Advancing digital health equity: Directions for behavioral and social science research

Abstract

The field of digital health is evolving rapidly and encompasses a wide range of complex and changing technologies used to support individual and population health. The COVID-19 pandemic has augmented digital health expansion and significantly changed how digital health technologies are used. To ensure that these technologies do not create or exacerbate existing health disparities, a multi-pronged and comprehensive research approach is needed. In this commentary, we outline five recommendations for behavioral and social science researchers that are critical to promoting digital health equity. These recommendations include: (i) centering equity in research teams and theoretical approaches, (ii) focusing on issues of digital health literacy and engagement, (iii) using methods that elevate perspectives and needs of underserved populations, (iv) ensuring ethical approaches for collecting and using digital health data, and (v) developing strategies for integrating digital health tools within and across systems and settings. Taken together, these recommendations can help advance the science of digital health equity and justice.

Behavioral and social sciences are needed to advance the science of digital health equity and justice.

Implications.

Practice: Digital health tools are being used in a variety of settings. Additional work is needed to make sure these tools are effectively implemented and meet the needs of the intended end-users.

Policy: Research using a digital health equity framework will better inform decision makers and support technologies that promote equity and do not create or exacerbate health disparities.

Research: Approaches used by behavioral and social scientists can contribute to the digital health field by facilitating interdisciplinary team science with a focus on human behavior, whole person health, intervention development, the impacts of power and privilege on health, digital health literacy and engagement, the ethical use of health data, and real-world integration in multiple settings.

INTRODUCTION

The field of digital health is multifaceted, complex, and has grown in popularity due to its potential for improving access to health-related information and healthcare, reducing healthcare system inefficiencies, improving the quality of care, lowering healthcare costs, and providing more personalized health care experiences [1]. Additionally, the COVID-19 pandemic has augmented digital health expansion and has brought focus to digital inclusion to promote health care equity [2]. However, it is important to ensure digital health innovations are informed by the evidence base and do not increase or create new health disparities. The goals of this commentary include underscoring the strong potential population health impact of digital health approaches, while pointing to some of the risks associated with digital health research. We emphasize the importance of centering equity in all digital health research, policies, and practices, and outline five recommendations to (a) cultivate equity-focused behavioral and social science researchers and (b) improve access and use of digital health services to address disparities and promote digital health equity.

DEFINING DIGITAL HEALTH AND THE NEED FOR DIGITAL HEALTH EQUITY

What is digital health?

“Digital health” is an umbrella term that refers to a wide range of tools and technologies that can be leveraged to improve individual and population health. Multiple definitions have been put forth, from organizations such as the World Health Organization [3], U.S. Food and Drug Administration [4], and the Healthcare Information and Management Systems [5]. In this commentary, we draw on the work of Ronquillo and colleagues [1], which provides a definition of the term, subcategories that comprise digital health, key goals, and the broader vision for what digital health hopes to achieve. The authors define digital health as:

[T]he use of information and communications technologies in medicine and other health professions to manage illnesses and health risks and to promote wellness. Digital health has a broad scope and includes the use of wearable devices, mobile health, telehealth, health information technology, and telemedicine.

Ronquillo and colleagues [1] also summarized categories and goals of digital health products and services. The categories include remote sensing and wearables, telemedicine and health information, data analytics and intelligence, health and wellness behavior modification tools, bioinformatics tools (-omics), patient-physician portals, do-it-yourself (DIY) technologies for diagnosis, treatments, and decision support, among others. Key goals include improving the quality of outcomes of care and service, population health, patient experiences, physician, and other non-physician provider experiences, and addressing health disparities.

Promises and unmet expectations of digital health

The COVID-19 pandemic has been a significant turning point for digital health research and services. It has “catalyzed a seismic shift in health care delivery” [6], para. 1], and substantially changed how digital health products and services are used (e.g., [7]). Multi-sector changes included the development and commercialization of digital products, pivoting to telehealth options (either via phone or videoconferencing apps), and policy flexibilities, such as the Centers for Medicare and Medicaid Services issuing blanket waivers for requirements and participation conditions in Medicare, Medicaid, and the Children’s Health Insurance Program [8].

