Assigment, A+ Work, No plagarism, On time, Original Work

REVIEW

Challenges and recommendations for

wearable devices in digital health: Data

quality, interoperability, health equity,

fairness

Stefano CanaliID*, Viola Schiaffonati, Andrea Aliverti

Department of Electronics, Information and Bioengineering, Politecnico di Milano, Milan, Italy

* [email protected]

Abstract

AU : Pleaseconfirmthatallheadinglevelsarerepresentedcorrectly:Wearable devices are increasingly present in the health context, as tools for biomedical

research and clinical care. In this context, wearables are considered key tools for a more

digital, personalised, preventive medicine. At the same time, wearables have also been

associated with issues and risks, such as those connected to privacy and data sharing. Yet,

discussions in the literature have mostly focused on either technical or ethical consider-

ations, framing these as largely separate areas of discussion, and the contribution of wear-

ables to the collection, development, application of biomedical knowledge has only partially

been discussed. To fill in these gaps, in this article we provide an epistemic (knowledge-

related) overview of the main functions of wearable technology for health: monitoring,

screening, detection, and prediction. On this basis, we identify 4 areas of concern in the

application of wearables for these functions: data quality, balanced estimations, health

equity, and fairness. To move the field forward in an effective and beneficial direction, we

present recommendations for the 4 areas: local standards of quality, interoperability,

access, and representativity.

Introduction

Devices that can be worn on our bodies and track several activities and parameters—wearable devices—are increasingly sold and used in the general population. One of the main areas of use

of wearable devices is health, including biomedical research, clinical care, personal health prac-

tices and tracking, technology development, and engineering. In this context, the use of wear-

ables for health has been connected to several promises and benefits for a more digital,

personalised, preventive medicine [1–3]. At the same time, crucial work has identified and dis-

cussed technical and ethical challenges in the extended use of wearables for health, including

accuracy, privacy, security, cyber risks [4–7]. Yet, most analyses have focused on either of

these areas of discussions, thus framing technical and ethical considerations as largely separate

issues. As a result, the connections between specific technical solutions and ethical consider-

ations remain underdiscussed: This is a problem, as we will show that many challenges of the

PLOS DIGITAL HEALTH

PLOS Digital Health | https://doi.org/10.1371/journal.pdig.0000104 October 13, 2022 1 / 13

a1111111111

a1111111111

a1111111111

a1111111111

a1111111111

OPEN ACCESS

Citation: Canali S, Schiaffonati V, Aliverti A (2022)

Challenges and recommendations for wearable

devices in digital health: Data quality,

interoperability, health equity, fairness. PLOS Digit

Health 1(10): e0000104. https://doi.org/10.1371/

journal.pdig.0000104

Editor: Shelagh Mulvaney, Vanderbilt University,

UNITED STATES

Published: October 13, 2022

Copyright: © 2022 Canali et al. This is an open

access article distributed under the terms of the

Creative Commons Attribution License, which

permits unrestricted use, distribution, and

reproduction in any medium, provided the original

author and source are credited.

Funding: This research was funded by the

Fondazione Silvio Tronchetti Provera (Project: Etica

e tecnologie emergenti) and Regione Lombardia

(Project: BASE5G, ID 1155850). The funders had

no role in study design, data collection and

analysis, decision to publish, or preparation of the

manuscript.

Competing interests: The authors have declared

that no competing interests exist.

wearable context can be addressed only partially through technical solutions. In addition, the

epistemic (knowledge-related) contribution of wearables to the collection, development, appli-

cation of biomedical knowledge has only partially been discussed. As a result, there is a lack of

understanding of the specific uses and functions that wearables can and should fulfil for digital

health—yet this is crucial to identify the role of wearables in digital health and beyond and

assess their ethical and social impact in relation to specific uses.

In response to these considerations, in this article we start by providing an epistemic over-

view of the main functions of wearable technology for health: we discuss monitoring, screening,

detection, and prediction. The role of these functions is clear when looking at the use of wear-

ables in concrete cases (Table 1), for example, the context of Coronavirus Disease 2019

(COVID-19). The COVID-19 pandemic has been discussed as a crucial catalyst for the use of

wearable technologies in the biomedical and health domain and we will use it as a source of

uses and cases to illustrate our points throughout the article [8,9]. On the basis of this over-

view, we discuss specific issues and concerns that are connected to the use of wearables for the

identified functions of monitoring, screening, detection, and prediction. We focus on 4 main

areas of concern (data quality, balanced estimations, health equity, and fairness) and propose

recommendations and possible solutions (local standards of quality, interoperability, access,

and representativity). On the basis of our overview and analysis of these challenges, the recom-

mendations we propose enable us to better understand the actual impact, benefits, and risks of

wearables and improve their application for digital health (Table 2). In this way, as a group of

researchers with different areas of expertise (biomedical engineering and research, philosophy

and ethics of science and technology) but working in the same department, we develop an

interdisciplinary account of wearable technology and its contribution to digital health.

Table 1. The main functions served by wearables for health, with examples.

Functions Examples

Monitoring - Pulse monitoring [10]

- Advanced tele-monitoring [20]

- COVID-19 symptoms and long-term effects monitoring [21]

Screening - Atrial fibrillation screening [14]

- Sleep apnea screening [13]

- Cardiovascular disease screening [22]

Detection - Physical activity levels detection [23]

- Pre-symptomatic detection of COVID-19 infections [17]

- Seasonal influenza detection [24]

Prediction - Prediction of mortality and clinical risk [16]

- Prediction of COVID-19 infections [17]

- Prediction of exacerbations of chronic obstructive pulmonary disease [19]

https://doi.org/10.1371/journal.pdig.0000104.t001

Table 2. Summary of the identified areas of concern and key issues and proposed recommendations.

