Health and fitness wearables: past, present, and future
Health and fitness wearables have evolved from simple pedometers (step counters) that track basic levels of activity and fitness to complete health monitoring solutions that can be used to monitor all the vital signs of the human body in a variety of ways.
In this article, Mark Patrick from Mouser Electronics traces that evolution by reviewing the different sensors that make these features possible before looking at some emerging applications for healthcare sensor technology.
People have always been interested in knowing how many steps they have taken over the course of a day – indeed, Leonardo da Vinci is known to have sketched out a concept step-counting device. The credit for creating the first prototype pedometer goes to the horologist and inventor Abraham-Louis Perrelet in 1770. The first modern version, the Manpo-Kei, was created by Dr. Yoshiro Hatano in 1965 whose aim was to encourage people to walk 10,000 steps per day. Early pedometers were worn in a strap around a person’s waist and had a mechanical switch to detect motion (implemented using a lead ball that moved back and forth or a pendulum that hit a stop contact) in combination with a counter. In 1921, the polygraph (more commonly known as the lie-detector test) was invented. This was the first device to have sensors that could measure heart rate, blood pressure and galvanic skin response (GSR).
While the purpose of the polygraph was quite different and it was not deigned to be mobile, it could perform many of the functions included in today’s health and fitness wearables. In the 1980s, the accelerometer was first applied to activity tracking and multi-axis accelerometers are commonly found in today’s wrist-worn devices, allowing wearers to track the number of steps taken, calories burned, and the amount of time spent sleeping.
In 1977 Finnish professor Seppo Säynäjäkangas, an avid skiing enthusiast, founded a company called Polar arising from his interested in using electronic technology to measure heart rate. In 1979 he filed a patent for a fingertip heart rate monitor followed in 1982 by the first wireless wearable heart rate monitor, the Polar Sport Tester PE2000. This consisted of a chest patch radio transmitter, with an integrated electrocardiogram (ECG), and a separate wristwatch receiver that displayed real-time data about the wearer’s heart rate. This was a game changer for the personal fitness arena and played a large role in increasing the popularity of high-intensity interval training. Early wearables devices were intended to provide indicators about overall activity and fitness levels but could not be used for medical purposes.
Figure 1. Polar Sport Tester PE2000 (Source:https://www.wareable.com/fitness-trackers/the-origins-of-the-fitness-tracker-1234)
Addition of vital signs sensors
Present day wearables are vastly different when compared to their modest predecessors. Apart from step-counting, calorie-counting and sleep tracking they can now also allow provide users with insights into other human physiological vital signs (blood pressure, body temperature, and blood oxygen saturation) at their choice of different body locations. While the wrist and chest are still the most common, ear-worn devices are becoming increasingly common (so-called hearables). This ability to measure vital signs has been made possible by the inclusion of a range of advanced sensors including:
Optical: These sensors, like the AFE4420YZR from Texas Instruments, are used to perform photoplethysmography (PPG), a simple and low-cost optical technique that detects blood volume changes in skin tissue. Colour or sometimes infrared LEDs are used to illuminate the skin of the device wearer.
The reflected light is detected using photodiodes which convert the light signal into an electrical current which is in turn sent to a microcontroller that calculates heart rate and blood oxygen saturation. Advanced optical sensors use algorithms to compensate for the motion of the wearer and changing light conditions that could adversely affect readings.
Figure 2. Wrist-based health sensor platform featuring the MAX32664 biometric hub (Source: https://eu.mouser.com/new/maxim-integrated/maxim-maxrefdes103-health-sensor/)
Biopotential: These sensors, like the MAX30001 from Analog Devices, detect tiny electrical signals in the human body and can be used to produce an electrocardiogram (ECG) – a graphical display of cardiac behaviour.
Temperature: While general purpose temperature sensors have varying levels of accuracy depending on their application, clinical grade body temperature sensors like Texas Instruments’ TMP117AIDRVR must have an accuracy better than 0.2°C
Until recently, the quality of the sensors has precluded their use in medical applications. This was not a major problem for consumers who used health and fitness wearables to provide them with general guidelines about their health and wellbeing without paying too much attention to the accuracy of the readings from their device. However, many manufacturers are now producing medical-grade sensors that designed for use in devices intended for professional healthcare applications once they have attained the relevant certification from the appropriate regulatory body (like the Food and Drug Administration).
Medical grade sensors can also assist with the development of more portable and easier to wear devices. For example, the Holter is currently the most used heart monitor by healthcare practitioners. However, it is bulky which makes it impractical to wear more than 24 to 48 hours. While this may be sufficient to capture some aspects of cardiac behaviour, in some cases, it is not long enough to fully detect the presence of life-threatening conditions like arrhythmias. Smaller, less invasive wearable heart monitors will make it easier to detect these conditions by making it possible for healthcare practitioners to monitor patients remotely for longer periods of time with no need to return their sensing device to a clinic for data to be uploaded and analysed.
Connecting medical grade devices worn by countless people to the Cloud and continuously uploading huge volumes of data will also open the possibility of applying machine learning techniques to uncover hidden trends and patterns in this trove of information. This will in turn assist with the development of algorithms which further improve the accuracy of sensor readings. For example, the MAX32664 by Analog Devices is a low-power biometric sensor hub family which can communicate directly with several different sensors to analyse readings in different use cases. An interesting future application which may be possible using this sensor hub will be the measurement of blood pressure using PPG techniques on the fingertip (and possibly other body parts) of a device wearer. Such a development would be welcome as it would possibly replace the current requirement to wear a bulky mechanical cuff, which is not practical for permanent wear.
Recently, wearable healthcare applications have been developed based on the use of electrochemical sensors like the MAX30134 analogue front end which is now being used in blood glucose monitors to continuously monitor blood sugar level in diabetes patients. Readings are sent wirelessly to a patient’s phone to alert them if they need to administer a dose of insulin to themselves. Future developments may see the monitoring device connect wirelessly to a wearable insulin pump which then automatically dispenses the appropriate dose to the patient.
The evolution of electronic wearable devices continues to happen at pace. While they will always be popular as consumable devices which allow individuals and athletes to monitor their general health and wellness, sensor accuracy has improved to the stage where they can now be used in clinical grade devices and equipment. This will be transformative in how healthcare is delivered in the future, allowing clinicians to remotely gather information and monitor the health of their patients without the need to see them in person as often, potentially making healthcare delivery more efficient and effective.