Despite the potential of digital health to transform healthcare and close gaps in access, quality, efficiency, and health outcomes, the benefits have been uneven. Disparities in smartphone ownership and broadband access persist [9], and the COVID-19 pandemic spotlighted the existing and increasing “digital divide,” which transcends access (e.g., [10, 11]). For example, despite the overall increase in telehealth services during the pandemic, video telehealth visits rates (as compared to audio only), were highest among young adults (ages 18–24 years), and individuals earning at least $100,000 annually who were privately insured and/or identified as White. Video telehealth visits were less frequent among persons without a high school diploma, who were ages 65 years and older, and/or identified as Latino, Asian, or Black [12]. However, digital exclusion may impact a wider range of health behaviors and outcomes, including COVID-19 incidence, mortality, and vaccination [13]. These examples underscore just a few of the ways in which persons with lower socioeconomic status, racial, and ethnic minority groups, and older individuals are systematically excluded from the benefits of digital health tools and may exacerbate health inequities [14]. Digital health options will continue to advance with increased sophistication. Thus, now is the time to ensure that rapid developments in digital health tools and resources, broadly, maximize their potential for improving access, quality, and outcomes, and avoid perpetuating or increasing health disparities [14, 15].

Equity as a guiding principle for digital health research and services

Digital health justice has been defined as “the equitable opportunity for everyone to access, use, and benefit from digital health, to achieve their greatest standard of health and wellbeing” [16], para. 3]. The problem of digital health injustice was raised prior to the pandemic (e.g., [17]). The term “digital health equity” recently emerged in the scientific literature (e.g., [15, 18, 19]), and is used with increasing frequency in both the academic and private sectors (e.g., [20, 21]). Crawford and Serhal [18] proposed the Digital Health Equity Framework (DHEF), which brings together factors of health equity, digital health equity, and digital determinants of health. Subsequently, Richardson and colleagues [22] developed the Framework for Digital Health Equity, identifying key digital determinants of health. These frameworks are important because they recognize the complexity of multi-level factors that influence health and wellbeing. Moreover, their ecological perspectives allow for an examination of the ways that power and privilege operate within and across contexts that influence health (e.g., [23]). Collectively, concepts, such as digital justice, and frameworks, such as DHEF, highlight the importance of anchoring future digital health research in equity as necessary [19] to understand and alleviate the root causes of digital health disparities and advance equitable access, meaningful engagement, and health-promoting experiences.

BEHAVIORAL AND SOCIAL SCIENCE RESEARCH RECOMMENDATIONS FOR ADVANCING DIGITAL HEALTH EQUITY

The behavioral and social sciences have been leaders in the field of digital health (e.g., [24–30]), and digital health equity (e.g., [15, 16, 18–23]). Through their focus on understanding human behavior, training in diverse theories, methods, and measures, and application of their work to important social issues, behavioral and social scientists have important roles to play in advancing digital health equity. Yet, more work remains to ensure that digital health lives up to its promise. A multi-pronged and comprehensive approach is needed to maximize the potential of digital health to reduce health disparities and promote health equity. We offer five recommendations with the goals of: (i) helping to build a community of behavioral and social science scholars for whom equity is a central tenet of their work, and (ii) driving research that tackles barriers to digital health equity and enhances the real-world impact of digital health tools. These recommendations include centering equity in research teams and theoretical approaches, focusing on key barriers to access, uptake, and usage, methodologies that elevate the voices and needs of historically underserved groups, ensuring digital health data are collected and used in ethical ways, and implementation science and planning for real-world impact at the outset of research initiatives to maximally be “of use” to society (e.g., [31, 32]).

Recommendation 1: Centering equity in research teams and theoretical approaches

Understanding and addressing digital health inequities requires diverse, multidisciplinary teams, capable of drawing on a variety of theoretical approaches and lived experiences. Equity must be a central focus—and cannot be an afterthought, add-on, or addressed via a brief consultation. Scholars in health equity-focused behavioral and social sciences are uniquely suited to complement the technology developers and big data scientists through their expertise in human behavior, intervention development, and specific physical and mental health conditions. They can investigate the psychological and social constructs that underpin health disparities and investigate strategies to promote equity. They are also trained in and can draw upon theories and frameworks that center equity and intersectionality. The use of intersectionality theory [33, 34] has been extended to many domains, including psychology (e.g., [35, 36]), technology studies [37], and digital health [38]. In the context of health, critical theories, such as intersectionality theory, are useful for exploring the ways that social positions uniquely intersect with each other, as well as systems and structures, to shape experiences of power and privilege, and in turn, influence health. It is important that scholars new to the field of health equity become well-versed in relevant theories and frameworks, be willing to examine their own positionality and biases, commit to collaboration and community building, ensure that equity is the driving force in their research, and dedicate sufficient time to do high quality and sustainable research to avoid “health equity tourism” [39]. Research should be conducted in a way that is grounded in an ongoing commitment to health equity and elevates and amplifies the scholarship of individuals and teams who have been working in the health equity space over an extended period of time [39, 40].

