This review was based on the Arksey and O’Malley framework for scoping reviews [23,24] and was informed by contemporary guidance from the Joanna Briggs Institute [25]. The 6-stage framework includes defining research questions, identifying relevant studies, study selection, charting the data, summarizing, and reporting the results. Additionally, to increase the trustworthiness of the findings, an expert advisory group was consulted and their feedback incorporated into the final manuscript.

This review is based on a published protocol [21]. Differences between the protocol and review included the addition of a fourth research question and the change of language from the word “codevelopment” which could specify just the development phase, to the more encompassing “coproduction” which includes all the phases of research and evaluation. Finally, to provide more specificity, the word “methods” was used in lieu of “tools.” We maintained rigor by fulfilling the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) checklist (Checklist 1).

Eligible articles for this scoping review were peer-reviewed papers published in the English language. The Participant, Concept, and Context framework was used to structure the inclusion criteria (Table 1) [26].

The search strategy aimed to locate published articles and was developed in consultation with a professional librarian (CG). An initial limited search of MEDLINE and ACM Digital Library was undertaken to identify articles on the topic, from both the health and information technology literature. The text words contained in the titles and abstracts of relevant articles, and the index terms used to describe the articles, were used to develop a full search strategy for the targeted databases (Multimedia Appendix 1).

The search strategy, including all identified keywords and index terms, was adapted for each included database or information source. In total, 5 databases (Cochrane Library, Medline ALL on Ovid, CINAHL on EBSCOhost, PsycInfo on Ovid, and Web of Science) were searched from 2010 to December 2023. This period was selected as it coincided with the introduction of smartphones and extended through COVID-19 when digital technology was widely implemented into the community. The reference lists included articles, and any identified systematic reviews were hand searched to ensure completeness. Given that digital health inequities affect priority communities across high-income and low- and middle-income countries, articles from all countries were eligible for inclusion in the review. To improve the feasibility of the review, traditional imaging that has been digitized and data-driven technologies (blockchain and AI-based algorithms) were excluded as these are not routinely used in the delivery of clinical health care services.

Following the search, all identified citations were collated and uploaded into the Covidence software (Veritas Health Innovation) for removal of duplicates and management of the subsequent stages of the review screening and data extraction processes [28]. Dual and independent screening was conducted for 10% of titles and abstracts (CVR, SS, and KB). When agreement between the reviewers was equal to or greater than 80%, single screening was conducted for the remaining yield (CVR). Potentially relevant articles were retrieved in full.

The full text of selected citations was assessed in detail against the inclusion criteria by 2 or more independent reviewers (all authors). The reasons for excluding full-text articles that did not meet the inclusion criteria were recorded and reported. Any disagreements that arose among the reviewers at each stage of the selection process were resolved through discussion, or with an additional reviewer (KB or CVR).

Data was extracted from articles by a reviewer (all authors) using a data extraction tool developed and piloted for the project. The data were then checked by a second reviewer (CVR), and any disagreements were addressed by a third reviewer (KB). The data extracted included specific details about the participants, concept, context, study methods, study location (classified using the World Bank Country Classification [29]), and key findings relevant to the review questions and mapped to the digital innovation pathway through the CeHRes roadmap for the development of eHealth technologies (Multimedia Appendix 2).

Methods and approaches were identified and deductively mapped to the CeHRes roadmap for the development of eHealth technologies. To identify methods that were successfully used with priority populations, we sought to identify any “validation” of method use. Primarily this was through any reference to acceptability of the methods use, from the perspective of priority populations. However, literature also indicated we may find examples of researcher reflexivity, or the practice of critically reflecting on one’s own role, assumptions, and biases, during the innovation process [30]. This data were extracted and presented as a typology. Finally, adjustments to the research or methods to make them more suitable for the priority population were extracted from papers and thematically analyzed using an inductive process.

The scoping review identified a total of 106 articles [31-136] that met the inclusion criteria and were included in the final analysis. Our search produced 4743 articles from databases or registers and an additional 131 articles from other sources. Of the total, 650 articles were eliminated after being identified as duplicates. In total, 4224 articles were abstract- and title-screened. Of these, 439 articles were advanced to full-text review, and 106 articles were ultimately included in the scoping review. Figure 1 shows the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flow diagram, which details the reasons for exclusion.

