Improving Access to an Increased Range of Health Services and the Potential of Cost-Saving

Digital health solutions hold the promise to bridge health inequity by expanding the range of health services to a wider population, be it in high-income or low-income settings (Rahimi, 2019). In particular, a review on mobile health—which is an important domain of digital health—has shown to result in substantial impacts in low- and middle-income countries by improving treatment adherence and appointment compliance among the patients, as well as supporting data gathering and the development of support networks among the health workers (Hall et al., 2014).

In terms of cost, the question of whether digital health solutions will increase healthcare cost due to their rising demand in meeting previously unmet needs, or reduce overall healthcare expenditures due to the lower operational costs, is debatable (Rahimi, 2019). However, preliminary evidence on the use of digital health tools to facilitate the recovery of specific clinical conditions such as cardiovascular diseases have shown potential for cost-saving (Bhardwaj et al., 2021; Jiang et al., 2019). A systematic review on the use of digital health interventions in the management of cardiovascular diseases reported an improvement in quality-adjusted life-year (QALY) in 43% of the studies as compared to standard care (Jiang et al., 2019). Another simulation study that examined the use of smartphone application, smartwatch, and wireless blood pressure monitor to support medication adherence, patient education, vital signs monitoring, and care coordination in acute myocardial infraction recovery had shown that when this intervention was delivered along with standard care, it resulted in significant cost-saving and improved QALY (Bhardwaj et al., 2021).

Maximizing Population-Wide Health Impacts Through Behavioral Modifications

Through customized intervention and personalized messaging, digital health tools and interventions can target behavioral changes at the individual level to reap population-wide health impacts. A bibliometric analysis on the use of digital health technology to change health behavior reported that digital technology–enabled behavioral change interventions aimed at promoting physical activity and dietary change have seen a steep rise since 2001 (Taj et al., 2019).

Digital health interventions delivered via mobile phone apps have shown to be effective in managing a range of acute and chronic conditions by promoting patient self-management and behavioral change, and have led to better physical and mental health outcomes (Willems et al., 2021). Through digital phenotyping that primarily employs passive data collection such as a smartphone capturing a person’s physical location, individual activity, voice, and speech, as well as a person’s interaction with their mobile device through typing and scrolling, digital data can be gathered to understand and predict behavioral health outcomes of interest (Marsch, 2021). Furthermore, digital segmentation, which is based on the use of digital technology behavior to sort and reach the right audience in a targeted manner by predicting their engagement with online advertisement and behavioral change campaigns, has been demonstrated to be quite successful in health behavior modifications among the youth and young adults (Evans et al., 2019).

Controlling and Managing Public Health Emergencies

The importance of digital health solutions in controlling and managing public health emergencies has been elevated since the COVID-19 pandemic that started in November 2019. Since the start of the pandemic in December 2019, various digital health solutions have been deployed. These range from artificial intelligence for surveillance, digital platforms for communication and data management, digital structural screening, and the applications of internet of things to respond to different situational challenges during the pandemic (Budd et al., 2020; Gunasekeran et al., 2021; Willems et al., 2021). To effectively curb virus transmission, many countries have designed and launched their own national contact tracing apps to enable more effective surveillance to support social distancing guidelines, to streamline the public health protocols for quarantine/isolation, and to detect infection clusters (Budd et al., 2020). Clinically, a comprehensive digital health management system that includes the use of telemedicine or virtual conferencing to monitor patients with mild symptoms recovering at home would allow healthcare professionals to better respond to the evolving needs of care management during the pandemic (Anthony, 2021). Besides, a comprehensive digital health management system would also allow the healthcare workers to escalate protocol remotely if a patient’s symptoms worsen and require acute management in a hospital setting (Willems et al., 2021).