Recommendation 2: Focusing on issues of digital health literacy and engagement

Behavioral and social science research is well-suited to address two key challenges that have limited the reach and impact of digital health technologies: digital health literacy and engagement. These areas are important because they reflect dimensions of access, adoption, and sustained use, and contribute to what Davies and colleagues [14] have termed “digital exclusion.” Engagement with a digital technology is influenced by multiple factors, including the specific digital health tool, its purpose, context of use, as well as health literacy—all which may play a role in effectiveness [41]. These topics are also complex and lend themselves to exploration via multiple levels of analysis using a comprehensive framework, such as the National Institute on Minority Health and Health Disparities Research Framework [42] and the recently published Framework for Digital Health Equity, which is an expansion on the NIMHD framework [22]. More research is needed to systematically understand and eliminate barriers to successfully using digital health tools and create tailored solutions that enhance engagement (e.g., creating a customized, internet-delivered program for Black women with insomnia [43]). The research approaches within the behavioral and social sciences are important for doing the work of advancing digital health inclusion by generating data to inform and highlight these and other challenges and opportunities. They offer expertise in user-friendly and intuitive designs and content that promotes digital health literacy and engagement (e.g., [44, 45]), as well as looking beyond the specific digital tool to understand and address areas such as effective messaging, equitable access, and community needs [2, 10, 46].

Digital health literacy is an important precursor to harnessing the broad spectrum of digital technologies for improved health and wellbeing, and a lack of digital health literacy has been linked to health disparities among medically underserved groups (e.g., [47, 48]). van Kessel and colleagues [49], building upon the work of Paige et al. [50], defined four competence levels of digital health literacy, including: (i) functional (ability to read and write about health using digital devices); (ii) communicative (ability to control, adapt, and communicate collaboratively about health in online environments); (iii) critical (ability to evaluate the relevance, trustworthiness, and risks of sharing, and receiving health-related information through digital sources); and (iv) translational (ability to apply health-related information from the digital ecosystem in different contexts). Collectively, these four domains reflect the multi-dimensional and complex nature of digital literacy and point toward its status as a “super determinant” of health that goes beyond merely digital or health literacy alone [49]. The Digital Opportunities for Obtaining Resources and Skills (DOORS) program for teaching digital literacy and the Digital Navigator Program for training new team members to help implement digital health technology into diverse care settings exemplify work that can be done to promote digital health literacy for those who need care, as well as how providers in care settings can implement practices to overcome digital health barriers [51–53]. At NIH, the All of Us research program has also emphasized the importance of digital health literacy, and provides resources for individuals, citizen scientists, and community-based settings, such as libraries [54]. Digital health literacy can also be addressed through intervention designs that minimize complexity and provide health information via formats that leverage mobile device capabilities and reduce user burden, such as short videos and text messaging [55]. Notably, enhancing digital health literacy may also be a way to foster trust in digital health systems, a distinct but related concept in the digital health space (e.g., [56]). Until health care systems can democratize access to reliable information (e.g., [57, 58]) and digital health literacy challenges are addressed, it does not matter if digital health solutions can be widely scaled—they will still be inaccessible and unusable by those most in need.