We extracted and reported key study characteristics summarized in Multimedia Appendix 3 to support analysis of publication trends. As seen in Figure 2, the number of papers exploring digital health equity published per year trended upwards throughout the review period from 1 in 2013 to more than 20 in 2023.

Most papers were from high-income countries (n=93) [31-123], with the United States producing the most studies of any country (n=58) [31,32,35-43,47-52,56-60,62-64,66,68,69,71,73,76,77,79,81,82,84-87,89-91,93,94,98-100,105,111-120], followed by Australia (n=11) [31,32,35-43,47-52,56-60,62-64,66,68,69,71,73,76,77,79,81,82,84-87,89-91,93,94,98-100,105,111-120], the United Kingdom (n=6) [44,65,72,75,102,106], and Canada (n=5) [54,61,74,96,108]. Coproduction with priority populations in low- and middle-income countries was low compared with high-income countries, with only 14 studies across upper-middle-income (n=5) [124-128], lower-middle-income (n=7) [45,129-134], and low-income countries (n=2) [135,136]. Moreover, 1 study conducted research in both a high-income and a lower-middle-income country [45].

Across the 106 articles [31-136] reviewed, 16 priority population groups were involved in the coproduction of digital health technologies at least once (Figure 3). These categories are drawn from the United Nations Human Development Programme Report (2016), and they are analytic aids rather than mutually exclusive groups. Individuals may belong to both categories, and the overlapping identities may result in multiple experiences of marginalization. Recognizing this, we have addressed coproduction of digital health technology with people who may have overlapping identities later in this paper [137].

People marginalized by race or ethnicity were included most often (n=38) [37,40,42,43,46-48,50-52,55,58-60,63-65,68,69,75,80,82,84-87,91,93,94,99,100,102,111-114,116,118], with women, girls, and pregnant people [37,40,42,46,48,50-52,63,64,68,70,79,84,86,87,96,105,115,116,120], people in vulnerable locations [32-34,36,41,45,47,53,56-58,72,73,83,88,89,97,103,107,117,131,136], people with disabilities [35,38,39,44,55,61,65,67,71,74-78,81,83,91,106,108-110,112,113], and people of lower socioeconomic status [47,48,53-55,60,63,65,68,71,79,82,90,91,94,96,100,101,115,122,123] were each the focus in more than 20 articles. People experiencing homelessness [72], people marginalized by occupation [117,131], and people marginalized by gender identity [102,119,126] were represented least, each included within 3 or fewer articles. Some articles focused on coproducing DHTs with participants who held multiple identities (Multimedia Appendix 4). A total of 42 articles [32-36,38,41,44,45,53,54,56,57,61,62,66,67,69,71-73,75,76,78,81,83,88-90,93,95,98,102,103,106,108,109,112,114,117,131,136] focused on 1 specific priority population; however, on average, articles centered participants who held 2 intersecting identities that may have a compound effect on inequitable experiences of and outcomes from using digital health technologies (Range: 1‐4).

Across the 106 papers [31-136], all DHTs being coproduced with priority communities had a consumer-facing component. With an estimated 95 definitions of digital health in the literature and multiple frameworks describing different modalities, classifying DHTs in a succinct schema proved challenging. Initially, we tried classifying DHTs using guidance from the World Health Organization [138], the Digital Therapeutic Alliance [139], and the National Institute for Health and Care Excellence Evidence Standards Framework for Digital Technologies (United Kingdom) [27]. All proved insufficient for defining direct-to-consumer technology classification with in-depth detail.

Based on advice from our international expert advisory committee (refer to Acknowledgments section), we have developed our own classification system for consumer-facing DHTs. For direct-to-consumer DHTs, we focused on the technology type or the platforms used to deliver the technology. Key parameters for classification included the engagement mechanism or tasks undertaken by users (eg, information, digital comics, and electronic distress monitoring tool), the user engagement model describing who used the technology, the primary function the technology addressed in the health system, and the clinical condition or health domain where the technology was applied. Refer to Multimedia Appendix 5 for the full results and Table 2 for a summary of the technology type, user engagement model, and primary health function.