Monitoring: Effective Disease Screening and Pandemic Control

The foremost benefit of digital health technology is to promote effective screening of diseases in a timely manner. In the recent years, telemedicine has emerged as a reliable tool to provide healthcare services and consultation remotely. It has advanced from the use of telephone calls and short message services or emails, to digital communications that enable virtual diagnosis (Cahn et al., 2018; Dorsey & Topol, 2016). This advancement allows remote monitoring of patients, improves administration and management of healthcare, and greatly reduces the need for individuals to physically visit clinics when their health concerns can be managed via virtual consultations (Mitchell & Kan, 2019; Senbekov et al., 2020). During the COVID-19 pandemic, telemedicine has helped to reduce the possibility of disease transmission by allowing healthcare providers to remotely look after individuals who are in quarantine (Senbekov et al., 2020). Telemedicine was also shown to benefit patients with chronic illnesses and required close monitoring of their conditions (Orozco-Beltran et al., 2017; Wootton, 2012; Zhang et al., 2021). Telemedicine is a digital revolution that changes our mode of healthcare provision during the COVID-19 pandemic. Before the pandemic, the healthcare industry was predominantly structured based on the conventional mode of in-person integrations between patients and clinicians. However, during the disease outbreak, increased physical contact facilitated the spread of virus and was thus minimized. The use of telemedicine in this scenario allows the delivery of care services through minimal physical interaction between the healthcare providers and the patients (Keesara et al., 2020; Mann et al., 2020; Monaghesh & Hajizadeh, 2020).

With the integration of digital technologies to the health system, diseases can be screened by applying digital biomarkers—objective measures of health or biological response collected by digital devices—to determine one’s physiological responses to therapeutic intervention. Nowadays, wearable devices (such as finger‐worn sensors and biometric skin patches) and smartphone apps (such as cognitive assessment and typing behavior) are able to passively measure certain physiological changes in the body without requiring the patients to wear any sensors. Such screening processes can be carried out either actively or passively by the users, enabling more frequent, objective, and sensitive measures of disease progression (Dockendorf et al., 2021).

A study conducted by Merck and Koneksa Health showed that the measurement of heart rate and blood pressure changes via mobile health was comparable to standard in‐clinic measures and provided enough sensitivity to detect treatment differences. This breakthrough suggests that applying digital devices for screening can provide meaningful information regarding an individual’s health condition (Dockendorf et al., 2021; Huang et al., 2020).

Digital monitoring can also contribute to disease management of chronic illnesses such as diabetes and heart diseases. Digital sensors and smart devices enable long-term and remote monitoring, thereby holding the promise to either prevent or reduce the chances of the occurrence of life-threatening conditions such as stroke and heart attack (Alwashmi, 2020; Senbekov et al., 2020). AliveCor’s smartphone electrocardiogram is the first Food and Drug Administration (FDA)–approved digital health sensor that can be used to detect atrial fibrillation. Other than smartphones, this application can also be incorporated into Apple Watch, further enhancing its monitoring capacity (Mesko, 2018). BlueStar is another FDA-approved application that can act as a diabetes management platform. Patients can adjust their insulin dosing by using the insulin calculator on the platform based on physicians’ prescriptions (Sharma et al., 2018).

Essentially, digital health devices can measure and record a wide range of health parameters such as the user’s vital signs, physical activities, dietary intake, energy expenditure, and glucose level. This wealth of information is not only helpful for patients’ self-monitoring and monitoring from physicians, it can also be fed into the health system to identify public health challenges and service gaps (Huat et al., 2019).

Digital monitoring also reduces the barrier for individuals inclined to seek help for mental health issues. Besides improving the accessibility to consultation and counseling, it could potentially reduce the stress and anxiety of affected individuals (Cao et al., 2020). For instance, users reported satisfaction and positive attitudes toward the use of digital health in addressing their mental health concerns as they felt that such an approach provided them more privacy and encouraged them to respond truthfully as compared to physical consultation (Liem et al., 2020).