Areas of concern Key issues Recommendations References

Data quality - Variability of sensors, data collection practices

- Lack of contextual information

Local standards of data quality [27,29,34]

Balanced estimations - Overestimation

- Overprediction

Interoperability of wearable data [35,41,44]

Health equity - Unequal access to benefits

- Digital and technological divides

Access to wearable data and interpretation [28,50,52]

Fairness - Exclusion of portions of the general population

- Unfair wearable datasets

Representativity of wearable data [2,9,36]

https://doi.org/10.1371/journal.pdig.0000104.t002

PLOS DIGITAL HEALTH

PLOS Digital Health | https://doi.org/10.1371/journal.pdig.0000104 October 13, 2022 2 / 13

An overview of wearable technology for digital health

There is currently an abundance of uses of wearable devices for health. Focusing on the episte-

mic contribution of wearables to the collection, development, application of biomedical

knowledge, we develop an overview that looks at the current context of wearable technology.

While this is not a systematic representation of all possible or future applications of wearable

devices, we identify 4 main functions that wearables are currently used to serve in the health

context: monitoring, screening, detection, and prediction (Table 1).

We identify monitoring as the basic and fundamental function served by wearables, per-

formed by wristbands, patches, watches, clothing. Monitoring is the practice of continuous data

collection, focused on members of a population, which can be the general population or a spe-

cific subset of individuals. Wearables are considered particularly efficient to fulfil this function

because they can track a number of various biomedical processes depending on the types of sen-

sors available and can be used for continuous and remote monitoring—as wearables can be

worn constantly, they are ideally placed to collect data continuously. In this way, wearables can

deliver a significant improvement to remote and tele-monitoring [10,11] and have been used in

this sense to monitor crucial physiological metrics for COVID-19 such as heart rate, physical

activity, oxygen saturation, as well as long-term effects [8,12]. In this context, wearable devices

have also been applied in coordination with other tele-health systems for remote monitoring for

individuals at risk that could easily shift to hospitalisation and to assist remote diagnosis.

On the basis of these monitoring capabilities, we identify 3 other main functions that wear-

ables can serve. Screening is the identification of specific conditions and individuals associated

with this condition within datasets collected through monitoring. The use of wearables for this

function is usually based on passive sensors that measure motion, steps, light, pressure, sound,

etc. [3]. For example, wearable garments have been used to monitor individuals during sleep

and screen for individuals suffering from sleep apnea [13]. A close function related to screen-

ing is detection. When wearables monitor specific conditions in populations, they are often

used to detect conditions and alert individual users. Detection is the analysis of wearable data

collected through monitoring in order to investigate possible patterns and features that can be

interpreted as indicators and markers of specific biomedical conditions. For example, a combi-

nation of smartwatches and dedicated bands has been used for heart rate monitoring and auto-

matic detection of atrial fibrillation [14]. The integration of wearable data with symptom data

has been presented as a way to improve the identification of COVID-19 positive versus nega-

tive cases [15]. Detection is also the function where we see an intersection with both monitor-

ing and screening: for example, smartwatches have been used to monitor populations for

irregular pulse and, on this basis, screen for individuals potentially suffering from atrial fibril-

lation as well as identify the condition [10]. A final diagnosis of a condition can thus be based

on detections performed by wearables, although wearables currently cannot perform diagnosis

as a consequence to technical and regulatory limitations.

The fourth function we identify is prediction, the inference of future trends and/or events of

interest for the biomedical study of populations based on monitoring. Just a few wearable

devices are currently used for prediction in the health context, for example, to predict mortal-

ity, readmissions, and clinical risk [16]. In the context of COVID-19, wearables have been

tested for the retrospective detection of infection and prediction of COVID-19, days before the

presence of symptoms [17]. Other examples include the use of accelerometer data from wear-

able devices to predict biological age and mortality [18] and respiratory rate data to predict

exacerbations of chronic obstructive pulmonary disease [19].

These 4 functions are often intertwined and interconnected in concrete contexts and in

many cases the same device can perform more than 1 function. Still, identifying different

PLOS DIGITAL HEALTH

PLOS Digital Health | https://doi.org/10.1371/journal.pdig.0000104 October 13, 2022 3 / 13

functions is a crucial step to understand the actual impact and goals of using wearable technol-

ogy for health. Depending on whether we use wearables to predict or monitor health, different

assumptions, uses, and standards will be necessary. In addition, understanding which func-

tions wearables can and do serve currently helps us to make sense of their possible limitations.

As we will see in the next sections, this overview enables us to see how the use of wearables for

health is currently limited by crucial challenges, which impact different functions in different

ways. The remainder of the paper will be dedicated to a discussion of these challenges, as they

emerge in concrete uses of wearables to serve the functions we have identified.

Local standards for data quality

In our overview, we have identified monitoring as the fundamental feature at the basis of the

functions of wearable technology for health. Monitoring is a promising application of wear-

ables thanks to their abilities for constant and personal data collection, but a key concern is

data quality. Quality is a crucial feature of scientific data, which needs to be evaluated to war-

rant the reliability of scientific claims. Data quality is also one of the fundamental values of

research ethics and the social goals of biomedical research—high-quality data are considered

the basis for benefits at the clinical level and beyond [25]. Yet, the variability of sensors and

lack of consistency of data collection in the wearable context make it difficult to coordinate

and assess quality. In addition, the lack of contextual information on the ways in which wear-

able data are collected, classified, and interpreted raise concerns on the possibility of assessing

quality.

A first issue that makes it difficult to assess data quality in the wearable context is variability.

Wearable data are usually collected by different types of devices or different sensors, if not

through different data collection practices. For example, the measurement of metrics such as

oxygen saturation can vary substantially in terms of location (e.g., wrist, finger, ear) and types

of devices (e.g., watches, rings, earphones) employed for measurement [1,11]. This level of var-

iability makes it difficult to have common standards to assess data quality: The same parameter

is often measured with very different sensors, which employ different processing techniques,

and may even render different results [3]. One way of responding to these concerns is regula-

tion, which should make sure that wearables can be used as reliable and high-quality sources

of data. In this context, the push is to regulate wearables as medical devices on the basis of clin-

ical validity [26]. Clinical validity is a crucial step for the adoption of wearable technologies for

health and also for the regulation of the quality of wearable data.