Engagement with digital health interventions has strong potential to be addressed by behavioral and social scientists. Broadly, engagement with digital health tools, particularly mobile apps, tends to be low (e.g., [59]). Nonetheless, it is linked with health outcomes. For example, in a meta-analysis on digital health approaches to physical activity conducted by Mclaughlin and colleagues [60], digital intervention engagement, as measured by subjective experiences measures, activities completed, and number of intervention log-ins, was found to be positively associated with physical activity. However, engagement and the benefits of it are not always uniformly experienced across populations. Webb Hooper and colleagues [61] investigated engagement with an internet-based tobacco cessation program and found that all racial and ethnic minority groups were significantly less likely than Whites to create an account, and African Americans were less likely than Whites to log-in to a web-only service. Among participants of the National Cancer Institute’s (NCI) SmokefreeTXT program for smoking cessation, Black participants were more likely than Whites to complete the program but were less likely to engage with the program through weekly assessments and had lower abstinence rates [62]. Understanding how to define engagement (e.g., [63]), the mechanisms that underlie it, strategies for motivating engagement, and under what conditions incentivizing engagement with digital health tools at the systems-level are most effective [64] are important areas for consideration. Research is also needed to explore the ways that community health workers, providers, care teams, healthcare agencies, and other health-focused organizations might best use strategies, communication techniques, and digital tools to best meet the needs of the populations that they serve (e.g., [65]).

Recommendation 3: Using methods that elevate perspectives and needs of underserved populations

Research methods applied by behavioral and social scientists have the potential to amplify and integrate the perspectives and needs of underserved populations into current and future digital health innovations. Mixed methods approaches can be helpful in gaining critically important real-world, contextual understandings, and gaining a deeper and more nuanced understanding of the relevant constructs (e.g., [66]). Qualitative approaches to explore questions of how, why, and under what conditions digital health tools are used (or not used) may be of particular importance for identifying and addressing barriers to digital health uptake and engagement, and in turn, improving health outcomes. Participatory research methods can also help ensure that digital health interventions and tools are grounded in the needs of the populations for which they were created.

Co-design is an excellent approach for ensuring that researchers and all potential end-users of a digital health product are collaboratively and iteratively creating resources that are easy to use, engaging, and impactful [67–70]. For example, the University of Pittsburgh, funded by the National Institutes of Nursing Research, developed the Equity Design Thinking Educational Series, which provides a comprehensive examination of the ways in which health equity can be advanced through the research and design process [71]. Involving participants and scientists in the design process, in both research and industry settings, can shape how the health challenge to be addressed is defined, the range of possible solutions, and how success can be measured. The co-design process can also empower participants to determine the outcomes that matter most to them at the individual, family, and community levels.

Mixed methods and participatory, human-centered design approaches are time-intensive, yet offer significant advantages for advancing digital health equity. Such methods are useful for challenging assumptions, elucidating root causes of barriers, as well as increasing engagement and utility. These approaches may also potentially allow for more meaningful, engaging, and precise interventions that can reduce inequities (e.g., [72]). In short, understanding the complex and lived experiences of research participants and applying methods that thoughtfully and respectfully conceptualize challenges and solutions can facilitate the development of digital health technologies that meet the needs of underserved populations.

Recommendation 4: Ensuring ethical approaches for collecting and using digital health data, understanding limitations, and mitigating biases

Digital health technologies can generate large amounts of temporally dense data, often in real-time, and from multiple sources. These data can be incredibly useful for measuring the social determinants of health and their impact on health, when the data are collected thoughtfully and for that purpose [73]. Digital data can be aggregated from sources outside of traditional care settings and create large datasets necessary for developing algorithms to predict behaviors and recognize medical emergencies, such as the early detection of stroke (e.g., [74]). This potential for transformative healthcare raises ethical and data safety considerations to ensure that end-users are well-informed about potential risks, that missing data are appropriately characterized, and that potential biases in the data are understood and mitigated.

The NIH Common Rule [75] provides guidance on the elements that must be included in informed consent to participate in research, yet in practice, there is variability in the extent to which consent forms and processes accommodate varying participant digital literacy levels (e.g., [76]) or are designed explicitly to improve understanding and privacy protections (e.g., [77, 78]). Informed consent for digital health research requires particular attention to participant awareness and power relations to make sure that the use of data does not increase the vulnerabilities of research participants, and inadvertently those of their friends, families, or other community members via secondary data uses [79, 80]. Participants must understand how their data will be used, by whom, who owns it, and potential risks. There are efforts underway to enhance the readability and improve the overall accessibility and usability of digital health research consent forms using a human-centered design approach [81], and more work is needed to make sure that researchers and participants fully understand the technologies and how their data may be used, both intentionally and unintentionally. Behavioral and social scientists and bioethicists (e.g., [82]) can assist by developing and evaluating strategies for researchers, their institutions, and governing bodies to center equity and ensure true informed consent.