From a macro perspective, digital health solutions have proven to be indispensable in the control of the COVID-19 pandemic. Surveillance provides a big picture of the disease outbreak. However, within the community, contact tracing serves as a critical step in containing the spread of the infectious agent. It identifies the disease carriers as well as individuals who have been in contact with the carrier. Thus, this approach serves as a primary measure of containment that terminates the chain of human-to-human transmission. As most epidemic/pandemic-causing infectious agents are highly transmissible, with our existing population density and fluidity, traditional paper-based contact tracing could not keep up with the transmission rate. However, with the help of technologies such as smartphones, Global Positioning System, and Bluetooth, digitalized contact tracing has contributed significantly to disease containment. For example, in Singapore, the use of Bluetooth technology in the mobile app has been used to identify individuals who have been in close contact with persons who have been diagnosed with COVID-19. This allows the authority to identify potential disease clusters and isolate potential cases as early as possible (Alwashmi, 2020; Lai et al., 2021; Tom-Aba et al., 2018). Similar applications have been developed and adopted in countries such as Australia, New Zealand, and France (Horstmann et al., 2021; Sharma et al., 2020; Touzani et al., 2021; Tretiakov & Hunter, 2020).

Other than identifying potentially infected individuals, contact tracing platforms can also be used to identify potential transmission hotspots and assess the effectiveness of public health interventions such as movement restrictions and mass vaccination programs (Budd et al., 2020).

Moreover, contact tracing platforms are often equipped with automatic alert systems that send reminders or message alerts to those likely to be ignorant about the lockdowns or movement restrictions during the pandemic (Radanliev et al., 2020).

Decision Support: Promoting Behavior Modifications and Optimizing Resource Allocation

Many digital health applications are implemented at the health agency level to manage the population’s health effectively. As health awareness grows and demand for the improvement of well-being increases among the population, this will inevitably generate a marketplace for health and fitness wearables. By using mobile apps, biosensor-equipped clothing, and wearable fitness devices, people can be more actively engaged in the process of changing their current lifestyle in order to gain more control over their personal health (Greiwe & Nyenhuis, 2020; Ledger, 2014; Montgomery et al., 2018). At the core of these “smart” devices, the internet of things enables sensors to power almost any ordinary object in people’s daily lives. The ability for the integration of this technology into every part of our life demonstrates its potential in changing our behaviors (Baig et al., 2019; Haghi et al., 2017; Montgomery et al., 2018; Piwek et al., 2016).

For individuals with health concerns, a digital health device serves as a tool to enhance their health consciousness. For example, it can remind the users about their medication intake, dietary restrictions, and sleep-wake routines. Most of the devices also provide real-time actionable feedbacks which users can access and examine. This empowers users by providing greater understanding of their health conditions, thus enabling them to make better decisions about their health (Greiwe & Nyenhuis, 2020; Montgomery et al., 2018; Patel et al., 2012; Stone et al., 2020; Van Hooren et al., 2020). In the longer run, digital wearables may modify certain disease-causing habits, potentially leading to reduced healthcare costs of chronic diseases such as diabetes, obesity, and hypertension.

Behind the interface of the digital health devices, companies are capturing data on the users’ contextual activity, health, and emotional state. The aim is to tie offline data to online behaviours and connect medical and clinical data with nonmedical behavioural and demographic information to infer and predict health behaviours and conditions. Simultaneously, predictive analysis using the existing data generated by the user was used to design a customized feedback mechanism for the user (de Arriba-Pérez et al., 2016; Gupta, 2015; Hernandez & Zhang, 2017; Montgomery et al., 2018).

In essence, the use of novel technologies creates a continuous data flow—from the collection of health information of the users through various devices to receiving that information via various intelligent communication platforms, clinicians can assess the health conditions of their patient in real time, allowing them to provide reliable, efficient, and personalized health care to their patients (Senbekov et al., 2020). Besides enabling two-way communication between physicians and their patients, these health data can be aggregated for the examination of population health profiles to facilitate better decision making in the allocation of health resources.