Yet, clinical validity as an intrinsic feature of quality is not enough on its own. Extensive

work in philosophy of science has shown that quality is not only an intrinsic feature of data.

Quality is a contextual component of data: depending on a specific use and context of use, con-

siderations of data quality might change [27]. For example, it is clearly crucial to know that a

wearable device has been clinically validated to collect high-quality data on heart rate [22].

However, using a wearable to monitor heart rate and detect COVID-19 infection on this basis

constitutes a new context of use, where considerations of data quality may be different. For

example, a certain number of false positives might be considered good enough for fitness

tracking or even remote monitoring of heart patients, but it might not be enough when wear-

ables are used to detect COVID-19 and suggest isolation and quarantines. The ethical and

social burdens of poor quality or unreliable data change depending on the context. For patients

who rely on wearables to track the health of their heart after bypass surgery, the quality of data

is more serious than for users tracking fitness activities: In the context of heart patients, the

collection of low-quality data about severe health problems can be a very serious burden, lead-

ing to unnecessary anxiety [28]. For users with limited financial resources, wearables can be an

PLOS DIGITAL HEALTH

PLOS Digital Health | https://doi.org/10.1371/journal.pdig.0000104 October 13, 2022 4 / 13

inexpensive tool to keep track of their health when other health services are too expensive and

difficult to access [2]. As such, wearables can improve access to healthcare, but wearable tech-

nology might constitute the main health service available for these users and poor data quality

will unequally impact them more than others. This is why data quality should be considered a

contextual property of data that needs to be constantly considered as the context of use

changes significantly.

In turn, in order to assess data quality for a specific use, knowledge of contextual features of

data collection is crucial. For example, it is crucial to know which experimental procedures

and protocols were applied, which sensors and techniques were used, and which questions and

hypotheses were investigated during data collection. As questions and hypotheses change from

general heart monitoring to COVID-19 detection, for example, it is crucial to know the origi-

nal experimental procedures and questions to understand whether data quality remains the

same—access to contextual features of data practices is crucial to assess quality and ensure the

reliability and trustworthiness of data [29]. The problem is that access to these contextual fea-

tures is often not available in the wearable context. For example, the collection of heart rate

data from wearable devices is usually covered by the opt-in of users to general medical studies,

which are organised by private companies and large research bodies, such as the Apple Heart

Study created by the collaboration between Stanford University and Apple. While this type of

study was of course validated and regulated as a clinical trial [30], there is little information on

data collection, analysis, storing, and access and this makes it difficult to assess quality. In addi-

tion, in many cases biomedical researchers cannot even download data directly from the

device and have to go through proprietary archives. As a result, because of commercial inter-

ests, very little information on how the data are collected, classified, and interpreted by the

device is shared throughout the process. This is an issue for researchers, but also users and

patients. The lack of access to contextual information about data collection makes it difficult

for users to interpret the data to take action on their health and can eventually lead them not

to trust and use the technology [28,31,32].

On this basis, we need more contextual information and coherence for wearable data quality

[33]. Contextual information can be used to understand the specific features and needs for the

assessment of data quality in the wearable context. In this direction, common and local standards

of data quality can be developed to overcome current limitations and gaps in the wearable market.

For example, the framework provided by FAIR (Findability, Accessibility, Interoperability, and

Reuse) can be used as a basis to discuss future developments in this direction [34–36]. We do not

see these as hard compliance standards set by standard organisations, but rather the result of a

bottom-up process of coordination and assessment, as a way of fully appreciating the contextual

dimensions of data quality. As we have seen, depending on the specific context of use, standards,

requirements, and burdens of data quality can change. This is why data quality standards need to

be local and could first be created for the research context, where knowledge of the contextual

components of data collection is crucial to assess data quality. Yet, standards of data quality can

clearly be crucial for regulators, institutions, industry, and users too, and standards could be

adopted by the institutions, journals, repositories of specific research communities. Again simi-

larly to FAIR data, the institutions of specific research communities could be in charge of manag-

ing, updating, and assessing standards and informing individual users of their existence and

application, thus presenting data quality as a fundamental issue for digital health.

Interoperability for balanced estimations

As we have seen in our overview, detection and prediction are among the main functions for

which wearable devices are currently used in health. In turn, detection and prediction are

PLOS DIGITAL HEALTH

PLOS Digital Health | https://doi.org/10.1371/journal.pdig.0000104 October 13, 2022 5 / 13

fundamental activities at the basis of the production and use of scientific knowledge. Yet, issues

affecting balanced estimations in screening and prediction raise concerns on the grounding

and validity of wearables as detection and prediction tools.

In the COVID-19 pandemic, several models have been used to predict the development,

spreading, and impact of the pandemic, but they have also been at the centre of several cri-

tiques concerning their uncertain assumptions and limitations [37–40]. Wearable devices have

been proposed as potential solutions to some of these issues [8]. For example, data from Fitbits

have been used to detect elevated signals at the level of heart rate and temperature—these are

possible symptoms of COVID-19 that can be identified in advance or just when more explicit

symptoms surfaced [17]. This is an extremely promising use of wearables, but the status of pre-

dictions based on wearable data raises challenges. Applications of wearables for the detection

of COVID-19 are severely affected by overestimation, the issue where non-problematic condi-

tions and abnormalities are systematically detected or predicted as problematic. For example,

it is often difficult to differentiate between COVID-19 and seasonal influenza and cases of

standard influenza on the basis of wearable data—elevated heart rate can be interpreted as a

symptom of respiratory illness more generally and, as a result, wearables have wrongly

detected and predicted COVID-19 infections [17,24].