The most frequently cited health domain was mental health, which appeared 16 (15.1%) times across a range of contexts. This included specific subdomains, such as peripartum depression [120], posttraumatic stress disorder or bipolar disorder [38], anxiety and depression [45,86,87,101,124], and digital services [56,57,74,76,92], as well as general mental health support in relation to parenting, nutrition, chronic illness [63], perinatal health [64], and other well-being services where the condition was not specifically described [80,97]. Next most frequently cited was technology to support people living with HIV, appearing 11 (10.4%) times, highlighting prevention [51], clinic attendance and service provision [66,127], care outcomes for at-risk communities [31,62,105,119,126,128], and disease management [94,131].

Physical activity was the target of 6 (5.7%) interventions, reflecting its role as a key behavioral intervention across health domains [48,103,111,117,121,123]. Technology was also applied to diabetes management in 6 (5.7%) studies, including general management [60,71,82,85] and gestational diabetes management [68,116]. Smoking cessation was also targeted with 5 (4.7%) technologies [59,88,89,114,115]. General health appeared 5 (4.7%) times, typically indicating non–condition-specific or whole-person health promotion approaches [53,90,118,132,134]. Finally, 5 other clinical conditions or health domains had 3 technology applications each, namely, breast cancer or health [37,42,52], prenatal care [40,46,47], managing disability [67,77,78], complex chronic conditions [33,34,75], and multimorbidity [65,69,91].

A total of 58 unique research methods were described in 106 articles [31-136] across the 5 stages of the CeHRes roadmap to coproduce DHTs with priority populations. These numbers partially correspond to the number of publications per stage (Figure 4) with multiple methods used in some papers, and several papers and methods used across multiple stages.

Most methods were used in the design stage (n=30), with contextual inquiry second (n=18), and value specification third (n=16). Operationalization (n=6) and summative evaluation (n=2) were poorly represented in the results, with fewer methods than at other stages. Although most methods were present in the design stage, contextual inquiry had the best ratio of methods to studies, demonstrating that researchers are interested in and have methods to use in the early stage of coproduction before design and development. Refer to Multimedia Appendix 6 and Figure 5 for the full details of methods per stage and their descriptions.

Interviews emerged as the most used method, used in 58/106 (54.7%) studies and across 4 of the 5 stages. Focus groups were used across 4 of the 5 stages and used in 41 (38.7%) of studies. Card sorting [65,91,121], member checking [55,97], yarning [70], and vignette or scenario techniques [53,65,67,97] were used across 2 of the 5 stages. Community advisory groups was the only method used across all stages of the roadmap, demonstrating the flexibility of the approach, although it was only used in 9 studies [35,36,58,66,74,97,111,122,124].

Contextual inquiry methods focused on understanding the lives, roles, and needs of stakeholders, as well as the broader sociotechnical environment in which the technology would be deployed. Innovative techniques such as photovoice [97], home tours [65], body mapping [97], and custom-developed or coproduced surveys measuring technology use, attitudes, and barriers were used.

In the value specification stage, the emphasis shifted to identifying what mattered most to users and stakeholders. These methods helped prioritize values and translate them into actionable design goals. Techniques included workshops [36,55,84], voting rounds [95], persona development [67,123], use-case scenarios [106], and feedback on paper prototypes [107]. Notably, member checking [55,97] and community advisory input [35,74,124] were used to validate interpretations and ensure alignment with participant expectations.

Design methods aim to translate insights into tangible prototypes and assess their appeal, usability, and clarity. These ranged from basic wireframes and mock-ups [106] to more complex testing using eye-tracking software [86,114], accessibility assessments, and validated instruments, such as the System Usability Scale [36,38,41,51,73,94,98,127], DISCERN questionnaire [106], and the Suitability Assessment of Materials [106]. Participatory techniques, such as role-playing [64], digital storytelling workshops [84], and participatory design meetings, ensured user involvement remained central.

For operationalization, the focus was on implementation planning and business model development. This phase drew on stakeholder mapping [75], logic modeling [75], and surveys that explored perceived barriers and facilitators to adoption [129], in addition to maintaining continuity with earlier-phase engagement methods like interviews and community advisory groups [58].

