Wearable technology and the IoMT

The Internet of Medical Things (IoMT) is flourishing, but will it help or hinder health professionals in keeping us fit and well?

The jogger who ran past you this morning was more than likely being followed. Not by some nefarious menace, but rather a digital device, which might just be motivating the wearer to beat an elusive personal best.

Wearable connected devices are becoming ubiquitous, and the popularity of fitness trackers is a major driver of this growth. Research conducted by Parks Associates suggests that more than 78 million wearables were sold worldwide in 2015; by 2019, the fitness tracker market is set to top $5bn.

Fitbit, Jawbone, Xiaomi and Garmin have all become household names thanks to an ever-increasing thirst for knowledge about our own physical performance. The range of measurable variables is already astounding, from basic metrics like the route and speed of your travel, to heart rate, respiratory rate, skin temperature, ECG (heart activity) and even posture during activities. And that’s just the start; many new devices are looking at more detailed information such as pulse and breathing variability, mood and EEG (brain activity).

Despite the medical terminology, it’s worth observing that these fitness trackers are still consumer products. They may inspire design trends and provide fun ways to follow sporting progress, but their rightful place is in the gym or on the field of play, not the GP’s office. However, a certified medical Internet of Things – known as the Internet of Medical Things (IoMT) – may soon create a step change in the quality that healthcare professionals can offer. The directions each sector takes in 2017 may well determine whether the two eventually converge.

One clear trend on the consumer side is a movement away from wristbands and towards earphones. It seems that very soon, many of us will have data coming out of our ears in a very literal sense.

Leading this shift is the Dash, the first truly wireless earphones, which set a high benchmark in 2016 as one of the world’s most powerful wearable microcomputers. The ear buds boast 27 unique sensors that can measure vital signs while playing music and augmenting the user’s incoming and outgoing communications. And if you think squeezing so much technology into an earbud is impressive, Nikolaj Hviid, founder of parent company Bragi, thinks much more is yet to come. “The number of sensors in wearable devices is set to increase rapidly – before the end of the decade they will contain hundreds,” he says.

While stressing the product’s commercial nature, Hviid has no doubt in-ear monitoring technology will have a huge impact on healthcare. “The way people can be helped or enabled to do something that wasn’t previously possible is mind-blowing,” he says. “From assistance for Alzheimer’s patients to understanding complex knee injuries, wearable computing will transform how we understand pharmaceuticals, rehabilitation and preventative care.” It’s not hard to see how tech companies like Bragi are responding to customer demand now in ways that will lay the foundations for doctors to meet patients’ needs further down the line.

While all-in-one lifestyle devices such as the Dash pack in ever more sensors, Collette Johnson, director of medical and healthcare at Plextek, believes medical devices used specifically for monitoring purposes may move in the opposite direction. “Lots of sensors are useful for trend data, but when looking at disease prevention a narrower range of indicators may be preferable,” she says. “Product developers are also getting better at asking ‘why am I measuring this?’ I think we’re likely to see the emergence of more discrete devices that exploit the innovative form factors available on the consumer wearables market, such as OMsignal’s smart bra.

Historically, the medical community hasn’t considered commercial wearables useful for research or treatment because they typically display unreproducible results on a design-focused interface rather than providing clinically valuable raw sensor data that can be rigorously tested under experimental conditions. Startlingly, some comparisons between fitness wearables have shown large variations in accuracy between devices, with error margins of up to 25 per cent. Reliability is an issue if consumer devices are ever going to cross over into consideration for any medical applications. In fact, the one clear negative impact wearables may pose for GPs is a misunderstanding by patients regarding the reliability and potential for self-diagnosis with these types of device.

Nevertheless, commercial innovation will certainly benefit healthcare professionals long-term, and acceptance of this fact is growing. For example, well-designed devices and associated front-end portals can prove very effective at steering user behaviour. Wearable manufacturers commonly use social influence strategies alongside other persuasive techniques, including gamifying activities with competitions and challenges, or reinforcing positive behaviours in the form of virtual rewards. These strategies are all designed to increase user engagement. Employing similar approaches could prove incredibly useful for addressing two major areas for healthcare professionals: adherence and compliance.

Typically, patients forget to take medication, skip doses or generally fail to follow doctor’s orders in myriad ways. This can extend recovery time and may be result in re-admittance to the hospital. Re-admittances are a severe burden on healthcare systems, and hospitals work hard to avoid them. Nurses spend a great deal of their time taking readings from patients, but they can’t monitor people around the clock to see if they are following directions. However, wearables can. By tracking adherence and compliance, patients will have more ownership of their health while the information at a GP’s disposal will improve, allowing them to make better judgements on treatment strategies. As with product design, IoMT is hanging on to the coat tails of preceding commercial successes.