This is a crucial epistemic issue for testing the validity of using wearable data to perform or

assist prediction, but is also significant from an ethical point of view. For example, health

resources and personnel may be diverted from actually problematic situations towards overes-

timated issues, thus creating imbalance in health treatment and access [41]. Erroneous predic-

tion and detection can also create unnecessary stress in patients, raising concerns on the

implementation of wearables for health [7,42]. In addition, the burdens of overestimation may

also be unequally distributed over different types of social groups, policy contexts, healthcare

services. For example, estimation issues in the context of COVID-19 might be overcome with

access to molecular or antigen tests, which can differentiate between influenza and COVID-

19. In this sense, it could be argued that overestimating infections might thus be better in light

of the precautionary principle. However, access to fast COVID-19 testing is not equally avail-

able and distributed in the world and has often become expansive, especially when infections

surge. Policy decisions might require a person to isolate if their wearable device has detected a

possible infection (as we have seen with the use of contact-tracing smartphone apps), which is

potentially harmful for them and their family, especially if remote work is not an option and

wages might be lost. These issues are even more severe in the context of wearables and other

digital health solutions. These technologies are presented as key opportunities for parts of the

world with limited or non-existing health services [43]. However, if other technologies and ser-

vices that might help overcome overestimation are limited and not available (e.g., fast COVID-

19 testing), this poses even more significant constraints on the accuracy and estimation of

wearables and other digital health technologies. Unsurprisingly, recent work by political insti-

tutions, such as the EU Commission, on the internet of things technologies such as wearables

has concluded that overestimation is among the main issues for the adoption of wearable tech-

nology [2].

In order to overcome these concerns, we propose to focus on the interoperability of wear-

able data as a crucial way forward. Interoperability is the possibility that data can be integrated

and used together with other types of data [35]. Several philosophical, historical, and sociologi-

cal studies of the role of data in science have highlighted that the value of large volumes of data

for research lies in the possibility of integrating and linking different datasets [44]. In this

sense, a high level of interoperability is key to exploit the benefits of new and large datasets,

such as those collected with wearable technology. A low level of interoperability makes it diffi-

cult to integrate wearable data with other health data and thus compare and balance results

PLOS DIGITAL HEALTH

PLOS Digital Health | https://doi.org/10.1371/journal.pdig.0000104 October 13, 2022 6 / 13

collected by different devices, sensors, approaches. In turn, making sure that wearable data are

interoperable can make it easier to compare results obtained through other means and assess

the extent to which overestimation might be a problem. Data interoperability is also connected

to interoperability at the software and hardware level of wearable technology. For example, the

integration of wearables into health services is currently challenging because the additional

staff required to assist patients with the technology might need to be trained differently, as soft-

ware and hardware solutions are different between devices [45]. In turn, interoperability stan-

dards are also crucial for data storage and thus to include wearables in health services, for

example through personal and electronic health records, which is currently very costly [3], and

to deal with cyber risks, for example by highlighting transparency and accountability in health-

care infrastructures [46,47]. Ensuring that a wearable device is interoperable is thus an essen-

tial way to approach overestimation and the promise of providing more personal and precise

healthcare in digital health.

Access for health equity

One of the defining features of wearables is their ability to be worn on our bodies. This means

that they can often be personal devices, in the sense that they might fulfil a personal need of the

user (such as tracking fitness activities and exercise) as well as be used as personal and fashion

objects (such as rings and wristwatches). As such, wearables play a crucial role towards an increas-

ingly personalised, precise, and person-centred medicine. In this way, wearable technology is

uniquely positioned to move in the direction of one of the goals of digital health: expanding access

to health services and thus improving health equity. Health equity is about making sure that differ-

ent users are equally provided with services and care as part of their interactions with the health

system, as defined in several policy initiatives such as the Thirteenth General Programme of

Work of the World Health Organization (WHO). While we agree the contribution of wearables

to these goals is promising, issues connected to access raise significant concerns.

As we have seen, wearable devices can clearly provide data that are personal to user needs,

issues, and concerns [48]. However, the extent to which individual users can access the benefits

of this data collection seems unequally distributed. Users with more digital literacy and socio-

economic resources are disproportionately advantaged to access benefits from the use of wear-

ables as tools to detect and predict states of health and disease [49,50]. In addition, the use of

wearables and other digital health tools for monitoring in the context of public health efforts

might raise concerns about surveillance, in different ways for different social groups. Histori-

cally, members of marginalised social groups have been targeted by health surveillance and

monitoring with unclear benefits and sometimes harmful results. For example, COVID-19

surveillance and policy restrictions have disproportionately affected structurally disadvantaged

social groups [51]. If wearables as digital health technologies are to be made part of public

health policy and campaigns, access to the technology needs to be ensured as much as access

to clear benefits from the use of the technology. Currently, the benefits of health monitoring

through wearables are disproportionately available to consumer technology companies, rather

than individual users. Most wearables available on the market are developed and sold by some

of the largest corporations in the world, such as Apple and Google. The increasing collection

of health data through wearables by consumer technology creates clear economic and political

benefits for these corporations, which can use the data for marketing and advertising. Individ-

ual users do not necessarily have access to these benefits of data collection or at least not at the

same level [52].

In addition, even for those who can and do use wearables, other issues of access raise con-

cerns on health equity. As we have seen, contextual information on the collection,

PLOS DIGITAL HEALTH

PLOS Digital Health | https://doi.org/10.1371/journal.pdig.0000104 October 13, 2022 7 / 13

classification, interpretation of wearable data is usually not shared by data providers and device

manufacturers. This is an issue for health equity: epistemically, information of the ways in

which wearable data are analysed for detection and screening is crucial to interpret data and

translate results into significant actions of health promotion for individual users. Without this

information, users can struggle to understand why the analysis of wearable data leads to the

detection of a condition and how they can act upon this function. This can also create doubt

and anxiety, as users do not know the extent to which the data are reliable and are unsure

about the actions they can take to counter possibly alarming conclusions [28]. In other words,

this creates a situation of health inequity. For some users, the collection of wearable data can

be a source of actions to improve their health, but for others barriers to data access can create

new burdens.