Education: Improving Health Knowledge and Awareness

In the medium to longer term, digital health technology would become a useful tool that health agencies can leverage to improve the delivery of safe and efficient population-level health initiatives to enhance health awareness among the citizens. For instance, the Australian government has launched My Health Record—a secure online summary of key health information for people and their healthcare providers. Through this one-stop platform, people can view their digitized health records such as medication records, allergies and diagnoses, hospital discharge summaries, medical subsidies, and billing information (Makeham, 2019; 2020; Senbekov et al., 2020).

The use of digital platforms as part of a comprehensive public health intervention is also observed for other communicable diseases such as human immunodeficiency virus (HIV) and other sexually transmitted infections (STIs). Digital platforms have been used to educate users about the importance of protected sex, deliver preexposure prophylaxis education, promote STI/HIV testing, and accelerate care and treatment for patients (Cao et al., 2020; Jones et al., 2019). Digital health interventions are especially effective toward the younger population. The appointment of social media influencers with a large number of followers to advocate for healthy sexual behaviors allows the message to receive more public attention. This peer-led intervention has been seen as a persuasive approach for on-campus health promotion and dating apps (Cao et al., 2020; Fernandez et al., 2019; Forsyth et al., 2018; Lau et al., 2019). Besides, the crowdsourcing approach can also increase public engagement for health promotion. This digital platform allows users or participants to be actively involved in the discussion and effectively collect feedback from the rest of the community (Fitzpatrick et al., 2018; Tang et al., 2018).

Data Privacy and Erosion of Trust

One of the most common governance challenges in the implementation of digital health technology is data privacy. At the health providers’ end, the breach of data privacy is usually a combination of negligence/human error, malicious intent, or the lack of an ethical framework to manage these oversights (Agboola et al., 2016; Andanda, 2020; Chernyshev et al., 2018). One of the most common digital health applications that is vulnerable to data privacy is the electronic patient record (EPR). The EPR stores longitudinal clinical data, laboratory, and diagnostic test results across various healthcare facilities, and can be linked through data portals in wide networks to facilitate clinical decision making, prescription refilling, appointment scheduling, and public health case reporting, as well as research applications. Hence, the wealth of data that EPR stores naturally predisposes it to such violations for the purpose of monetization (Birnbaum et al., 2018). In addition, the lack of robust privacy legislations or frameworks to govern the entire lifecycle of health data obtained from digital health apps is another common problem that results in data privacy violation. In spite of the adoption of General Data Protection Regulation in Europe since 2018, and the prescription on data privacy by the Health Insurance Portability and Accountability Act in the United States, there remain ambiguities surrounding data access and data ownership in the healthcare arena. Besides, there are no robust policies surrounding data destruction in the event that a user uninstalls an app or a manufacturer/developer goes out of business (Gordon & Stern, 2019).

The issue of lack of prescription on data destruction policies, including how much compensation that health providers or manufacturers will need to provide to individuals should their data be sold or repurposed for research use, is very much related to the immortal nature of the digital health footprint, which has raised longstanding debates about privacy on digital health information particularly, on genetic data privacy. Genetic data share this quality because the chain of custody is long, digital copies of genetic data can exist in multiple platforms, and chances are these data will be used perpetually as there is no set expiration date and no clear method to destroy these data at the moment (Grande et al., 2020). The immortal nature of genetic data that had raised debates among bioethicists can be best illustrated by the famous case of Henrietta Lacks, a black female patient whose cervical cancer cells were taken without her consent back in the 1950s in the United States. Her cervical cancer cell lines were later shared with multiple other research laboratories and applied extensively, which fueled key discoveries in multiple fields including cancer, immunology, infectious diseases, and mostly recently, on the development of COVID-19 vaccines (Nature, 2020). Genetic data are highly personal data and data scientists have shown that an individual can easily be identified from a population genetic database even without personal identifiers (Grande et al., 2020).

