According to an article in the scientific journal The Lancet, a group of scientists has proven the usefulness of wearables to improve public health, since the information obtained through these devices can be compared with a series of characteristics that indicate a healthy state.
As more and more people are using this type of technology, scientists believe that the digitalization of healthcare can especially help predict disease and infectious outbreaksThis is undoubtedly a challenge that, as the current Covid-19 pandemic has demonstrated, must be addressed globally.
However, there are factors that must be addressed and evaluated so that the data stored by wearables can be used for the benefit of individual and collective health, which we will discuss below.
<< Apps and wearables as distance sensors, the key to prevent contagion in companies >>
Predictive models with wearables
The potential of wearables in healthcare depends on the development of predictive models of
machine learning
so that the data collected follows clear rules and can be monitored and replicated in hospitals and clinics.
In fact, for these models to be reliable they have to overcome these biases:
- Bias in the prediction domain. When those responsible for these processes do not master the operation of machine learning
- Bias in outcome measurement. When the data are measured subjectively, without considering exceptions to the rule
- Bias in the scope of analysis. When the sample size is small and the training, validation and test data sets are separate
- Participant dominance bias. When the participants enrolled are not representative of the general population
Regarding the last bias, it is important for technologists and physicians responsible for innovative health care programs to recognize that the spectrum of physiological parameters considered as normal is only a reference.
Therefore, two things are needed at the same time: wide population ranges, in order to have a solid data base; and specific information on each individual, in order to know the fluctuations in his or her health.
Therein lies the difficulty of reproducing predictive models of diseases and illnesses on a global scale, because there are not only differences between individuals, but also between ethnic groups, communities and societies.
Wearables usability
The wearables available on the market that can be used for medical purposes are varied:
- Smartwatch
- Bracelet
- Ring
- Headset
- Adhesive patch
- Garments
Each of them offers different advantages and disadvantages, so the focus should be on the reliability of the information they collect, so that both health systems and users trust them and use them in favor of health.
Currently, the sensors in these handheld devices can track:
- Physical activity
- The position of the body
- Heart rate and rhythm
- Temperature
- Oxygen saturation
- Variation in skin properties
Another challenge for the identification of a given disease through wearables is for people to be consistent in their use and then explicitly report any abnormalities they observe in their regular health status.
It is worth mentioning that the information provided by wearables should not be considered clinically reliable unless there are adequate validation tests, which ideally should be performed by the health authorities of each community.
There are also some scientists who recommend the use of non-portable sensors for longitudinal studies (analyses that are applied to specific groups of people repeatedly over several years).
These technological tools are not placed on the body but in some space of the home, and through their sensors they can detect the vital signs and other physical aspects of the people living there.
Scientific support for the use of wearables in health care
When data from a population’s wearable devices are pooled, it is possible to track viral activity in real time, as well as improve disease predictions.
For example, Kinsa smart thermometers, used in the United States, have been effective in predicting the activity of respiratory diseases such as COVID-19.
Data from wearables also help detect trends, form an outbreak calendar and identify geographic foci of infection.
This complements traditional clinical and laboratory surveillance, as well as other innovative metrics such as the rise of certain health or disease terms in internet searches and geolocation and mobility data.
For each new metric added to wearable sensors, both companies and institutions need to investigate what can be learned from the measured aspect; that is, analyze why upward and downward changes occur, and then assess whether they may be associated with a disease.
In addition, in the event of an epidemic or pandemic, monitoring via wearables would be useful to identify the response of immune systems to drugs or vaccines, including their efficacy and side effects.
Uncertain future in the use of wearables in health care
If any benefit can be drawn from the current pandemic context, it is the series of technological advances that have been made in health care, especially for the care of vulnerable groups and physically distant populations.
However, the historical lack of contextualized methods has contributed to the low adoption and scalability of technological innovations in healthcare.
The Internet is not accessible to 40% of people living in emerging economies, and close to one billion people do not have electricity service. and close to one billion people do not have electricity service.
Therefore, companies and institutions looking to increase the number of wearable users need to work collectively to close the gaps in these areas:
- Racial prejudice
- Non-representative databases
- Economic inequality
- Digital illiteracy
- Access to and quality of health services
- Collection, use and protection of personal information
To this end , it is vital that the technological infrastructure reaches all corners of the world. Data from wearables and other devices can then be leveraged to minimize current disparities and ensure public health services.
As wearables continue to evolve, collaboration and sharing of data and predictive models between scientists and technologists will have to be strengthened, first within a territory and then internationally.
In short, more critical analysis is required to offer wearables that do support individual and public health needs, while creating a better user experience.
Does your company use wearables to measure any physical aspect of customers? Have you ever used a wearable? Do you think that if its use becomes widespread it may have more benefits than harm in terms of public health?
Comment in the space below and subscribe to my blog to learn more about machine learning for business, as well as other topics of innovation and scientific technology applied to business.
Originally published in Jorge Pérez Colin Blog
