Over the past decade, the health care system has increasingly relied on digital technologies to educate, organize, and support people in managing their health [1]. Digital health can be described as “the field of knowledge and practice associated with the development and use of digital technologies to improve health” [2]. Patient-facing technologies such as telehealth systems, web-based services, or smartphone apps are now widely used to support the prevention, diagnosis, treatment, and self-management of health [3]. There is a large body of evidence demonstrating the effectiveness of digital health technologies (DHTs) for the treatment and prevention of various health conditions across multiple settings, including communities, primary care, and hospitals [4-6]. Furthermore, the use of DHTs has been shown to address other barriers that patients commonly experience with the health care system, such as poor access to health care providers and increased costs [7].

Despite substantial investment and an increasing number of DHTs available to health care consumers, their uptake has been limited among some population groups [8]. These include lower–socioeconomic-level groups and people from culturally and linguistically diverse (CaLD) communities, with the latter including racial or ethnic minority groups, immigrants and refugees, and First Nations people [8-15]. Studies have repeatedly reported that CaLD populations have lower uptake and use of DHTs than non-CaLD populations [8,16,17]. For example, US-based studies have found that African American and Latino people are less likely to use digital technologies for health care than White Americans [12,18], whereas other studies have shown less acceptability of DHTs among immigrants [12,13]. Indigenous and First Nations people are also less likely to use DHTs because of difficulties in access and low cultural appropriateness of the technology [19]. As such, DHTs may widen the existing inequalities within our health care system by creating a digital divide that affects the effective and equitable delivery of care [10].

This lower uptake of DHTs may, in part, be owing to a failure to meet the cultural, linguistic, or health literacy needs of these diverse population groups [20]. When designing DHTs, it is important to gain an understanding of the perspectives of the end users themselves (eg, consumers, patients, and carers) on the factors influencing their use of digital health [17]. Qualitative research approaches offer the opportunity to explore these perspectives in depth, with studies conducted among CaLD populations identifying several barriers to and facilitators of the uptake of digital health, including technical literacy, access to the internet, and acceptance of services [17]. Although this evidence is important for understanding barriers and facilitators from a consumer’s perspective, these qualitative studies are often small or focused on a single digital health intervention and, as such, may not provide representative data [21]. Furthermore, although existing qualitative reviews have explored barriers to and facilitators of the uptake of DHTs, few have explored the reasons influencing their use among CaLD populations [19,22-24].

This study aimed to review and synthesize qualitative literature to determine the barriers to and facilitators of interacting with DHTs from the perspectives of CaLD population groups. In this review, CaLD populations were defined as those who were (1) born in countries where the official language differs from that of their current country of residence or their language spoken at home is not the official language of the country where they reside, (2) First Nations or Indigenous populations (who may or may not speak English), or (3) populations who were described in studies as “ethnically/racially diverse” or “ethnic/racial minority” [11]. We acknowledge the considerable heterogeneity among these groups; however, our intention was to capture population groups that are generally underserved in the health sector based on their cultural, linguistic, or ethnic background to identify the appropriateness of digital health for these groups and highlight areas for improvement [19]. We also acknowledge a recent recommendation that First Nations or Indigenous populations be reported separately from CaLD findings in research projects [11]. However, the small sample size of many studies with First Nations or Indigenous people indicates a need for caution when drawing community-wide conclusions, as recommended in other studies [25,26]. Therefore, for the purposes of assessing this review, we grouped these populations as we hypothesized that they may share many of the same concerns in terms of barriers and facilitators regarding DHTs.

A systematic review and thematic synthesis of qualitative studies were conducted following the guidance of the Enhanced Transparency of Reporting the Synthesis of Qualitative Research framework and reporting guidelines [27].

Inclusion criteria were defined according to the Population, Intervention, Comparison, Outcome, and Study Type framework (Textbox 1). Additional inclusion criteria included publication in a peer-reviewed scientific journal in English and being accessible in full text. Protocols for research studies, book chapters, systematic reviews, and theses were excluded.

An electronic search using 3 web-based databases (Ovid MEDLINE, Embase, and CINAHL) was performed to find studies published between January 2011 and June 2022. Terms for digital health were combined with terms for cultural or linguistic diversity, ethnic minority groups, or Indigenous or First Nations people and terms related to barriers to accessing digital technologies. Ovid MEDLINE search terms are shown in Multimedia Appendix 1. Reference lists from eligible studies and systematic reviews were also searched. The results were imported into the Covidence systematic review software (Veritas Health Innovation), where each title and abstract was independently screened by 2 researchers, with discrepancies resolved by consulting a third researcher. Full texts of potentially relevant articles were then inspected for inclusion by 2 researchers based on the aforementioned inclusion criteria.