Several approaches have been proposed in recent years to counteract the burdens of health

inequity [53]. A way forward for these challenges in the wearable context is an expansion of

both the access to the data and related interpretation tools. More access to data can partially

counter the economic and political power of technology corporations [54]. Access to interpre-

tation, in turn, can empower users, enabling them to make sense of the trustworthiness, qual-

ity, and actionability of the functions provided by wearables. We see these as goals that should

be part of health campaigns and public health policy involving wearables and other digital

health technologies. Crucially for health equity, however, access to technology should be

approached critically, in light of considerations of the specific social and political context of

use. For example, some members of the general population might not be interested in tracking

their health or might find it confusing, alienating, guilt-inducing, stressful. The specific use

and position of wearables as digital health technology needs to be openly and critically dis-

cussed to ensure that those who choose not to be part of the movement are not unequally

treated and loose access to other health services.

Representativity for fairness

Wearables are at the centre of several attempts to make health more mobile and digital. As we

have seen in our overview, wearables are technologies that can track and collect digital data on

various daily activities and provide users with individual monitoring and screening in connec-

tion to other digital tools and services. Wearables can also be ways of further developing

remote detection and prediction, without the need to interact with other health services. In the

digital health context, this use of digital devices and services such as wearables is connected to

various benefits. For example, digital health is often framed explicitly as an opportunity to shift

the medical knowledge system towards the representation of the majority that is typically

excluded from more traditional research methodology [55]. While wearables are clearly prom-

ising tools to achieve these crucial goals, we raise concerns on their fairness. In the health con-

text, fairness is close to the notion of equity and related attempts for the equal distribution of

services and care. Yet, fairness is also about the just treatment of individuals when they interact

with health services and thus about making sure that people are not treated in unjust ways in

healthcare because of bias, discrimination, lack of consideration [56]. We argue that current

use and features of wearables disproportionately target some members of the general popula-

tion and exclude others, thus creating issues of fairness.

Thanks to wearable and other digital devices, data points such as steps have been tracked

for almost a decade, at a scale that is unprecedented when compared to more traditional and

preceding data practices. In these ways, more generally, wearables are contributing to the

increasing datafication of activities and aspects of our lives. But they are contributing to their

medicalisation too, as the possibility of quantifying and measuring these activities and aspects

PLOS DIGITAL HEALTH

PLOS Digital Health | https://doi.org/10.1371/journal.pdig.0000104 October 13, 2022 8 / 13

renders them as new areas of research and intervention. In the health context, current pro-

cesses of datafication and medicalisation are contributing to a re-configuration of health, by

expanding the limits and remits of biomedical research, producing new markers of health and

disease, redefining what counts as health data, broadening the categories of influential stake-

holders, and involving and empowering more individuals [32]. Datafication and medicalisa-

tion through wearables can thus create various benefits by uncovering new health needs and

issues of specific communities. Consider, for example, the role that patient groups have played

throughout the COVID-19 pandemic in raising concerns on the limitations and diverse

impact of public health interventions and raising awareness on the long-lasting effects of

COVID-19 infection, which are now known as long COVID. Enabling patients to track their

own health individually and actively can provide them with more powerful tools and empow-

erment in this direction.

However, current uses and applications of wearable technology for health focus only on

some members and groups of the general population, thus rendering the use of the technology

unfair. For example, consider the framing of wearable technology as a crucial tool for the

remote and constant monitoring of the elderly and patients that need to practice social dis-

tancing, avoid hospital visits but require monitoring [8]. Looking at current figures on the

adoption of wearable devices, members of the population that fit into these categories are

severely underrepresented and excluded by the application of this technology [2]. This is

highly problematic from the point of view of fairness: Wearable technology seems to exclude

the users that arguably would benefit the most from the use of wearables. Children are also an

interesting type of users in this sense. Age groups including young adolescents and children

have increasing access to digital technologies, including wearables. Yet, the adoption of wear-

able technology in children can vary substantially, for example depending on whether they use

other technologies (e.g., smartphones are normally gateways for wearables), where they live,

and the socioeconomic status of their family. The new contribution of these age groups to bio-

medical research is an exciting opportunity of wearable technology, potentially enabling the

retooling of medical knowledge system to represent groups that are currently excluded and

underrepresented [55]. At the same time, the opportunity of further introducing wearable

technology in these age groups needs to be balanced against ethical reflections about security,

privacy, intrusiveness. More generally, the cases of the elderly and children suggest that, how-

ever, large and extended wearable datasets may be, wearables usually target some social, eco-

nomic, age groups more than others. This is crucial because excluding important and large

parts of the general population can lead to biased and underrepresentative datasets, which do

not give us a good picture of population health, thus creating a weak and unsound basis for

knowledge claims and focusing health policy only on few members of the population.

Thus, issues of fairness raise concerns on the legitimacy of using and recommending wear-

able technology for health in the general population. To overcome these challenges, more

focus needs to be given on the representativity of various members of the general population

in wearable technology and digital health. We see the focus on representativity as one of the

steps of the assessment of data quality and fairness [36], which should be one of the first steps

for discussions on using wearables as part of health promotion and public health programmes

too. In addition, focusing on representativity can also be a way of taking into account the con-

text around the use and introduction of wearable technology. In communities and parts of the

world with limited availability of fast and inexpensive testing, for example, early detection of

pre-symptomatic COVID-19 is not as useful or might be useful only for some members of the

population, thus creating issues of fairness. Consider one of the prime areas of application of

wearables for health: the tracking of physical activity to suggest interventions and behavioural

change [57,58]. Wearables can be powerful tools in this context—yet alerting a person that

PLOS DIGITAL HEALTH

PLOS Digital Health | https://doi.org/10.1371/journal.pdig.0000104 October 13, 2022 9 / 13

they have been sedentary might not be as useful, if they do not have opportunities or services

that can make them more active.

Conclusions

In this paper, we have discussed various implications of wearable technology for digital health.

First, we have identified functions that wearable technology currently serves in biomedical

research and clinical care as a way of specifying the epistemic contribution of wearables to the

development and application of biomedical knowledge through monitoring, screening, detec-

tion, and prediction (Table 1). On this basis, we have discussed a number of challenges that are

connected to these functions, particularly at the level of data quality, estimations, equity, and

fairness. As a way to overcome these challenges, we have introduced recommendations and

possible solutions based on local standards of quality, interoperability, access, and representa-

tivity (Table 2). Our analysis has thus been aimed at improving our understanding of the posi-

tion and relations between wearables and other biomedical technologies and data sources, as

well as ways to approach their adoption and regulation.