Medical device development has traditionally been much slower to market than commercial efforts. Unfortunately, time-to-market is not the only thing that has lagged. User experience has often proved inferior to consumer products. But with innovations in rapid prototyping moving so quickly, the landscape is evolving rapidly.

Diabetes is one area of real potential for IoMT. Keeping blood glucose, blood pressure and blood fat levels under control greatly reduces the risk of developing complications, but that’s no easy task. One of many emerging solutions is the Dario All-in-One Smart Meter. It includes a simple glucose meter, disposable test strip cartridge and lancing apparatus that connects to your mobile device. The smart meter makes it easier to manage the condition, automatically logging results, and keeping patients connected to their caregivers; this means fewer interactions with healthcare professionals, and more valuable exchanges when these are required. Innovations in this space even extend to wearable organs. Pancreum Genesis has developed a wearable pancreas with smart insulin delivery for diabetes patients.

Not all devices have to be strictly ‘wearable’ to help with regular monitoring. One great example is the Breezhaler, developed by Novartis in partnership with Cambridge Consultants to overcome issues with drug delivery for chronic obstructive pulmonary disease patients. “One problem with using any drug delivery device is that under or overdosing can severely limit the benefit of using the device at all,” explains Chris Humby, programme manager at Cambridge Consultants. “One reason to add intelligence to the Breezhaler is to understand if the patient is using it correctly; the other is to ensure they’re adhering to the correct schedule.”

The inhaler employs ‘sound signature’ technology, based around an algorithm that listens to the distinctive rattle and whoosh sounds that indicate correct usage. With smart processing the predictive algorithms can understand the fingerprint of the rattle, identifying correct inhalation, if the capsule is empty, as well as information about the volume and rate of inhalation and other medically significant information. “One of the most exciting things we’ve found as engineers working on this project is that the algorithm is tiny and very ‘sleepy’,” says Humby, “meaning the sensor awakes just before use and processes data highly efficiently and sends it wirelessly to a hub or smartphone, before returning to sleep mode”. In fact, Humby is describing edge computing, and if doctors are to gain any actionable insights from real-time monitoring, edge computing is likely to provide part of the solution.

Deploying edge computing means only the most useful, actionable data is shipped to the cloud, freeing up a huge amount of capacity and improving efficiency. Without having to process vast swathes of mostly unusable data, product developers can use edge computing to only analyse the most meaningful information and apply the intelligence accrued to enhance device development, resulting in better quality of patient care in turn.

IBM Research identified the significance of edge computing to IoMT early. Its Cognitive Hypervisor project employs edge to tie several loose ends together, including general ongoing concerns relating to battery life, data security and standardisation. “Our objective is to improve the performance of IoMT devices by combining data streams from multiple IoMT devices in the Cognitive Hypervisor,” says Bruno Michel, a research scientist at IBM. “This means devices no longer require such substantial communication capabilities; instead they only need to communicate with the Hypervisor, which runs analytics to extract significant patterns from the data stream, improving patient-doctor interactions by relaying important information the patient may not even be aware of.”

In the eyes of many chief innovation officers, regulation will continue to be a major obstacle for IoMT, but not everyone agrees that this status quo should be accepted. Professor Jeremy Watson heads Petras, a UK-based Internet of Things research hub comprising nine leading universities which are spending three years exploring critical issues in privacy, ethics, trust, reliability, acceptability, and security. He is also President of the IET.

“The regulatory issues around IoT need to be clarified; we’re not talking about drugs trials remember – so it is possible to introduce devices that have standard levels of protection,” he says. “Procurement policies could therefore hold the key, with medics and healthcare trusts essentially maintaining the security of entire systems through their purchasing decisions.”

With IoMT innovations starting to converge, what will the sector look like in five years’ time? It’s hard not to see large tech companies such as IBM making significant contributions due to their high levels of investment in IoT and machine learning: “Our long-term goal is to create a context-specific patient model, that will allow us to understand how patients function when healthy and how they react to their external environment in day-to-day cases,” Michel explains. “We can then more precisely monitor how the patient is progressing towards an accurate baseline state – something that will prove highly valuable to doctors.” With teams at Apple and Google also working on major healthcare projects, the journey towards fully personalised health certainly looks to be getting closer.

The case for continuous monitoring in instances of chronic disease is very strong, but what ultimately happens when re-admittance figures drop, and savings are made by keeping patients in their homes? The initial major benefit may be to triage – the system by which doctors prioritise patients. Physicians will be better placed to treat more urgent cases, saving more lives in the process. However, in the longer term, we may see the fabric of the role itself change. Technological advances have allowed nurses to take on tasks traditionally conducted by higher-skilled physicians, but as machine learning starts to outstrip human diagnostic abilities, many doctors may very well start to more closely resemble data scientists and engineers.

Originally featured on E&T