Data were extracted within Covidence using the following headings: study description (country of study and study design), participant description (percentage of female participants, mean age, and CaLD group), the type of digital health or DHT discussed and a brief description (if relevant), and results (extracted as relevant quotes from study participants and as text from the results section of each paper where authors identified barriers to and facilitators of interacting with DHT).

We conducted a qualitative thematic synthesis as described by Thomas and Harden [28]. A thematic synthesis treats qualitative findings as data for analysis [28]; these data include both direct quotes from participants and summaries and interpretation of findings from study authors. We used an inductive approach, whereby themes related to interacting with digital health were identified from within the data. We followed the three thematic synthesis steps by Thomas and Harden [28]: (1) free line-by-line coding, (2) organization of these codes to construct descriptive themes, and (3) development of analytical themes. In step 1, quotes and text were coded (“named”) according to their meaning and content. One author (LW) coded all studies individually. New codes were added as necessary as the researcher continued re-examining the data. Two authors (LW and AB) discussed the codes and revised the coding schema as needed. In step 2, similar codes were grouped together to create descriptive themes related to barriers and enablers; at this point, the synthesis still followed the original findings of the included studies. The frequency of the themes was recorded as the number of times the theme was mentioned across the included articles. In step 3, these descriptive themes were interpreted to develop inferences about barriers and enablers and implications for future research (“analytical themes”). To examine the hypothesis that First Nations or Indigenous populations share many of the same barriers and facilitators regarding DHTs as other CaLD populations, we performed a subgroup analysis using the aforementioned data synthesis approach.

Quality appraisal was undertaken using the 2018 version of the Mixed Methods Appraisal Tool developed for use in systematic reviews of mixed methods studies [29]. This tool is designed for the appraisal stage of reviews that include qualitative, quantitative, and mixed methods studies and applies 5 specific quality criteria to each study design, with results presented as a percentage score for each study indicating the proportion of criteria met. For quantitative studies, quality criteria included randomization (if appropriate), representativeness of participants, completeness of outcome data, bias, confounding, and statistical analysis. For qualitative studies, quality criteria included appropriateness of the approach used; data collection and analysis methods; and coherence between data sources, analysis, and interpretation. For mixed methods studies, criteria were related to both the aforementioned quantitative and qualitative aspects plus the rationale for using a mixed methods design and adequate integration of the qualitative and quantitative phases of the study.

Table 1 describes study and participant characteristics for all studies. Most of the included studies were conducted in the United States (17/34, 50%), Australia (6/34, 18%), and Canada (4/34, 12%), with 32% (11/34) of the studies using a mixed methods design and 6% (2/34) being randomized controlled trials, the latter including interviews with intervention participants to explore the cultural acceptability of the tool [30] and barriers to and enablers of the use of the app [31]. Nearly all studies (32/34, 94%) were conducted in the community, with 6% (2/34) conducted in primary care or outpatient clinics [32,33].

The sample sizes ranged from 5 to 197 (median 23), with an age range of 16 to 74 years. Most of the health topics considered were related to mental health (11/34, 32%), chronic health conditions (8/34, 24%), or lifestyle factors such as smoking or obesity and nutrition. In total, 12% (4/34) of the studies did not relate to any health topic. There was considerable variation in terms of the CaLD backgrounds of the participants. A total of 26% (9/34) of the studies focused on Indigenous or First Nations communities from the United States [32,40,53], Australia [31,56,60], Canada [47,57], and New Zealand [52]. Of the remaining 14 US studies, 3 (21%) were conducted solely in African American communities [35,39,46]; 2 (14%) were conducted among Hispanic or Latino groups only [30,51]; and 6 (43%) were conducted in mixed communities such as Latino, African American, and Euro-American communities [34,36,37,50,55,59]. Of the 17 US studies, 1 (6%) was conducted among Tongan American communities [61], and 2 (12%) were conducted among Korean immigrants or Korean American individuals [45,62]. Immigrants were also represented in 15% (5/34) of the studies from other countries, including Israel [20], Australia [42,54], Canada [43], and Hong Kong [41]. Refugees were represented in 6 studies: 2 (33%) from Sweden [48,49]; 1 (17%) each from Australia [33], Germany [58], and Canada [44]; and 1 (17%) conducted across Egypt, Germany, and Sweden [38].