Finally, summative evaluation assessed real-world uptake and impact, using built-in feedback mechanisms (user engagement data via Google Analytics and a Feedback button on the homepage) [44] and continued engagement with community advisory groups to monitor performance and acceptability.

In total, 13 (12.3%) studies reported acceptability of their methods when coproducing with priority populations. Acceptability in this context is defined as any subjective or objective evidence to understand if the method was effective and acceptable for the priority population. A total of 5 distinct techniques were identified, varying in formality and depth, ranging from experience surveys and evaluation via comparison with current best practice [36] to researcher reflexivity [91,104,110] and assessment using a culturally led framework [70].

The most used validation technique was community consultation, where the research methods were reviewed and adjusted by the priority population before the research was conducted. Table 3 outlines each technique, provides examples of how it was applied, and lists the corresponding references. Together, these techniques reflect a diversity of strategies used to assess methodological acceptance in digital health innovation involving priority populations.

Across the 106 papers [31-136], 17 methodological adjustments were implemented to ensure the coproduction process was inclusive, culturally safe, and responsive to participant needs. These were inductively themed into (1) participant safety and comfort, (2) enhancing literacy, (3) peer-support, and (4) social, cultural, and demographic considerations (Table 3).

In total, 40 (37.7%) studies used at least 1 research adjustment. Henson et al [70] used the most adjustments (n=5) to identify how older Aboriginal and Torres Strait Islander women use DHTs to enhance their own health: using yarning rather than interviews or a focus group, taking a strengths-based approach, peer-led recruitment, a peer researcher, and using an Aboriginal project governance group. Enhancing literacy was a key consideration when adjusting for priority population inclusion. Out of the 38 studies, 27 (71%) that used adjustments improved involvement in the coproduction of DHTs by conducting the methods in the language of the priority population (n=25), adapted methods or instructions for low literacy levels (n=6), and/or providing paper alternatives (n=1). Refer to Table 4 for the full results; some studies used multiple methods.

Multiple systematic, scoping, and narrative reviews have focused attention on the opportunity of digital health innovations to reduce inequity and improve health outcomes for priority populations. Evidence is mounting that coproduced technologies, where users and community partners are consulted throughout the research, development, and evaluation process, are more acceptable to priority populations than those created without them [4,36,53,141,142]. The aim of the scoping review was to systematically identify and assess published methods and approaches used during digital health innovation to promote equitable inclusion of priority communities at every stage of the CeHRes roadmap, from conceptualization to evaluation [21,22]. Existing reviews demonstrating the practical application of equitable methods have largely focused on the design stage of technology development [16-19]. Therefore, we advance and expand the practice of including priority communities in digital health innovation across the whole innovation pathway by cataloging 58 practical methods, 4 approaches, and 17 adjustments. Indeed, for professionals working across health care, health informatics, digital health, and technology development, this review provides practical, implementable strategies to center equity throughout the entire technology innovation pathway [143-145] in Multimedia Appendix 7.

Coproduction of digital health innovations with priority populations increased between 2013 and 2023. This review demonstrates the early-to-mid stages of innovation, particularly the design stage, encompassed the most studies and methods. This is concordant with the proliferation of literature on co-designing DHTs [146]. Additionally, we show that both during conceptualization and when specifying value, priority groups can be included, enabling them to be involved before an innovation is designed or developed into a solution. This aligns with recent advice from research funding bodies to create authentic public and patient partnerships from project inception in order to improve research relevance, enhance research impact, strengthen the research process, and increase trust and engagement [147].

Although the CeHRes roadmap originates from Western research traditions, its application does not inherently reinforce inequity when priority communities are meaningfully engaged from the earliest stages of innovation. Involving priority communities from project inception—particularly during contextual inquiry and value specification—allows their knowledge systems, priorities, and lived experiences to shape the development of digital health technologies. In this way, the CeHRes roadmap can function as a flexible analytical structure rather than a prescriptive framework, supporting inclusive coproduction rather than imposing external assumptions.

