Smart watches, rings, and a growing array of patches are adding more functionality and being used across a growing set of applications.
Doctors have been using advanced technology for years, but the growing trend is for consumers to use devices at home and have direct access to their data. Watches and rings that were once primarily used for counting steps or registering sleep patterns can now read blood pressure, heart rate, blood oxygen, body temperature, and other early signs of illness. Meanwhile, various patches are under development and gaining approval from the Federal Drug Administration to measure particaular health concerns.
“Recently, the FDA has been approving over-the-counter continuous glucose monitoring (CGM) patches,” said Shantanu Bhalerao, vice president of Bluetooth products at Infineon. “This means regulation is moving where more and more consumerization of health care will happen. There is a desire to put more data in the hands of the user.”
Others agree. “Medical devices are getting smarter for multiple types of applications, and there’s a trend toward more wearable devices, as well as home-use medical devices,” said Ryan Bauer, director, medical device and pharmaceutical solutions at Siemens Digital Industries Software. “Patient care is expanding beyond just the hospital and the doctor, and there are a couple challenges with both. As you miniaturize, the devices typically need to be wearable or transportable, so power is always a big concern. How long can they last? How do you recharge it? Where can it be used in the home? Then, if you’re moving into the implantable side of things, they’re not necessarily accessible. This might be implanted for a long time, and there are implantables that are non-powered sensors. There are devices that have powered sensors, like pacemakers, but they’re also moving into applications such as joint replacement. You can now get a smart knee that has sensors and batteries in order to relay information about your gait.”
A key driver for adoption is that medical professionals and patients want access to the information. “They expect that as they get into their medical care, they can have data and feedback and understand the performance and help that guide their therapy,” said Bauer.
Home-use medical devices can come in many forms, depending on what data is being captured. “There are multiple ways to skin the cat here,” said Prakash Madhvapathy, director of product marketing for Tensilica audio/voice DSPs at Cadence. “One is in the ear, which is a good place to catch signals, because it’s right near the cerebrum, and the signals may be stronger there. There is a lot of research going on in that direction, along with a smart watch or a patch that is applied to the human body. Near the heart is another place for such applications. It could also be that you have an AR [augmented reality] device, such as eyeglasses, that fit right over the ears. Sensors also could be at the tips of the glasses, looking for signals from the brain. The most unobtrusive ways are the smart glasses or the earbuds, because those are things that people wear every day, and they’re used to it.”
One of the main benefits of an unobtrusive device is that it can continuously monitor for anomalies. People may be willing to wear a more noticeable or cumbersome device once they know they have an issue, but not merely for early detection.
“The problem needs to be detected without the knowledge of the person,” said Madhvapathy. “That way you know a larger portion of the population is being monitored continuously and here the key word is continuous, because if you have to go to a clinic to get tested, that’s once a year, and they only do certain general tests. But if you have something that you’re wearing that is looking for not one but multiple symptoms of different types, then it would be useful to you in your daily life, and you will not have the stigma that you’re wearing something that people will look at you strangely and wonder if you have a problem.”
For example, a stigma around hearing aids has prevented many older people from adopting the technology, but that is unlikely to be the case with modern earbuds such as Apple’s AirPods with hearing health features.
“If the ear pods themselves do a certain monitoring without their knowledge, or even with their knowledge, nobody else needs to know about that, and the signal can be processed locally by a DSP with AI functionality,” said Madhvapathy. “In case there’s a problem, it can set off a very private alarm sent either to the phone or spoken in the ear to the user.”
Artificial intelligence, machine learning, and other tools have been used in the medical space for some time and applications are surfacing every day for better ways of filtering data to provide better insight than a human is capable of, such as screening for breast cancer. The FDA is also actively seeking ways to best manage the new tools.
“The FDA is holding meetings on AI and starting to provide guidance on how to incorporate AI into your medical devices,” said Siemens’ Bauer. “Some of it’s been around for a while. You can look at what they do with diagnostic imaging – reading those data sets and forming a model to do segmentation identification. But the applications are moving beyond that, and they’re moving to the edge as well, for medical devices. It’s definitely a hot topic there. There’s a lot of potential with AI for medical devices and the data sets are huge to pull from and learn.”
Devices and patches can gather data, but they also can create haptic sensations beyond vibration. For example, Infineon partnered with Theranica on a migraine patch.