The quality rating for each study is shown in Table 1. A total of 65% (22/34) of the studies had a rating of ≥80%, indicating that they met at least 4 of the 5 criteria for that study design. Of the studies that met fewer criteria, reasons included limited description of the qualitative approach and methodology used by the qualitative designs. For quantitative designs, reasons included small or nonrepresentative samples and limited description of the tools used to assess outcomes.

A total of 62% (21/34) of the included studies described a DHT (Table 2). Of these 21 interventions, 4 (19%) were classified as SMS text messaging services, including cognitive behavioral therapy for depression [34], health-promotion messages to new mothers [36,37], and checking participants’ use of illicit drugs [47]. In total, 29% (6/21) of the studies were classified as internet-based interventions; these included 67% (4/6) of mental health interventions [38,49,54,58], a series of skill-building videos for participants with HIV [30], and a multimedia intervention to improve sleep [58]. A total of 24% (5/21) of the interventions were classified as Android or iOS apps, including 20% (1/5) of the studies that evaluated 2 apps: a mental health intervention based on motivational intervention and a suicide prevention app [56]. A further study evaluated an app for improving mental health [60], and apps were also used to support smoking cessation [31] and teach hypertension and diabetes management [41]. One intervention was classified as a digital health survey offering screening and feedback for mental health issues before meeting with a midwife [33]. A total of 10% (2/21) of the interventions were classified as videoconference consultations; these included a telepsychiatry consultation [62] and a remote nursing assessment [35]. Other interventions included a web-based portal where patients could access their medical history and message their health care provider [50]. A total of 6% (2/34) of the studies used websites as their technology: one providing learning modules on breast cancer [46] and one tailored to users’ beliefs in relation to vaccination for human papillomavirus [51]. In total, 38% (13/34) of the studies did not include an intervention (data not shown). Instead, participants were asked for their views on mobile technology and health apps [39,61]; their web-based health information–seeking behavior and needs [40,42,52,57,59]; their use and intentions regarding telehealth [32]; their attitudes toward the use of apps in health-related research [55]; their views on health technology more broadly [43]; and their perspectives on internet-based mental health [44], social media postings [53], and technology-based cancer support programs [45].

This study identified several analytic and descriptive themes influencing access to and uptake of DHTs among CaLD populations, including Indigenous groups. We found that cultural factors affected all the identified themes to some degree and that cultural and linguistic perspectives should be considered in the design and delivery of all DHTs, with this best served through the inclusion of the target communities at all stages of development.

The dominant analytic theme identified in this review was “Using Technology,” with accessibility and digital literacy (the skills required for technology use) being the most common barriers observed across most cultural groups. Although both factors are also noted as barriers to the uptake of DHTs among the wider population [21], this review suggests several mitigation strategies for culturally diverse populations. These include the delivery of DHTs via socially normative and familiar technologies (eg, mobile phones) and the provision of culturally tailored education programs. Improving access to free internet in public spaces (eg, community libraries) may also be a useful strategy. This access should address concerns about privacy noted among some population groups such as African American individuals [65] and also the irregular availability of public internet services [10,66]. The second most frequent analytic theme was “Language,” and for many participants across a range of cultural groups, lower literacy in English or in their own language was a barrier to the use of DHTs. This finding has also been observed in other systematic reviews [21]. The use of good health literacy principles when designing DHTs, such as plain language, limited text, clear layout, and use of appropriate images, would go some way toward addressing these language issues [24]. The delivery of content through multiple modes (audio, video, and text) could also enable users to understand basic messages without language or literacy barriers [22]. The use of vernacular languages within DHTs could also be considered as this may be more understandable to many users [67]. The theme of “Culture” identified the importance of ensuring that community sensitivities are considered and highlighted the value of seeing one’s own culture accurately represented. Although this analytic theme was less frequent than others, cultural-specific factors were noted within many other themes, such as fear of technology among Greek or Italian migrants and Indigenous people, triggering of past traumas through culturally inappropriate images for refugees, and the importance of co-design with Indigenous people. Finally, trustworthiness was also a common theme, including distrust of web-based information. There is limited literature on cultural or linguistic factors influencing trust in web-based information even in relation to the COVID-19 pandemic. Some comparative studies have suggested that cultural theories may underpin differences in trust between different populations (eg, individual vs collective societies) [68,69], whereas others report that an underlying mistrust of the medical system may also influence attitudes toward web-based health information [70,71]. This is an area requiring further research, including whether DHTs that are linguistically or culturally appropriate are seen as more trustworthy and also whether alternative strategies such as having community leaders champion the use of DHTs may be effective at enhancing cultural relevance [72].