Throughout the article, we have applied an integrated approach for the discussion of wear-

ables for health, which we see as a starting point for more work. In recent years, philosophers,

sociologists, and ethicists of science and technology have started to work more closely in col-

laboration with biomedical scientists, engineers, and practitioners. An increasing number of

publications is the result of collaborations between science scholars and scientists; philosophi-

cal work is increasingly relevant and cited in science journals [59]. In this context, approaches

such as ELSI (Ethical, Legal and Social Issues) and E2LSI show the need for a systematic inte-

gration of epistemic, ethical, legal, and social considerations [60]. This is a particularly impor-

tant step to take in the context of new and evolving technologies for digital health, as

important decisions are being taken now on their regulation, inclusion in healthcare pro-

grammes, and use in research. Our work in this article provides a first step for thinking about

these as integrated issues.

References 1. Aliverti A. Wearable technology: role in respiratory health and disease. Breathe. 2017; 13:e27–e36.

https://doi.org/10.1183/20734735.008417 PMID: 28966692

2. Shin G, Jarrahi MH, Fei Y, Karami A, Gafinowitz N, Byun A, et al. Wearable activity trackers, accuracy,

adoption, acceptance and health impact: A systematic literature review. J Biomed Inform. 2019;

93:103153. https://doi.org/10.1016/j.jbi.2019.103153 PMID: 30910623

3. Sim I. Mobile Devices and Health. N Engl J Med. 2019; 381:956–968. https://doi.org/10.1056/

NEJMra1806949 PMID: 31483966

4. Lupton D. M-health and health promotion: The digital cyborg and surveillance society. Soc Theory

Health. 2012; 10:229–244. https://doi.org/10.1057/sth.2012.6

5. Sharon T. Self-Tracking for Health and the Quantified Self: Re-Articulating Autonomy, Solidarity, and

Authenticity in an Age of Personalized Healthcare. Philos Technol. 2017; 30:93–121. https://doi.org/10.

1007/s13347-016-0215-5

6. Owens J, Cribb A. ‘My Fitbit Thinks I Can Do Better!’ Do Health Promoting Wearable Technologies Sup-

port Personal Autonomy? Philos Technol. 2019; 32:23–38. https://doi.org/10.1007/s13347-017-0266-2

7. Predel C, Steger F. Ethical Challenges With Smartwatch-Based Screening for Atrial Fibrillation: Putting

Users at Risk for Marketing Purposes? Front Cardiovasc Med. 2021; 7:615927. https://doi.org/10.3389/

fcvm.2020.615927 PMID: 33521064

8. Seshadri DR, Davies EV, Harlow ER, Hsu JJ, Knighton SC, Walker TA, et al. Wearable Sensors for

COVID-19: A Call to Action to Harness Our Digital Infrastructure for Remote Patient Monitoring and Vir-

tual Assessments. Front Digit Health. 2020; 2:8. https://doi.org/10.3389/fdgth.2020.00008 PMID:

34713021

PLOS DIGITAL HEALTH

PLOS Digital Health | https://doi.org/10.1371/journal.pdig.0000104 October 13, 2022 10 / 13

9. Channa A, Popescu N, Skibinska J, Burget R. The Rise of Wearable Devices during the COVID-19 Pan-

demic: A Systematic Review. Sensors. 2021; 21:5787. https://doi.org/10.3390/s21175787 PMID:

34502679

10. Perez MV, Mahaffey KW, Hedlin H, Rumsfeld JS, Garcia A, Ferris T, et al. Large-Scale Assessment of

a Smartwatch to Identify Atrial Fibrillation. N Engl J Med. 2019; 381:1909–1917. https://doi.org/10.

1056/NEJMoa1901183 PMID: 31722151

11. Ding X, Clifton D, Ji N, Lovell NH, Bonato P, Chen W, et al. Wearable Sensing and Telehealth Technol-

ogy with Potential Applications in the Coronavirus Pandemic. IEEE Rev Biomed Eng. 2021; 14:48–70.

https://doi.org/10.1109/RBME.2020.2992838 PMID: 32396101

12. Radin JM, Wineinger NE, Topol EJ, Steinhubl SR. Harnessing wearable device data to improve state-

level real-time surveillance of influenza-like illness in the USA: a population-based study. Lancet Digital

Health. 2020; 2:e85–e93. https://doi.org/10.1016/S2589-7500(19)30222-5 PMID: 33334565

13. Contini M, Sarmento A, Gugliandolo P, Leonardi A, Longinotti-Buitoni G, Minella C, et al. Validation of a

new wearable device for type 3 sleep test without flowmeter. Rowley JA, editor. PLoS ONE. 2021; 16:

e0249470. https://doi.org/10.1371/journal.pone.0249470 PMID: 33861776

14. Bumgarner JM, Lambert CT, Hussein AA, Cantillon DJ, Baranowski B, Wolski K, et al. Smartwatch

Algorithm for Automated Detection of Atrial Fibrillation. J Am Coll Cardiol. 2018; 71:2381–2388. https://

doi.org/10.1016/j.jacc.2018.03.003 PMID: 29535065

15. Quer G, Radin JM, Gadaleta M, Baca-Motes K, Ariniello L, Ramos E, et al. Wearable sensor data and

self-reported symptoms for COVID-19 detection. Nat Med. 2021; 27:73–77. https://doi.org/10.1038/

s41591-020-1123-x PMID: 33122860

16. Burnham JP, Lu C, Yaeger LH, Bailey TC, Kollef MH. Using wearable technology to predict health out-

comes: a literature review. J Am Med Inform Assoc. 2018; 25:1221–1227. https://doi.org/10.1093/

jamia/ocy082 PMID: 29982520

17. Mishra T, Wang M, Metwally AA, Bogu GK, Brooks AW, Bahmani A, et al. Pre-symptomatic detection of

COVID-19 from smartwatch data. Nat Biomed Eng. 2020; 4:1208–1220. https://doi.org/10.1038/

s41551-020-00640-6 PMID: 33208926

18. Pyrkov TV, Slipensky K, Barg M, Kondrashin A, Zhurov B, Zenin A, et al. Extracting biological age from

biomedical data via deep learning: too much of a good thing? Sci Rep. 2018; 8:5210. https://doi.org/10.