At the users’ end, ignorance as well as the lack of awareness on the privacy frameworks of various digital health applications may also contribute to the violation of data privacy and the erosion of trust. For instance, when a person downloads a digital health app in their smartphone, he or she may be unaware of whether a set of privacy policies or agreements exists in the app. This makes them incognizant toward the extent of data that they have “authorized” the developers to access, and whether private information are sold or sent to a third party (Filkins et al., 2016). For instance, a study analyzing mobile health apps in the United States revealed that 40% of the applications collected high risk-data that included financial information, full name, health information, geo-location, date of birth, and zip code (Filkins et al., 2016). Eighty-three percent of these applications stored their data locally without encryption, and only 50% of these applications have personal identifiable information encrypted (Filkins et al., 2016). Another study of diabetes apps for Android smartphones revealed that many diabetes self-monitoring apps routinely shared users’ information with third parties (Fleming et al., 2020). These are examples of data privacy breaches that are rampant due to the profitability that developers could gain from selling these big data to third parties, compounded by the lack of strong enforcement in data privacy governance.

The above discussed issues concerning data privacy and erosion of trust are similarly observed in low- and middle-income countries where regulatory mechanisms to govern data privacy are even more lacking (Gopichandran et al., 2020). In India, the Aadhaar biometric identification system has been subjected to various litigations due to data privacy breaches including linking digital health information to the Aadhaar identification system, leaking of its data to private telecommunication companies, and government websites inadvertently displaying individual data from the Aadhaar system (Gopichandran et al., 2020). Weak privacy laws and the lack of capabilities of many low- and middle-income countries to govern digital health data protection pose a fundamental challenge to the widespread applications of digital health technologies in those countries.

Safety

Patient well-being and safety are always at the heart of all safety discussions surrounding digital health applications (Dhingra & Dabas, 2020; Fleming et al., 2020). Many medical information embedded in the digital health apps may be misinterpreted by patients, especially when they are acquired from uncredible online resources without guidance from the health workers or manufacturers who developed the apps (Dhingra & Dabas, 2020). One prominent example is the various digital health apps developed specifically for diabetes patients. With the mushrooming of various apps to support diabetes patients ranging from those focusing on nutrition, to physical activity, to glucose monitoring, to insulin titration, and to insulin delivery, the danger is some of these apps may cause more harm than benefits without regular supervision and monitoring from healthcare workers (Fleming et al., 2020). However, with increasingly diverse apps that are made available and the high workload faced by many healthcare workers, the question of how they can keep abreast with all of these developments to better support their patients has become an issue. In addition, some applications may not have undergone rigorous research evaluations to ascertain their safety before deployment for general use, raising issues on patient safety (Agboola et al., 2016).

Furthermore, in the case of telemedicine, when patient care is delivered remotely, physicians or other healthcare workers may not detect subtle cues that they could detect from an in-person consultation. While digital health technology that powers telemedicine enables the health system to reap benefits that result from time and economic efficiencies, these virtual encounters could potentially undermine patient safety if safeguarding mechanisms are inadequate (Agboola et al., 2016). There is a fine line between patient empowerment and unregulated harm that these smart apps can do to patients, and health authorities and health implementers will need to keep this issue in check (Fleming et al., 2020).

An additional technical issue to ensuring patient safety is software quality. For instance, the longevity and transparency of the software that powers an app has been raised as a concern that is associated with patient safety. There is a risk to product safety that translates directly into patient safety when the software lacks maintenance and updates, which will predispose it to implementation flaws such as incorrect insulin-dosing recommendations by the manufacturers. Furthermore, when a manufacturer ceases operations, there is uncertainty as to what will happen to the technical support that was previously rendered to patients (Gordon & Stern, 2019).

Cybersecurity

A third governance challenge to the implementation of digital health technologies for public health improvement is cybersecurity. As the functions of digital health technologies are essentially configured by networking and software capabilities, this makes it vulnerable to cybersecurity risks. For instance, there have been reports that hackers were able to remotely control medical devices such as pacemakers and data from insulin pumps and modify their functions remotely (Fleming et al., 2020; Gordon & Stern, 2019). These cyber-attacks can potentially be life-threatening to patients and disrupt healthcare delivery.