“This is a different thing compared to a CGM, which is measuring something,” said Infineon’s Bhalerao. “You wear the migraine patch on your arm, and it does pain management without taking any medicine or drugs. It’s Bluetooth-connected and uses a Bluetooth MCU, and the basic idea is there are neural pathways and the ability to modulate the pain using sub-threshold pain signals. You use your phone to say how much migraine pain you’re feeling, and this device simulates some pain. You’re not actively feeling it, but it distracts your mind from the migraine pain.”
Devices and patches could therefore replace some medication. “I see more and more such things coming in the future, whether it’s detecting how much lactate you’ve built up, which is optical, or how you manage your nutrition as well as your exercise,” he said. “This medical patch area is going to be a very interesting space.”
Ansys also is exploring wearable optical sensors and helping designers address such challenges as compact system integration, precise optical path management, and real-world lighting conditions to ensure the wearables deliver reliable, actionable insights.
Other examples of recent wearable and med tech developments include:
Body and movement:
Blood and sweat:
Cancer:
Materials:
Fig. 1: “MouthIO” is a device with integrated sensors and actuators to capture health data and interact with a computer or phone. Photos courtesy of Sebastian Krog Knudsen via Aarhus University and MIT CSAIL.
Global shipments for wearables were expected to grow 6.1% in 2024, according to a December 2024 report by International Data Corporation (IDC). The smart watch market growth rate was expected to decline as the technology matured, while the hearables market was predicted to grow. Demand for smart rings and glasses is also expected to increase.
Chips, sensors, and signals
DSPs coupled with MEMS sensors are commonly used in wearable devices, along with other types of sensors, and/or electrodes, which are then connected to a data analytics platform via Bluetooth or Wi-Fi.
“For any sort of signals that you receive in a wearable environment, like a hearable device that can monitor signals in the ear, you’re basically now monitoring the brain’s activity from the ear. Those signals have a certain frequency, timing, amplitude, and characteristics,” said Cadence’s Madhvapathy. “For these, you will need a signal processor to be able to interpret those signals. You need to receive them, first of all, then interpret them in a way that the device can infer the health of the person. Once the brain signals are received in the ear to the MEMS or another sensor, you first amplify them, because the signals are very small, very tiny.”
Then, filtering the signals is essential to ensure that only the correct signals are being amplified, which is especially crucial when a device is monitoring for a health condition.
“That is hard to achieve, first of all, and you will end up also amplifying some amount of noise,” Madhvapathy said. “Along with the noise in the signal, you have to run it through another set of algorithms that can interpret those signals to say what the behavior of the signals is for a normal, healthy person, versus what is the behavior for an anomalous health condition that the person may be experiencing [such as Parkinson’s or Alzheimer’s]. The signals are very tiny, and the anomaly behavior represents itself in the signals in very subtle parts, so you have to now extract that anomaly.”
While early warning could be useful, a concern here is the chance for false alarms. “You must be very careful as to how you actually interpret those things, because you don’t want to go in there and find out five years later you had no problem,” he said.
Size is always a factor when it comes to wearables, and the smaller the better. “You want the device to not leak too much power when it’s sitting idle, and you also want it to have a very small area while providing all the DSP audio functions,” he said. Low power leakage is especially important for modern nodes such as 7nm. “Normally leakage is small, but if you are duty cycling it, and the leakage is still pretty large, then the duty cycle doesn’t help much, because by the time you come close to the next active cycle, you’ve lost a lot of charge already. That is not a good situation. Having a very low profile or low area DSP is very key to maintaining a longer battery life. At the same time, the design itself has to be properly adapted to any other power hogs, like the clock trees and the activity of the clocks and so on, are limited and turned off when certain blocks of the DSP are not active.”
One of the big challenges with wearables and small IoT devices is trading off the impact of processing data locally versus sending it somewhere else to be processed.
“We are getting into low-power processing solutions like disposable health care products,” said Satish Ganesan, senior vice president and general manager for Synaptics’ Intelligent Sensing Division. “If you have patches and things like that, you need to dispose of them in a certain time frame. The big question is whether you can apply all the things that you’ve already learned in terms of operating at low power and delivering the functionality you want and still make it feasible.”
