Reducing noise to improve accuracy of medical devices

Nick Pearne, director of business development, TransSiP, explains one aspect that can make wearable devices more accurate. 

Patients and doctors are increasingly relying on wearable technology to help monitor health and issue alerts. However, these devices are still challenged with providing data that is accurate enough to be relied upon in this way. In most low-risk cases, they may be good enough but how can we ensure better accuracy to truly help in those more critical cases? Could the answer lie in improving the power supply and reducing frequency noise?

The potential of wearable devices

Technology has been improving steadily over recent years and we have seen a great deal of innovation when it comes to wearable medical devices. At the same time, the rise in consumer adoption of wearable fitness trackers that double as health monitors means that this market is booming further. A recent study from Insider Intelligence found that more than 80% of consumers are willing to wear fitness technology. Insider Intelligence  also anticipates there to be more than 84 million health and fitness app users by 2022.

Increasingly these wearables include some level of health tracking, such as insights into heart rate, blood pressure, and dissolved oxygen (SpO2). There are of course also some wearables purely focused on health monitoring. Most of these devices can export the data collected to send to health professionals and they can be key for helping users manage their personal health and thus reducing hospital visits. No wonder insurers and companies are looking into how supplying devices to consumers could be beneficial.

Industry data and forecasts vary, but the Global Medical Devices market is expected to grow from approximately $450 billion in 2021 to some $700 billion in 2027. While the potential market for these kinds of wearables is huge, it can only work if we can ensure they are accurate enough and don’t risk lives with bad or delayed data. 

The challenge of accuracy

Technology is rapidly evolving and wearable devices can do so much more than ever previously. Many devices can do all the above in just one device. And when the accuracy is there, they are even life-saving or life-changing. However, for them to be relied upon for life-saving information and alerts, 24/7 reliability and accuracy is imperative. While steady progress is being made with the addition of richer feature sets and functionality to the basic heart rate monitors of the past, there are several contrasting trends which are going to make life more difficult for device designers.

Wearables are low power devices that need to be small enough to fit on a person’s wrist or other part of their body, while remaining accurate wherever they are. They also need to communicate: the supporting apps are where most of the action is in data processing, parameter tracking, progress to goal, and so forth. All of that in a small device is a big ask, with more functions packed into tiny form factors and complex digital processing and radio transceivers forced to cohabit in a tiny and noisy space.

There has also been a great deal of focus on improving battery life, fuelled largely by consumer demand for devices to last longer between charges. The problem is that this is driving on-chip voltage reductions, which means that the signal to noise ratio is degraded, resulting in more processing errors. This is what makes it challenging to ensure accuracy as the ability of an electronic system to do something is determined by the cleanliness of the environment in which it operates. At the same time, there is a growing trend to harvest energy from the user, from body heat or motion, for example. This requires complex power management circuitry, which itself draws power, as well as adding further noise.

The bottom line is that current trends mean that existing power management technologies are going to constrain development of accurate, precise, reliable, and long-lasting medical devices. To make it even more challenging, most device manufacturers are not aware of why that accuracy is not there and don’t therefore know how to solve it.

How can we improve accuracy?

If we can make the power conversion/regulation process less noisy and more efficient, this will significantly improve device accuracy and maintain a high signal-to-noise ratio, improving overall accuracy and performance of the device. For the designer this would mean a reduction in power management complexity and faster and fewer design spins to get the product working as it should be. It also has the potential to increase the battery life significantly and ensure that these devices can use energy harvesting without having a detrimental impact on accuracy. 

At TransSiP, we achieve this using multidimensional noise management. In any electronic system, you can get several dimensions of noise: frequency and time domain noise, as well as switching noise jitter, i.e. random noise events. So, if you want to clean the noise on a device, you need to target all of those areas. The result is a significant and measurable improvement in device performance, particularly under conditions of problematic signal acquisition- for example a weak GNSS signal environment, or a heart rate measurement even when compromised by a drop in blood pressure. 

Thanks to a unique ultra-wideband 3-dimensional filter technology, wrist-worn optical heart rate monitors can achieve limits of agreement acceptable for clinical applications. We compared those limits with major devices on the market and in different conditions and to a chest strap HR device. We found that other devices were wide of the mark in accuracy, especially during activity, whereas a wrist-worn device using our technology stayed very close to the results produced by a chest strap, even during activity (see diagram). 

Wearable devices providing health, fitness, and clinical monitoring can be life-changing for many people, whether as stimulus for improvements/changes in lifestyle or early warning of developing medical conditions. However, this will only be possible if the devices have a whole new level of precision, reliability and accuracy. Field data has shown that new power integrity technologies can help designers achieve these goals.