Improving equity in digital data used for health purposes requires a focus on the six Vs: virtuosity (equity and ethics of big data), volume (size of data), veracity (trustworthiness of data), variety (types of data), value (usefulness of data for decision-making), and velocity (speed with which data are collected/processed) [83]. Applying an equity lens to digital health data will assure the ability to report who is not represented in the data, how missing data are characterized, and avoid building digital health tools using datasets that are not representative of the intended end-users. Predictive algorithms should consider the diversity of the sample used to develop the algorithm, and its accuracy for population subgroups. Algorithms with lower accuracy for racial and ethnic and medically underserved groups have the potential to widen health disparities. The notion that algorithms can exacerbate racism and classism, and can cause damaging consequences, is not a new (e.g., [84–86]); however, the detection and impacts of biased digital health algorithms merits further exploration. Data missingness is also likely, and researchers should assess the patterns of missingness or whether the pattern of absent data indicate “structural missingness” [87]. Examples of structural missingness include economic hardship, type of device (e.g., older devices with fewer capabilities), participant burden, cultural competence of measures—which may explain incomplete data, rather than lack of participant interest or motivation. The types and patterns of missing data need additional attention and analyses, particularly if they differ across segments of the population (e.g., [88]).

Recommendation 5: Developing strategies for widespread adoptions and use of digital health tools within and across systems and settings

Building an excellent, evidence-based, and usable digital health tool is necessary, but not sufficient, for widespread, and equitable adoption and use over time. To date, there is relatively little research focused on how digital health tools can be most effectively integrated into healthcare settings [89] in ways that ensure equitable benefits to all patients. Dissemination and implementation science frameworks can facilitate the understanding of how to optimize and maximize the equitable use of digital technologies in various contexts.

There are established dissemination and implementation science models. The widely used Consolidated Framework for Implementation Research (CFIR [90]) considers intervention characteristics, as well as characteristics of the organization and individuals in the implementation process. CFIR has recently been extended to the implementation of digital health applications [91]. The integrated-Promoting Action on Research Implementation in Health Services model (iPARIHS [92]) applies to digital health research through its consideration of the influences of facilitation, innovation, recipients (individual and collective) and context (inner and outer) on implementation in healthcare settings [93] and within digital mental healthcare [94]. Assuring equity requires the perspectives and needs of the communities being served through community-engaged dissemination and implementation science [95]. Implementation science research must also examine how knowledge is produced, the roles of researchers in perpetuating inequities, and consider structural and systemic causes of disparities [96]. Community-engaged implementation science must be integrated into digital health research, products, and services from the outset to ensure that barriers in access, adoption, and use are prevented and/or adequately addressed. Additionally, researchers may consider partnerships within the digital health industry to help ensure that evidence-based practices are integrated into products, and services are scalable, sustainable, and actionable in real-world settings. Behavioral and social science can enhance digital health innovations in the research and design process for both technology expansion and in the start-up arena. The promise of digital health for reducing disparities and promoting equity will be strengthened by co-design processes grounded in human factors psychology [97] and rigorous study designs that evaluate both impact and scalability.

CONCLUDING THOUGHTS

In summary, the future of all forms of digital health requires critical consideration of inclusion to ensure disparities are not further exacerbated and to optimize the promise of these technologies to improve population health, including addressing current health crises (e.g., maternal morbidity, and mortality; youth mental health). Digital healthcare has the potential to be a catalyst for greater health equity [98], but technological innovations and novelty cannot overshadow the need for careful and thoughtful research on its health impacts. Approaches used by behavioral and social scientists can contribute to the digital health field by facilitating interdisciplinary team science with a focus on human behavior, whole person health, intervention development, the impacts of power and privilege on health, digital health literacy, and engagement the ethical use of health data, and real-world integration in multiple settings. Behavioral and social scientists can collaborate with health care systems to examine the impact of digital-policy initiatives to improve health outcomes for underserved populations and using a DHEF will also better inform decision makers and support technologies that promote inclusion and equity [2]. Digital health research should “avoid technocentric over-optimism” [99], and make sure that digital health technologies can increase reach and access, make healthcare more efficient and affordable, address health disparities, and seek to improve the health and wellbeing of the population. Taken together, the recommendations offered here highlight key opportunities to advance the science of digital health equity.