Common psychological distress and mental health concerns are widespread across populations, affecting large numbers of people at different life stages and placing a sustained burden on well-being and daily functioning [1]. Despite the high lifetime prevalence, a substantial treatment gap persists in traditional in-person services. Multiple, interacting constraints shape whether people initiate help-seeking. At the social level, stigma and limited visible role models for successful help-seeking can suppress motivation to approach services [2]. This absence of relatable exemplars is often more pronounced in cultural or contextual minority communities where norms do not encourage formal mental health care, which can reduce perceived usefulness or necessity and weaken attitudes toward help-seeking, with stigma and the fear of loss of face further suppressing intentions to seek support [3,4]. Regarding the psychological domain, beliefs about effectiveness and low confidence in navigating and accessing the most appropriate care can inhibit first contact [5,6]. Structurally, limited-service availability, travel time, out-of-pocket costs, and workforce shortages reduce practical access, especially in rural and resource-constrained settings [7]. Together, these barriers contribute to delayed or foregone care, meaning that people who could benefit do not receive support or receive it late.

Technology, as a rapidly expanding field, has been recognized as a promising approach for enhancing access to mental health care. Digital mental health tools (DMHTs) are proposed as one such approach, offering flexible, private, and lower-cost entry points that can complement in-person services [8].

DMHTs refer to technology-enabled support for mental health information, assessment, and self-management delivered through channels such as mobile apps, web platforms, SMS text messaging services, and chatbots [9,10]. With continuing advances in digital technology, the use of these tools within health care has expanded rapidly across countries and service contexts, spanning psychoeducation, symptom monitoring, structured therapeutic modules, peer support, crisis response, and tele-mental health [11,12]. In this review, DMHTs are considered in the context of adult users seeking self-care and well-being support rather than clinical treatment.

During the COVID-19 pandemic, research observed significant increases in the uptake of digital services, underscoring their potential role in maintaining access to support when traditional pathways are constrained [13]. In that review, self-guided courses showed the highest uptake across offerings, indicating demand for direct-to-public, on-demand options. Prior research has also highlighted benefits of self-guided DMHTs, such as anonymity, convenience, affordability, and scalability, which may be especially important for first steps toward support [14].

Across populations, DMHTs are frequently described as lowering perceptual and practical barriers to early support by offering more discreet, flexible, and place-independent access to help [12,15,16]. This potential appears especially valuable where stigma and cultural pressure inhibit help-seeking. Gendered expectations illustrate this dynamic. Studies consistently report that men face stronger stigmatizing norms around seeking professional help, rooted in beliefs that managing distress independently signals strength while seeking support signals weakness [17,18]. DMHTs may soften these pressures by allowing users to engage privately and at times and locations that feel safer than in-person services, including use at home, thereby preserving a sense of anonymity and control [19,20].

In communities where collective reputation and deference to public opinion are highly valued, fears of discrimination, “losing face,” and public stigma can deter formal help-seeking [12,21]. In such settings, accessible psychoeducation delivered digitally can improve mental health literacy, clarify symptoms and treatment options, and challenge stigmatizing beliefs, which in turn can strengthen intentions to seek help [22,23].

However, the potential benefit of DMHTs can only be achieved when individuals are willing and able to initiate and sustain engagement [24]. While DMHTs offer clear advantages in accessibility and flexibility, public acceptability remains mixed. Preexisting doubts about the effectiveness or rigor of digital interventions, along with prior negative experiences, are linked to lower motivation to try or persist with DMHTs [25,26]. When users view digital delivery as inferior to face-to-face care, they tend to rate tools as lower in quality, effectiveness, and efficiency, which undermines uptake. Persistent digital exclusion among minority and marginalized communities has also been documented [27]. A lack of culturally responsive content and insufficient attention to the needs of specific ethnic communities can further undermine perceived relevance and increase the burden of engagement [28]. Privacy and confidentiality concerns, especially uncertainty about data storage and sharing practices, are also prominent barriers [12,15]. Although privacy worries are not unique to digital contexts, remote services introduce distinct data risks that many users perceive as heightened and difficult to evaluate. Confidence in navigating digital tools further impacts individuals’ willingness to initiate use and to sustain engagement. When users have limited prior experience with self-initiated help or receive minimal guidance, complex or cluttered interfaces can quickly overwhelm them, lowering their willingness to continue [15,29].