One of the biggest motivating factors to such illicit means of obtaining health information via attacking digital health information systems is the profit that comes along with the extraction of these data. It was reported that the black-market value of digital health information is at least 10–20 times more than the value of data collected by credit card providers (Chernyshev et al., 2018). Extortion or demands are not always targeted at the individuals whose information were violated, but more at the health providers or authorities responsible for safeguarding their data (Chernyshev et al., 2018). The accelerating growth of smartphone use is fueling the intensity of cybersecurity attacks as well. The widespread “Android installer hijacking” that was publicly disclosed in March 2015 was estimated to impact approximately 50% of all Android users at that time. Attackers essentially replaced the Android app downloaded from third-party app stores with malware to gain access to the device and obtain sensitive data, including usernames and passwords, without users’ knowledge (Filkins et al., 2016).

In general, three major forms of cyber-attacks in the digital health space have been observed. The first is ransomware attack. One of the most famous examples of this is the WannaCry ransomware attack that happened in the United Kingdom in 2017, which paralyzed the health system for almost a week. The WannaCry cyberattacks, which resulted in the shutdown of several hospitals under the National Health Service and cancellation of about 600 surgeries and more than 19,000 appointments (Alami et al., 2019), exposed the fundamental lack of sustained investment in the healthcare industry to boost cybersecurity measures (Acronis, 2021). An impact analysis of this cyberattack reported a significant decrease in the number of attendees and admissions among the affected hospitals, which translated into £5.9 million lost in hospital activities (Ghafur et al., 2019). Such attacks are detrimental to a country’s health system and incur both direct costs in terms of revenue lost (e.g., ransom payment to the pirates/attackers and reparation payments to the affected patients, loss of income due to its inability to run health services during the period when the entire health information system was compromised) and indirect costs in terms of tarnished reputation among the public (Alami et al., 2019). The second form of cyber-attack is phishing. This attack is usually launched by sending malicious links or attachments to users via emails and tricking them into opening those attachments (Acronis, 2021). The third form of cyber-attack is privilege abuse, which emerges when an individual gains unauthorized access, raises permission to gain access rights, or uses legitimate permissions to access data or records either for personal gain or for malicious activities (Fimin, 2018). This is a tricky issue to govern as the demarcation on the right to access the health information system is often unclear. While some electronic medical record systems adhere to role-based access control policies to govern against privilege abuse, other systems may not share the same degree of granularity. Furthermore, in emergency situations that warrant open access to electronic medical records by different levels of health providers or health workers, a restrictive role-based access control policy may not be practical (Chernyshev et al., 2018).

Algorithmic Bias

Algorithmic bias has been raised as a key concern in the last decade following the proliferation of artificial intelligence and machine learning tools in the digital health ecosystem. Concerns have been raised on the dissemination of individual health data that can be used to calculate predictive health scores in the United States (Humphreys, 2020). For instance, the Affordable Care Act health risk score, which creates a relative measure of predicted health care costs for an enrollee, has been used as a proxy to suggest how sick a person is; likewise, the brand name medicine propensity score can be used as a proxy to predict how likely it is that one will choose brand-name medications over generic medications (Humphreys, 2020). These data can be traded by data brokers for commercial clients such as pharmaceuticals and insurers, who could use this information to feed the algorithms in artificial intelligence systems and conduct risk profiling or predictions that could potentially make biased decisions that are unfavorable to certain subpopulations with disadvantaged backgrounds. Besides health risk profiling, racial profiling by automated systems have also been raised. Studies have documented how machine learning and algorithms behind digital tools or internet search engines perpetuate racial biases and discriminations rather than eliminate them. This phenomenon, coined as “New Jim Code,” refers to the embedded, invisible, and injurious biases behind decision making in automated systems that are fed by biased datasets to begin with (Grande et al., 2020; Wilks, 2020).