1038/s41598-018-23534-9 PMID: 29581467

19. Singh G, Tee A, Trakoolwilaiwan T, Taha A, Olivo M. Method of respiratory rate measurement using a

unique wearable platform and an adaptive optical-based approach. ICMx. 2020; 8:15. https://doi.org/

10.1186/s40635-020-00302-6 PMID: 32449051

20. Rosenbaum L. Swallowing a Spy—The Potential Uses of Digital Adherence Monitoring. N Engl J Med.

2018; 378:101–103. https://doi.org/10.1056/NEJMp1716206 PMID: 29281504

21. Radin JM, Quer G, Ramos E, Baca-Motes K, Gadaleta M, Topol EJ, et al. Assessment of Prolonged

Physiological and Behavioral Changes Associated With COVID-19 Infection. JAMA Netw Open. 2021;

4:e2115959. https://doi.org/10.1001/jamanetworkopen.2021.15959 PMID: 34232306

22. Bayoumy K, Gaber M, Elshafeey A, Mhaimeed O, Dineen EH, Marvel FA, et al. Smart wearable devices

in cardiovascular care: where we are and how to move forward. Nat Rev Cardiol. 2021 [cited 2021 May

6]. https://doi.org/10.1038/s41569-021-00522-7 PMID: 33664502

23. DiFrancisco-Donoghue J, Jung M-K, Stangle A, Werner WG, Zwibel H, Happel P, et al. Utilizing wear-

able technology to increase physical activity in future physicians: A randomized trial. Prev Med Rep.

2018; 12:122–127. https://doi.org/10.1016/j.pmedr.2018.09.004 PMID: 30234000

24. Piroth L, Cottenet J, Mariet A-S, Bonniaud P, Blot M, Tubert-Bitter P, et al. Comparison of the character-

istics, morbidity, and mortality of COVID-19 and seasonal influenza: a nationwide, population-based ret-

rospective cohort study. Lancet Respir Med. 2021; 9:251–259. https://doi.org/10.1016/S2213-2600(20)

30527-0 PMID: 33341155

25. Cai L, Zhu Y. The Challenges of Data Quality and Data Quality Assessment in the Big Data Era.

CODATA. 2015; 14:2. https://doi.org/10.5334/dsj-2015-002

26. Ravizza A, De Maria C, Di Pietro L, Sternini F, Audenino AL, Bignardi C. Comprehensive Review on

Current and Future Regulatory Requirements on Wearable Sensors in Preclinical and Clinical Testing.

Front Bioeng Biotechnol. 2019; 7:313. https://doi.org/10.3389/fbioe.2019.00313 PMID: 31781554

27. Canali S. Towards a Contextual Approach to Data Quality. Data. 2020; 5:90. https://doi.org/10.3390/

data5040090

28. Lomborg S, Langstrup H, Andersen TO. Interpretation as luxury: Heart patients living with data doubt,

hope, and anxiety. Big Data Soc. 2020; 7:205395172092443. https://doi.org/10.1177/

2053951720924436

PLOS DIGITAL HEALTH

PLOS Digital Health | https://doi.org/10.1371/journal.pdig.0000104 October 13, 2022 11 / 13

29. Leonelli S. Global Data Quality Assessment and the Situated Nature of “Best” Research Practices in

Biology. Data Sci J. 2017; 16:32. https://doi.org/10.5334/dsj-2017-032

30. Turakhia MP, Desai M, Hedlin H, Rajmane A, Talati N, Ferris T, et al. Rationale and design of a large-

scale, app-based study to identify cardiac arrhythmias using a smartwatch: The Apple Heart Study. Am

Heart J. 2019; 207:66–75. https://doi.org/10.1016/j.ahj.2018.09.002 PMID: 30392584

31. Prainsack B. Personalized medicine: empowered patients in the 21st century? New York: New York

University Press; 2017.

32. Sharon T, Lucivero F. Introduction to the Special Theme: The expansion of the health data ecosystem–

Rethinking data ethics and governance. Big Data Soc. 2019; 6:205395171985296. https://doi.org/10.

1177/2053951719852969

33. Zhang J, Symons J, Agapow P, Teo JT, Paxton CA, Abdi J, et al. Best practices in the real-world data

life cycle. McGinnis RS, editor. PLoS Digit Health. 2022; 1:e0000003. https://doi.org/10.1371/journal.

pdig.0000003

34. Wilkinson MD, Dumontier M, IjJ A, Appleton G, Axton M, Baak A, et al. The FAIR Guiding Principles for

scientific data management and stewardship. Sci Data. 2016; 3:160018. https://doi.org/10.1038/sdata.

2016.18 PMID: 26978244

35. Blasimme A, Fadda M, Schneider M, Vayena E. Data Sharing For Precision Medicine: Policy Lessons

And Future Directions. Health Affairs. 2018; 37:702–709. https://doi.org/10.1377/hlthaff.2017.1558

PMID: 29733719

36. Leonelli S, Lovell R, Wheeler BW, Fleming L, Williams H. From FAIR data to fair data use: Methodologi-

cal data fairness in health-related social media research. Big Data Soc. 2021; 8:205395172110103.

https://doi.org/10.1177/20539517211010310

37. Saltelli A, Bammer G, Bruno I, Charters E, Di Fiore M, Didier E, et al. Five ways to ensure that models

serve society: a manifesto. Nature. 2020; 582:482–484. https://doi.org/10.1038/d41586-020-01812-9

PMID: 32581374

38. Maziarz M, Zach M. Agent-based modelling for SARS-CoV-2 epidemic prediction and intervention

assessment: A methodological appraisal. J Eval Clin Pract. 2020; 26:1352–1360. https://doi.org/10.

