruptions in data transmissions. In a streaming broadcast of a game or concert, a brief dropout may be annoying, but there is no need to re-broadcast the affected portion of the stream. For connected medical devices, a dropout in the transmission of a large data stream or series of files may require complex recovery behavior, and in a real-time surgery, a dropout may be fatal. Therefore, it is not enough to simply measure the dropout frequency. You must also validate that the device appropriately buffers, backs up, and re-transmits the data.

In summary, much of the value of connected medical devices depends on reliable connectivity. But realizing that value requires extensive testing and strong KPIs to deliver value to medical professionals and de-risk the patient experience.