We also note that fewer studies and methods focused on the implementation and evaluation stages. Although many digital health implementation papers focus on the clinical outcomes for priority populations, they fail to include these populations in the interpretation of the results. This is a clear gap and should be a focus of future research. Without a situated narrative and an understanding of implementation complexities that influence the equitable uptake, scale, spread, and sustainability of these digital health innovations, there is a risk they will not achieve the proposed equity aims [148].

Encouragingly, all priority populations defined by the United Nations Human Development Programme Report (2016) were featured in at least 1 paper [31,32,35,37,40,42,47,53,61,70,72,78,101,102,131]; however, as most studies were conducted in high-income countries, it is unclear to what extent these methods could be applied to other socioeconomic settings. Even within high-income countries, people experiencing homelessness, people marginalized by occupation (eg, sex workers), people marginalized by gender identity (eg, trans people), and Indigenous peoples were underrepresented. This research supports the use of adjustments to enhance participants’ research experience when coproducing technology. In some of these cases, the presence of multiple intersecting identities led to a greater number of research adjustments. Based on this review, the lack of consideration of intersectionality and how it applied in digital health coproduction is a key concern. Future research should recognize multiple identities and, therefore, multiple disadvantages, using research adjustments to overcome these disadvantages.

Priority populations’ health outcomes are driven by colonization, structural inequality, and ongoing discrimination, not biology [149]. To better understand these results, in the next section, we have reflected on them from an Indigenous Australian perspective through the lens of decolonization.

Decolonizing research in digital health requires a critical interrogation of the colonial structures that shape research processes, from topic selection to dissemination, and a deliberate recentering of Indigenous ways of knowing, being, and doing [150]. Decolonizing research involves “evaluating and dismantling and disrupting Western ways of knowing”—in this case the foundations of research practice as it stands [151]. In reviewing studies pertaining to Aboriginal and Torres Strait Islander communities, this scoping review identified multiple adjustments, methods, and approaches, such as Aboriginal project governance, yarning methodologies, peer researchers, strengths-based frameworks, and active community participation. Categorizing these elements proved challenging, as many could be simultaneously understood as an approach, an adjustment, or a method, and were often applied across multiple stages of the CeHRes roadmap. This difficulty in classification reflects the tension between Indigenous research practices, which are relational, iterative, and context-specific, and Western research systems that prioritize standardization, rigid categorization, and linear progression. As noted by Laycock et al [152], Indigenous research methodologies emphasize Indigenous control of the research agenda, alongside respectful processes for consultation and negotiation. These principles resist prescriptive methods and instead demand flexibility to meet local community priorities. However, fitting such approaches into frameworks developed within colonial systems risks reducing them to technical steps rather than honoring them as holistic practices.

There is growing evidence that recentering Indigenous ways of knowing, being, and doing strengthens research quality and relevance for First Nations peoples [153]. This is particularly critical when working with data relating to Aboriginal and Torres Strait Islander communities, whether sourced indirectly through literature reviews or generated in direct collaboration with communities. Many of the adjustments, methods, and approaches identified in this review are integral to culturally safe and methodologically robust Indigenous health research. However, embedding these approaches within the rigid frameworks and protocols shaped by colonial research systems remains challenging. While the principles of Indigenous health research provide essential guidance, they are not prescriptive, and methods must be responsive to local contexts, cultural protocols, and community priorities. In contrast, reliance on colonial (“Western-centric”) and institutionally governed research models risks marginalizing Indigenous priorities and knowledge systems, leading to epistemological mismatches.

Drawing on over a decade of public health research across New South Wales, Australia, Clapham [154] observes that these approaches often fail to align with Indigenous governance structures, values, and aspirations. These mismatches are compounded by differing modes of knowledge sharing—oral versus written traditions, distinct ways of doing, and diverse knowledge systems [150]. This epistemological disjunction is a global concern. Research frameworks within settler-colonial states routinely exclude Indigenous epistemologies, effectively coercing Indigenous scholars into Western paradigms, undermining epistemic justice and sovereignty [155]. Data practices exemplify this mismatch—colonial data mining often overlooks Indigenous frameworks, giving rise to the data decolonization movement that prioritizes Indigenous paradigms of data governance and self-determination [156]. Although this scoping review actively sought First Nations–specific literature, the processes used to identify, search for, and categorize codesign methods for digital health technologies may not have fully applied an Indigenous lens, highlighting the structural constraints of current research paradigms.