1111/jep.13459 PMID: 32820573

39. Fuller J. What are the COVID-19 models modeling (philosophically speaking)? Hist Philos Life Sci.

2021; 43:5.

40. Van Basshuysen P, White L, Khosrowi D, Frisch M. Three Ways in Which Pandemic Models May Per-

form a Pandemic. EJPE. 2021; 14. https://doi.org/10.23941/ejpe.v14i1.582

41. Vogt H, Green S, Ekstrøm CT, Brodersen J. How precision medicine and screening with big data could

increase overdiagnosis. BMJ. 2019: l5270. https://doi.org/10.1136/bmj.l5270 PMID: 31519649

42. Brodersen J, Siersma VD. Long-Term Psychosocial Consequences of False-Positive Screening Mam-

mography. Ann Fam Med. 2013; 11:106–115. https://doi.org/10.1370/afm.1466 PMID: 23508596

43. Pinto FJ, Piñeiro D, Banerjee A, Perel P, Pervan B, Eiselé J-L. World Heart Day 2021: COVID-19, digital

health, and tackling cardiovascular disease. Lancet. 2021; 398:1467–1468. https://doi.org/10.1016/

S0140-6736(21)02144-9 PMID: 34597535

44. Leonelli S, Tempini N, editors. Data Journeys in the Sciences. Cham: Springer International Publish-

ing; 2020. https://doi.org/10.1007/978-3-030-37177-7

45. Adler-Milstein J, Nong P. Early experiences with patient generated health data: health system and

patient perspectives. J Am Med Inform Assoc. 2019; 26:952–959. https://doi.org/10.1093/jamia/ocz045

PMID: 31329886

46. Radanliev P, De Roure D, Ani U, Carvalho G. The ethics of shared Covid-19 risks: an epistemological

framework for ethical health technology assessment of risk in vaccine supply chain infrastructures.

Health Technol. 2021; 11:1083–1091. https://doi.org/10.1007/s12553-021-00565-3 PMID: 34123697

47. Radanliev P, De Roure D. Epistemological and Bibliometric Analysis of Ethics and Shared Responsibil-

ity—Health Policy and IoT Systems. Sustainability. 2021; 13:8355. https://doi.org/10.3390/su13158355

48. Lu L, Zhang J, Xie Y, Gao F, Xu S, Wu X, et al. Wearable Health Devices in Health Care: Narrative Sys-

tematic Review. JMIR Mhealth Uhealth. 2020; 8:e18907. https://doi.org/10.2196/18907 PMID:

33164904

49. Prainsack B. Personalised and Precision Medicine: What Kind of Society Does It Take? In: Meloni M,

Cromby J, Fitzgerald D, Lloyd S, editors. The Palgrave Handbook of Biology and Society. London: Pal-

grave Macmillan UK; 2018. p. 683–701. https://doi.org/10.1057/978-1-137-52879-7_29

50. Staunton C, Barragán CA, Canali S, Ho C, Leonelli S, Mayernik M, et al. Open science, data sharing

and solidarity: who benefits? HPLS. 2021; 43:115. https://doi.org/10.1007/s40656-021-00468-6 PMID:

34762203

PLOS DIGITAL HEALTH

PLOS Digital Health | https://doi.org/10.1371/journal.pdig.0000104 October 13, 2022 12 / 13

51. Fiske A, Galasso I, Eichinger J, McLennan S, Radhuber I, Zimmermann B, et al. The second pandemic:

Examining structural inequality through reverberations of COVID-19 in Europe. Soc Sci Med. 2022;

292:114634. https://doi.org/10.1016/j.socscimed.2021.114634 PMID: 34883310

52. Prainsack B. Logged out: Ownership, exclusion and public value in the digital data and information com-

mons. Big Data Soc. 2019; 6:205395171982977. https://doi.org/10.1177/2053951719829773

53. Prainsack B. The political economy of digital data: introduction to the special issue. Policy Studies.

2020; 41:439–446. https://doi.org/10.1080/01442872.2020.1723519

54. Hummel P, Braun M, Dabrock P. Own Data? Ethical Reflections on Data Ownership. Philos Technol.

2020 [cited 2021 Jun 8]. https://doi.org/10.1007/s13347-020-00404-9

55. Celi LA. PLOS Digital Health, a new journal driving transformation in the delivery of equitable and unbi-

ased healthcare. PLoS Digit Health. 2022; 1:e0000009. https://doi.org/10.1371/journal.pdig.0000009

56. Olsen JA. Concepts of Equity and Fairness in Health and Health Care. In: Glied S, Smith PC, editors.

The Oxford Handbook of Health Economics. Oxford University Press; 2011. p. 813–836. https://doi.

org/10.1093/oxfordhb/9780199238828.013.0034

57. Laranjo L, Ding D, Heleno B, Kocaballi B, Quiroz JC, Tong HL, et al. Do smartphone applications and

activity trackers increase physical activity in adults? Systematic review, meta-analysis and metaregres-

sion. Br J Sports Med. 2021; 55:422–432. https://doi.org/10.1136/bjsports-2020-102892 PMID:

33355160

58. McDonough DJ, Su X, Gao Z. Health wearable devices for weight and BMI reduction in individuals with

overweight/obesity and chronic comorbidities: systematic review and network meta-analysis. Br J

Sports Med. 2021; 55:917–925. https://doi.org/10.1136/bjsports-2020-103594 PMID: 33731385

59. Khelfaoui M, Gingras Y, Lemoine M, Pradeu T. The visibility of philosophy of science in the sciences,

1980–2018. Synthese. 2021 [cited 2021 Jun 28]. https://doi.org/10.1007/s11229-021-03067-x

60. Lohse S, Wasmer MS, Reydon TAC. Integrating Philosophy of Science into Research on Ethical, Legal

and Social Issues in the Life Sciences. Perspect Sci. 2020; 28:700–736. https://doi.org/10.1162/posc_

a_00357

PLOS DIGITAL HEALTH

PLOS Digital Health | https://doi.org/10.1371/journal.pdig.0000104 October 13, 2022 13 / 13