There is a widespread consensus that wearable monitors will be a key part of delivering healthcare in the future. Bruce Keogh, the former CEO of the UK’s National Health Service (NHS), has said that the NHS envisages a major roll out of wearables as part of “a revolution in self-care”.1 In the US investment in digital health devices has increased from around $25m in 2011 to $312 in 2014, with a 125% increase in funding between 2013 and 2014.2 Technology giants who have previously shown little interest in the healthcare market such as Google, Microsoft, Apple and Samsung are all developing devices containing sensors, which can be used to track a wearer’s health.
Wearable monitors have been used as part of routine healthcare for decades. The first wireless monitor was developed in 1947, by Norman “Jeff” Holter, who transmitted a recording of his ECG, the heart’s electrical activity, via radio whilst cycling on a stationary bicycle. As a result of his pioneering work cardiac telemetry devices are now routinely used monitor the ECG (the heart’s electrical activity) of patients who are at high risk of developing cardiac problems. Similar devices exist for monitoring people out of hospital.
However, it is not until recently that that it has seemed feasible that they might be clinically useful for more a narrow group of, primarily diagnostic, applications. This has come about from developments in three key areas: hardware, data processing and analytical software and infrastructure.
The most visible change in the design of wearable monitors is their progressive miniaturisation. The monitor Holter used for his 1947 experiment was the size of a large hiking rucksack, weighed 38kg and needed to be strapped to his back. The telemetry monitors used in routine clinical practice are about the size of an old audiocassette Walkman and are worn in a belt holster or a bag slung around the neck. Cutting edge “digital patch” devices are smaller than a matchbox and can be worn unobtrusively under clothing for up to a fortnight without needing to be removed.
However, even the digital patches will seem bulky in comparison to the monitors of the future. Researchers have developed monitors that can be printed onto the skin, playfully disguising the electronic circuitry as a pirate tattoo.3 Currently these are only prototypes but they indicate the potential for monitoring to become truly unobtrusive.
Another approach to making monitors more subtle and easy to wear is through the use of smart garments. At the simpler end of the spectrum sensors can be embedded into clothing or held in special pockets. Many examples of this approach are available commercially, commonly for sports monitoring. A more advanced approach is to weave conductive and sensing elements into the fabric or design fabrics that have intrinsic properties that can be used to form sensors.
In parallel with monitors becoming smaller the range of sensors has increased, facilitating novel designs and making them more clinically useful. Optical sensors have enabled heart rate monitors to be incorporated into headphones. Blood pressure, traditionally measured using a bulky cuff worn around the upper arm can now be monitored throughout the day using a device that looks like a digital watch. Babies’ temperatures can now be continuously monitored using a small patch stuck on the side of their chest wall. Useful though today’s activity trackers are, many clinical applications will require measurement of the majority if not all of the five traditional vital signs – heart rate, breathing rate, blood pressure, temperature and blood oxygen levels.
Monitoring need not been done solely using wearable sensors. Other approaches are to collect data using sensors embedded into the environment, ambient monitoring, or combine environment and wearable sensors, pervasive monitoring. Currently these remain largely within the domain of research. Challenges remain in terms of making them scalable and ensuring that data is correctly attributed to the correct person when there are multiple individuals using the same space. However, it is likely that ambient and pervasive monitoring will develop into useful tools in the future.
Monitor innovation has not been confined to the sensor components of the electronics. Equally important is the standardisation and widespread adoption of wireless networking technologies. Historically, wireless monitors have relied on proprietary radio protocols to transmit data, limiting the use of these devices to areas where receivers have been installed. Newer generation devices can transmit data either via Wi-Fi or Bluetooth, vastly increasing the number of environments in which they can be used.
Particularly important is the development of Bluetooth Low Energy (also called Bluetooth 4), which improves on previous versions of the protocol by making device pairing easy and significantly reducing power consumption. Power conservation is one of the key reasons that motivate engineers to use proprietary radios or more obscure protocols such as Zigbee or ANT.
The adoption of standardised wireless protocols has largely been driven by the rapid development of wireless infrastructure. Wi-Fi and internet access have become more pervasive. Most NHS hospitals have institution-wide Wi-Fi networks. 77% of UK homes have broadband access.4,5 99.5% of UK premises have mobile internet reception via a 3G or 4G network. The existence of pervasive networking means that wearers can be monitored in real-time wherever they are.
Another transformative change, whose significance is yet to be fully realised is the rapid option of smartphones. 61% of the UK population own one.6 Most users carry their phone with them for large portions of the day. Many phones contain sensors that can be used for health monitoring without the aid of external devices. More importantly, through pairing with wearable monitors, smartphones can offload functions such as data storage, long-range data transmission, rich information display and complex computation onto a familiar and easy-to-use device. This enables the monitors to be smaller, cheaper and consume less power whilst increasing their usability and utility.
Recognising this potential the two major mobile operating system developers, Apple and Google have integrated the facility to store health-related information, including data from wireless monitors, into the iOS and Android operating systems. Cloud-based data aggregation platforms are also available from other manufacturers. Aggregation platforms, if widely adopted, add tremendous value as they reduce the amount of time that hardware developers need to spend on data storage solutions and they enable a rich ecosystem of third party applications to use data from wearable monitors without having to write code specific to each individual monitor.
Advanced software for processing and analysing the data from wearable monitors is the third piece of the puzzle that is required for wearables to transition from gadgets to a useful tool in healthcare delivery. Many of today’s health tracking applications simply allow entry and display of information. It is up to the user to draw their own conclusions, a potentially time-consuming and arduous process. For clinicians who may be caring for thousands of patients in their community, an unprocessed deluge of data is unlikely to facilitate better insights into the health of those whom they serve. In a hospital setting, years of experience with continuous monitoring of patients using traditional monitors, has shown that alerts based on simple thresholds generate a high number of false alerts. This results in a phenomenon known as alarm fatigue, whereby staff stop responding to the alerts.