Previous studies have identified structural barriers, such as inadequate infrastructure, poor integration with existing clinical workflows, user-unfriendly health information exchange interfaces, and a lack of trained personnel [157]. At the patient level, barriers include limited broadband access, lack of access to digital tools, and insufficient cultural or linguistic appropriateness of digital health solutions. Additional factors influencing adoption include awareness of digital health solutions, perceived benefits versus burden of technology use, accessibility, trust in digital health technologies, and user experience [158]. These findings highlight how failing to include priority populations in the design of digital health technologies may contribute to lower adoption and reinforce existing health inequities.

Future research should explore the use of existing and new methods focused on the implementation and evaluation stages of the CeHRes roadmap enabling a more relevant interpretation of the results generated from technology trials. Future research also needs to take into consideration the multifaceted identity of people, developing methods, and adaptions that take into consideration multiple disadvantages and intersectionality. The overlapping social positions and identities held by participants in this research may indicate the need for a specific intersectional theoretical framework to consider how digital health outcomes are influenced by interlocking systems of privilege and oppression [159]. Despite some attempts to validate methods, only 13 (12.3%) studies reported conducting any acceptability review of their methods [36,48,70,74,84,91,97,104,110,111,119,124,129]. Thus, future work should explore using the 5 identified techniques and how we can measure or monitor the impact of equitable approaches in research and development contexts.

There are several strengths and limitations to this review. Key strengths include multiple measures to ensure rigor, such as the involvement of authors from diverse disciplines and from high-, middle-, and low-income countries. The review was guided by a published protocol, incorporated a comprehensive search strategy designed and executed by a specialist librarian (CG), and used dual screening of full-text articles. Data extraction was conducted individually, with verification by a second reviewer (CVR), with disagreements addressed by a third reviewer (KB). Furthermore, the project benefited from oversight by an expert advisory group.

Limitations include the lack of studies from very marginalized priority groups and papers from low- and middle-income countries, and thus we did not analyze papers based on individual priority populations. This was likely because of only English language studies being included. While searching for relevant literature, the information specialist performed searches in 5 mainstream databases, as is common practice. Additional items were identified by citation tracking and searching gray literature sources. However, this search process was faulty because it did not fully compensate for the poor identification and indexing of health literature related to priority populations in the mainstream health databases [155]. As a result, the search did not retrieve 2 relevant articles that had been recommended by one of the authors (AMI) in our review team [122,123]. These 2 items were subsequently included and have improved the review’s coverage for and relevance to specific populations. In the PRISMA flow diagram, the category “References from other sources” could be expanded with an extra dot point: “Items recommended by expert advisors.”

This important lesson prompts us to be mindful of equity issues when conducting literature searches. Moreover, 2 extra resources that may merit wider use in Digital Health Equity research are the proprietary Emerging Sources Citation Index and the open-access website Directory of Open Access Journals. Research should broaden database coverage beyond the dominant biomedical indexing systems and explicitly include Indigenous and regional journals where indigenous research may be more visible. Search strategies should also account for the diverse ways equity concepts are described, moving beyond narrow keyword matches to incorporate culturally grounded terminology. Such steps will reduce the risk of excluding relevant Indigenous-led studies and strengthen the inclusivity and relevance of digital health equity reviews.

Involving priority communities across the lifecycle of DHT innovation is vital to avoid exacerbating inequities and is possible using the practical methods identified in this research. The most interesting and appropriate methods we identified were developed by priority communities themselves. Thus, we recommend that all digital health projects aiming to provide technology for priority populations include the intended end users in research and development from the start, taking on an appropriate knowledge lens (eg, decolonization) and using the adjustments requested by the community to enhance their coproduction experience. An effective way to achieve this is with a community advisor or advisory group. There is also a clear opportunity for educators and certifying organizations to incorporate these approaches into health institutions’ core curricula for digital health competencies addressing structural inequity.