A seminal study examining racial bias in the commercial prediction algorithm used to identify patients with complex health needs with the intention to facilitate more targeted healthcare programs for those who needed them most had revealed that when algorithmic risk score is conditional on cost, Black patients generated lower costs than White patients, but when the algorithmic risk score is conditional on health risk proxied by the number of comorbidities, at any given score a Black patient would present with more comorbidities than a White patient. The results suggest the occurrence of racial algorithmic bias, which could inadvertently result in health systems underserving Black patients when health needs are correlated with health costs without accounting for the mechanisms of health inequality that are inherent in the population (Obermeyer et al., 2019). Besides, another study also warned against the use of digital technology such as facial recognition software and predictive policing without consideration of the inherent demographic biases that have been inadvertently built into the system, especially in COVID-19 apps used for surveillance and contact tracing. Racialized groups and vulnerable subpopulations tend to be subjected to higher levels of scrutiny and suffer greater negative repercussions from racial profiling and disproportionate policing as compared to the general population. Without addressing these issues, digital health tools that operate based on big data analytics may aggravate the marginalization and stigma of the already vulnerable populations (Hendl et al., 2020).

Liability

While the issue of liability has been discussed in other autonomous systems such as autonomous vehicles (Taeihagh & Lim, 2018) and robotics for aged care (Tan et al., 2021), this issue has not received as much attention in digital health technology. In the discussion on liability, the question often revolves around which party should hold the ultimate responsibility in the event of data breaches, or when any harms or injuries to the users are inadvertently caused by a technology during its use. While logic may point to the device manufacturer as the party responsible for faulty product in hardware medical devices, liability issue in software-enabled medical devices may not be as clear cut. This is due to the fact the software embedded in the digital health device may contain off-the-shelf components that are not created by the device’s manufacturer and the software problems that surface at a later time by these off-the-shelf distributors may impact the entire pool of existing device users (Gordon & Stern, 2019). A clear liability regime that prescribes channels of redress or clear grievance mechanisms for all parties involved in the development, manufacturing, distribution, and implementation of digital health devices is thus necessary moving forward.

Autonomy

One of the most salient issues related to autonomy is a question related to access: Can patients have access to their health data for health monitoring purposes, and what is the extent to which their health data access—whether in summary/aggregate form or in more granular detail—ought to be granted? From patients’ perspectives, having the autonomy to access their health data on the digital health devices that they are using is useful for more frequent and effective self-monitoring. For instance, a heart patient with a cardiac device installed in his/her body may want to have access to his or her personal health data such as measurements of physiologic parameters, heart rate variability, or arrhythmia burden on a daily basis instead of a longer duration in order to inform their medication titration. Likewise, patients with frequent but nonspecific palpitations may want to correlate their experiences with the digital health device data more frequently than they would want to consult their clinicians. Even though these data are important to enable more effective self-management, the ability to enable more granular access of digital device data is often conditioned on the ability of the clinicians to review these data regularly (Cohen et al., 2020). The tension between autonomy of data access and ensuring patient safety may imply that there is no immediate or fast resolution to this issue.

Social Justice

Social justice is another fundamental ethical challenge that needs to be considered during the implementation of population-level digital health interventions. Specifically, it is of paramount importance to consider three issues pertaining to social justice—availability of digital health services or infrastructures that support them, affordability of these interventions, and access to technology (Brall et al., 2019). For some traditionally underserved communities and vulnerable populations such as the elderly, low-income population, and homeless people, fairness and equity will need to be addressed long before rolling out digital health solutions to a wider population. Other sociocultural barriers such as values, norms, health beliefs, and digital literacy that could decrease engagement with technology among certain populations should also be considered (Brall et al., 2019; Crawford & Serhal, 2020). Since November 2019, the issue of fairness and justice in the use of COVID-19 surveillance apps designed to control infection rates has also been raised. Research has shown that in some countries, these apps have subjected racialized groups to disproportionate policing and profiling, raising concerns on digital health technology exacerbating the marginalization of these subpopulations (Hendl et al., 2020).

The above sociocultural determinants and structural issues should be constantly addressed when implementing digital health solutions at the population level. This is so that every well-intended digital health solution will not risk widening health inequality in society, but instead fulfilling its promise to be a social leveler.