At the simplest level data from the sensors can be fed into a rule-based Clinical Decision Support System. Combinations of changes in multiple different parameters have the potential to significantly reduce false alert rates.7
An alternative approach is to use machine learning to identify patterns that are linked to deterioration. These may allow discovering of subtle insights, hitherto unrecognised by clinicians. An example of the success of such an approach is the C-Path algorithm,8 developed to analyse microscope slides of tissue taken from patients with breast cancer, which was recently accurately able to identify markers associated with increased survival. In doing so three, previously unknown characteristics of the tissue surrounding the cancer were discovered to be important.
Despite such advances there has not been significant clinical uptake outside of diagnostic monitoring. Evidence regarding the effectiveness of telemonitoring is mixed with more positive results in some conditions9 than others.10 In these trials remote monitoring forms only part of an overall package of intervention and it is difficult to separate out the effect of the monitoring from the effect of the remainder of the intervention. It is also worth noting that in the majority of trials monitoring consisted of spot checks of vital signs rather than continuous recording using wearable monitors. For wearable monitors to gain greater adoption as part of routine healthcare a number of hurdles must be overcome.
The simpler of these are the technical challenges. Successful monitoring of athletes, astronaut, military personnel and fire fighters, to name a few does not automatically translate into tools for monitoring hospital patients. There is a large difference in the ergonomic refinement that is sufficient to measure a motivated individual operating for a short period in a high stress environment and the refinement needed to monitor an anxious, unwell patient over a long period of time. In one hospital trial nearly 60% of patients asked for the wearable monitor to be removed within 48 hours of starting to wear it.11 Considerable improvements are needed to make monitors capable of monitoring multiple vital signs feasible for long-term use.
Far more complex are the social, cultural and organisational changes required for the wearable monitoring systems to be accepted by patients and clinicians. For any such system to be beneficial, the end user has to be convinced of its benefit. For the clinical community to be convinced the technology needs to advance to a reliable level and the response mechanisms need to be optimised so as to not introduce extra workload, cost and potentially new risks to the currently used healthcare mechanisms. For the patient to be convinced, systems must be simple to use and not require technical competence; they must be designed such that the user’s identity is not threatened: usability, aesthetics and utility concerns must be addressed, placing the patient in the centre of the design approach.
Finally cost effectiveness must be addressed. The majority of patients pass through hospital uneventfully being monitored intermittently every 6–12 hours. A review of the benefits of cardiac telemetry concluded that indiscriminate use of cardiac monitoring yielded little benefit.13 To motivate a change in practice, the costs and inconvenience of continuous monitoring must be brought so low that it is worth monitoring everyone to catch the few that might otherwise slip through the net.
Similar conclusions apply in the outpatient setting. An economic analysis carried out as part of the 2008 Whole System Demonstrator trial, the largest trial of telehealth in the world, concluded that telemonitoring was not a cost effective intervention and that care needed to be given to how such interventions should be funded because while the monitoring services may be run by the primary care services, the limited benefits observed were in reduction of the utilisation of secondary care resources.
Until the costs of monitoring become negligible or analytic tools are developed which are able to deliver significant clinical benefits if widespread continuous monitoring is available, further research is needed to clearly identify the subsets of patients who might benefit from continuous monitoring.
However, it may be that costs do not have to be entirely borne by the healthcare system. With prices falling and direct marketing of wireless monitors to consumers there has been a democratisation of the ability of individuals to measure their own physiology, further empowering them to take control of their own healthcare. Through careful facilitation of this trend it may be that significant benefits can be realised.
At the same time the potential benefits of these changes should not be over-estimated. There is a significant difference between those who are most likely to buy wearable monitors – young, affluent individuals – and those with greatest healthcare needs, the poor and the elderly. One size does not fit all. It may be that the initial beneficiaries are those best suited to being early adopters but care must be taken to develop monitors and packages of intervention that are appropriate for the needs of a wide variety of users, with an emphasis on those with the greatest need.
- Kim D-H et al. Epidermal electronics. Science. 2011;333(6044):838–43.
- Ofcom. http://media.ofcom.org.uk/facts/.
- Mobile Operators Association. http://www.mobilemastinfo.com/stats-and-facts/.
- Ofcom. http://media.ofcom.org.uk/facts/.
- Tarassenko L, Hann A, Young D. Integrated monitoring and analysis for early warning of patient deterioration. Br J Anaesth 2007;98(1):149–152.
- Beck AH et al. Systematic analysis of breast cancer morphology uncovers stromal features associated with survival. Sci Transl Med 2011;3(108):108ra113.
- Inglis SC et al. Which components of heart failure programmes are effective? A systematic review and meta-analysis of the outcomes of structured telephone support or telemonitoring as the primary component of chronic heart failure management in 8323 patients: Abridged Cochrane Review. Eur J Heart Fail 2011;13(9):1028–40.
- Bolton CE et al. Insufficient evidence of benefit: a systematic review of home telemonitoring for COPD. J Eval Clin Pract 2010;17(6):1216–22.
- Jeffs E et al. Wearable monitors for patients following discharge from an Intensive Care Unit; practical lessons learnt from an observational study. 2015. Submitted for publication.
- Sanders C et al. Exploring barriers to participation and adoption of telehealth and telecare within the Whole System Demonstrator trial: a qualitative study. BMC Health Services Research 2012;12(220), doi:10.1186/1472-6963-12-220.
- Larson TS, Brady WJ. Electrocardiographic monitoring in the hospitalized patient: a diagnostic intervention of uncertain clinical impact. Am J of Emerg Med 2008;26(9):1047–55.