Tech Talk: Could Wearable Devices Warn Of Sickness?

Advertisement
Advertisement

A new report has been published by a panel of California-based biological experts detailing how wearable devices might be used to provide early warning signs of sickness. In the report, the panel investigated how portable biosensors could be used for frequent measurements of human physiological changes, and how the collected data might be processed and interpreted to provide an early warning mechanism. The panel used a volunteer team of up to 43 individuals for two years taking a number of measurements from up to seven worn biosensor units. The collected data was combined with an understanding of real world circumstances in order to build a database of up to over 250,000 daily measurements.

The source document details how the body shows physical signs that all is not well, for example the heart rate, blood pressure and body temperature provides a basic status report of health status. If any of these readings are high, this could be the cause for concern. However, current medical practice and technology means that measurements are typically taken upon a visit to a physician's office and for a healthy person, this could be every other year. As such, patients could be suffering from an underlying issue for a number of months before one or more changes in the body is detected. Another issue is that measurements taken at a regular appointment are compared with the overall population typical results, but we all have our own unique characteristics and these are themselves variable: a single isolated visit cannot identify a trend. The document elaborates that the information obtained from body measurements is especially important for caregivers, where the patient is perhaps less able to report on discomfort or other symptoms. In total, the experiment utilized wearable biosensors to measure the heart rate, skin temperature, blood oxygen levels, a number of activity parameters (including steps, activity type, calorie burn and acceleration), weight and total gamma and X-ray radiation exposure.

In recent years, this technology has both been shrunk in size and cost such that they are now affordable, highly portable and provide "a promising approach to routinely monitor personalized physiological measurements." Wearables are able to measure and deliver data in real time to either the patient or the physician. Several data streams may easily be combined to provide a thorough picture of activities, such as walking, running and sleeping. According to the report, as at July 2015 there were over 500 different healthcare wearable devices, three times the number available in 2013. The market has also grown such that it had over 34 million sales in 2015. The research team ran the project using a combination of one regularly monitored individual equipped with seven sensors, and many other subjects wearing two or more sensors. This technique allowed the team to assess accuracy and ease of data collection. In addition to the biosensors, volunteers collected additional data using a smartphone and one conclusion from the experiment that a combination of a smartphone, smartwatch and two sensors was enough to provide comprehensive data.

Advertisement

The research team concentrated on easily and accurately recording biosensor readings in either real time or at a high frequency. This data was then used to establish daily patterns and measure a "personalized baseline." This data was then used to determine the differences in health states between different kinds of individuals, the example given being those volunteers with diabetes and those without. Another objective of the report was to "detect inflammatory responses and assist in medical diagnosis at the early phase of disease development" so as to improve medical care. As a side project the team were also able to gauge the technical abilities of a number of different sensors. A number of devices used during the research had already been validated for clinical-grade accuracy but the team still "performed extensive testing to assess the accuracy of the different devices against gold standard measurements" and against their Welch Allyn [WA] 6000 series instrument, "routinely used at the clinical laboratory services at Stanford University." The instruments used were deliberately picked for accuracy and the team found this to mostly be the case; but when one instrument was found to be less accurate the trending information was still viable for the study.

Some of the observations detected included how the blood oxygen level decreased during a high altitude flight, which the researchers noticed was inversely correlated with altitude. The higher the aircraft flew, the lower the blood oxygen levels. However, the team also noticed that during long haul flights of seven hours or longer, the human body appeared able to adapt and increase the blood oxygen levels. During and after the experiment, the team are working on producing "a computational algorithm for personalized disease detection" based on the data. The researchers observed a period of elevated heart rate and skin temperatures leading up to and during a period of Lyme disease. Other participants also showed similar physiological changes immediately prior to reporting being ill. The team discovered that it was relatively easy to determine the physiological differences between individuals who are sensitive to insulin, and those who are not, and the data from this experiment could become a way to determine the risk of type 2 diabetes.

Ultimately, the panel agreed that the data collected by wearable biohealth sensors could be useful for medicine, and that wearable sensors "are likely to play an important role in managing health." As such, as wearable devices become less expensive and contain more sensors, so the logical conclusion from this is that we may start to see clinically approved biosensor-equipped smartwatches, although these may need to be supplemented with one or more additional sensors around the body. Furthermore, we may see manufacturers introducing smartwatches or activity trackers able to provide an indication of an impending illness some days before we might otherwise detect it ourselves.

Advertisement