Feasibility also includes security, particularly with medical devices. So in addition to extremely low power, data at rest and data in motion all need to be secure. “In the embedded processor space, we implement Trust Zone functionality for particular data that sits in there, and you have to add that on top of whatever data you’re sending,” said Ganesan. “And on the connectivity side, how do you make sure there’s no man-in-the-middle attack that’s going through?”
Safety, security, and biocompatibility
Medical devices face much more stringent testing than consumer devices, but regulators are starting to approve more technology.
“When you’re doing your device design, you go through a systems-level view of risk assessment, requirements, definition, and test management,” said Siemens’ Bauer. “In terms of electrodes, if you’re imparting any energy to the body, that has to be defined and analyzed from a risk standpoint — and if your device is receiving energy from the body, how that impacts the performance of the device. Capturing it both ways is super important in a very structured risk requirements test and in the verification/validation view of it.”
For wearable medical devices, biocompatibility is important. “There are different levels of biocompatibility depending if there is skin contact, in the body, a mucosal membrane, those types of things,” said Bauer. “There are a series of tests to show that the human body will accept that device for that use case. Biocompatibility is one aspect of safety, and anything that’s implantable or on the body is subject to that.”
Security is also a top concern. “From a security standpoint, there’s a big effort right now with the FDA to advance cyber security in medical devices,” he said, citing final guidance from September 2023. “Part of it is when you submit your device information to get cleared or approved, and they expect you to provide things like your software bill of materials and how that relates to vulnerability management, your plans for that, but also this idea of a secure product development framework. This is a systems expectation of the regulators and medical device companies to show that they have control of the inputs and outputs from a system level of the device, so that it’s not just looking at the electronics, not just looking at the software or the mechanical, it’s all of these things together. Your body’s a part of that system so you have to look at the inputs and outputs from the body as well. More complicated today is we have got connected devices, or edge devices. They may not just interact with the body, but also all these systems of systems around them.”
Materials and their interactions must also be taken into account and biocompatibility requirements limit what can be used. “If you put something in the body, you also have to consider things that can be extracted or leachable,” said Bauer. “Or if the device is contacting any of the medicine or food that can enter the body, that is a pathway too.”
Environmental degradation over time is another challenge. “Any oxidization that could occur, and whatever cleaning chemicals and things that are used – that’s a big source of medical device failures,” said Bauer. “They’re breaking down, and that can provide a pathway for accessing the internal components, so normally you do a series of risk assessments with a hazard analysis at the top.”
The impact of temperature is also a consideration. “With medical devices, you specify the lifetime that it’s designed for, and test to that,” said Bauer. “But with wearables and home-use devices, you need to consider if it’s going to stay attached to the body and within the body environment, or if people are going to leave it in a car for a period in Arizona or Fairbanks, for example. “The devices experience those extremes, so that feeds into the system view and the requirements you test for up front.”
Based on that risk analysis, mitigations are defined in the requirements. “What we call in the medical device space is the idea of design control,” he said. “It’s formally going through this process, like a systems engineering approach to the design. From there in the mitigations, you go into your testing to prove that. All of that documentation then gets supported and provided and reviewed for your higher-risk device — which most of your electronic devices are — to the regulatory authority. That way, you have a second set of eyes looking at what you did, and you’re showing them that you covered all the bases. They may come back with additional information requests. Then you’re working on meeting additional requirements or investigations on that before you can bring it to market.”
Whatever the purpose of the device, multiple teams need to use product lifecycle management tools to bring all the components and functionalities together.
“PLMs are the backbone for collaboration with all the different domain disciplines,” said Siemens’ Bauer. “This means the electrical, mechanical, biomedical, clinical teams and, on top of that, the digital twins for doing multi-physics simulation, or electronics and thermal all feeds back and forth in the background, so we can manage that in context across the whole development cycle.”
However, each company keep records in different ways, and this presents challenges when collaborating or during acquisitions. The biggest challenge is keeping the information in context across domains within the organization, said Bauer. Each domain needs to understand the product and have that knowledge as they apply it to their design.
Conclusion
Home-use medical devices come in many shapes and sizes, from watches to rings, glasses, and patches. Combined with the endless abilities of AI to process data, there will soon be a wearable to assist, augment, or monitor just about any bodily function. The old stigma of needing help is disappearing as populations age and more people embrace technology to optimize their daily lives or manage health concerns.
Further Reading:
 |
 |
 |
 |
 |
 |
 |
Leave a Reply