Overall, there are considerable gaps in the evidence base for factors influencing interaction with DHTs among CaLD populations, including from the perspectives of community members [8,23,73]. This is surprising given the substantial investment in digital health in recent years [8]. These gaps include lack of evidence on health literacy and digital health literacy in CaLD populations, both of which are prerequisites for the effective use of DHTs. CaLD populations have been shown to have relatively low health literacy [74,75], whereas digital health literacy is rarely reported [23,76], although some European studies suggest it is lower among migrants and non–native language speakers [77,78]. Inadequate digital health literacy is compounded by barriers related to internet access, cultural and linguistic factors, self-efficacy in using digital technology, and concerns about privacy, all of which were factors identified in this review. Disparities in digital health literacy may lead to further widening of existing health inequalities, and there is a need to ensure that digital technologies are accessible and usable by all [10,65]. Therefore, a greater understanding of the digital health literacy needs of CaLD groups is required [23].

There are also few studies demonstrating the input of CaLD community members themselves during the development of DHTs. Participatory-based approaches such as co-design may be warranted to ensure that DHTs meet the needs of CaLD groups [79]. Co-design is a process in which targeted end users, stakeholders, and researchers work together on all aspects of intervention development, from needs assessment to content development and pilot-testing. In a recent meta-analysis, co-designed interventions were shown to have positive effects on health behaviors across multiple health conditions [80]. However, co-design of health technologies with culturally diverse groups is an emerging area [39,54,81,82], and there is little evidence on involving CaLD communities in the co-design of DHTs. A systematic review of eHealth interventions targeted at socially disadvantaged groups (including ethnic minority groups) identified that most interventions did not involve members of these communities in their development [76].

A further area for future research is understanding the factors influencing the digital divide. People from CaLD and First Nations or Indigenous communities are more likely to experience socioeconomic disadvantage [11], which will only compound many of the challenges associated with the use of DHTs identified in this review and further exacerbate the digital divide [17]. However, it is noteworthy that 15% (5/34) of the included studies highlighted participants’ motivation to engage with DHTs, including a stated willingness to improve their digital literacy skills [43]. Studies also highlighted the widespread use of mobile phones. Therefore, it is important that efforts to reduce the digital divide for CaLD populations do not just focus on access to technologies or assume a lack of motivation to adopt them. Rather, a greater focus on appropriate tailoring through understanding the cultural needs and assets within communities and then co-designing solutions is warranted [17].

Finally, it would be valuable to explore potential interactions between themes to enhance the value and effectiveness of DHTs. As noted previously, culture was a factor that influenced many of the themes identified in this study; however, the findings also suggested other interactions, such as using design elements to improve both social connectedness and engagement with content and the importance of a common language in creating positive interactions with clinicians during telehealth appointments.

This is the first qualitative systematic review to focus on the perspectives of CaLD populations. We used a comprehensive search strategy and a systematic, widely used method for synthesizing qualitative data. However, the limitations of this review must also be discussed. There was no consistent definition of CaLD populations in each of the reviewed articles, which might limit the generalizability and comparison of the findings [11]. The review included a broad range of cultural, linguistic, ethnic, and Indigenous groups that are heterogeneous with respect to multiple factors. Where possible, findings were reported separately to highlight barriers and facilitators unique to each population group and those that were common across several groups. Search terms for facilitators were not included. In addition, we were unable to contact the authors of the 12% (4/34) of the studies where full text was not available. As such, we may have missed some evidence in this review.

The combination of a health system that relies heavily on a patient’s skills in navigating the largely English-based digital system, with substantial proportions of the population for whom English is not their native language, leaves a potential gap in the ability of policy makers and health care providers to deliver health care in an equitable and effective manner. As technological innovations continue to become inseparable from health care, a greater understanding of the needs of CaLD populations will identify how digital health can be tailored to support improved health outcomes in these disadvantaged groups. The inclusion of CaLD community members in the design and development of DHTs may help ensure that cultural, linguistic, and other factors are considered and addressed from the outset. In turn, this may improve the potential of DHTs to be more acceptable, appropriate, and accessible to population groups currently at risk of not obtaining the full benefits offered by digital health.