Existing reviews on the use of DMHTs have tended to focus on tool formats and design characteristics [30-32] or on outcomes such as clinical efficacy and cost-effectiveness [33,34]. Syntheses that center specifically on determinants of user initiation and early engagement remain limited. Where engagement has been discussed, psychological determinants are rarely organized within a coherent theoretical framework [12,15,35]. Moreover, although theory has long been used to explain general help-seeking behavior, theory-informed examinations focused explicitly on digital mental health services are comparatively rare [36,37].

This study focuses on initiation and early engagement and integrates a theory-driven lens by applying the theory of planned behavior (TPB) to synthesize how attitudes, subjective norms, and perceived behavioral control (PBC) operate as barriers or facilitators in DMHT contexts. The review also analyzes design and access features in parallel and explores how these features intersect with perceived control. To our knowledge, this is the first systematic review to jointly examine theoretical and practical determinants of adults’ help-seeking intentions and early engagement with DMHTs.

We position the TPB as the primary lens for interpreting initiation and early engagement with DMHTs, while acknowledging the relevance of other behavior-change frameworks. In TPB, behavior is shaped by 3 proximal determinants, including attitudes toward the behavior, perceived social expectations (subjective norms), and the sense that one can carry out the behavior given available skills and opportunities (PBC) [38,39]. These determinants influence intentions, and intentions in turn guide behavior [40].

TPB has been widely used to predict health-related behaviors, and, in the mental health field, it has most commonly been applied to understand the factors shaping help-seeking intentions and behaviors [41]. In particular, it has informed the early stages of mental health support by clarifying how willingness to seek professional help may be influenced. Early work also applied TPB to examine doctors’ intentions to provide support within mental health services, highlighting its value for understanding clinical decision-making and its potential to improve service provision [42]. To date, TPB has been widely used in mental health research to examine help-seeking across a range of mental health concerns. Across settings, studies have consistently identified positive associations between TPB constructs, particularly attitudes, subjective norms, and perceived behavioral control, and intentions to seek professional psychological support [43,44]. In digital mental health, although relatively few DMHTs have been explicitly developed on the basis of TPB, the framework has increasingly been used to inform the evaluation of DMHTs and interventions. In this context, TPB has most often been applied to better understand, predict, and support users’ initial uptake and sustained engagement with digital interventions [45,46].

Other frameworks of behavior theories have also been used in previous studies to assist mental health services researchers. These include the Health Belief Model, which centers on individual beliefs and attitudes, perceived benefits and barriers, and cues to action and is frequently used in adherence research, including medication adherence [47]. Moreover, Social Cognitive Theory, which emphasizes outcome expectations, observational learning, and self-efficacy, is commonly used to inform skills training or disease-management interventions for serious mental illness [48].

TPB is adapted in this study, as it is aligned with this review’s scope. This study focuses on adults without a prior diagnosis using DMHTs for general well-being and self-care. Early decisions to try a tool and to continue beyond initial contact are intention-sensitive moments, for which TPB provides a concise account. Moreover, TPB maps onto key DMHT concerns; attitudes capture evaluations of content credibility, privacy, and perceived helpfulness; subjective norms reflect the influence of peers, family, and clinicians as well as broader stigma; and PBC links directly to usability, cost, connectivity, and time flexibility that condition whether people feel able to engage. We read study findings through the 3 TPB constructs to clarify how participants and the original authors understand the role of beliefs, social influence, and control in shaping intentions to initiate and sustain early engagement with DMHTs.

The limitations of TPB in digital mental health contexts are also recognized. Although TPB offers a useful framework for understanding intention-related processes, it gives less prominence to affective and automatic influences (for example, mood and habit), the intention-to-behavior gap, and broader structural or logistical determinants that may shape access to care [40]. TPB is therefore used in this review as a flexible interpretive guide rather than a strict predictive model. Within this framing, design and access-related determinants, although not directly specified within TPB, would be considered in terms of how they may interact with TPB-informed factors by shaping users’ perceptions, evaluations, and sense of control and influence initiation and engagement with DMHTs.

This review synthesizes factors influencing user initiation and engagement with DMHTs through 2 connected lenses, including psychological factors directly aligned with TPB and practical design and access features that may influence or intersect with TPB-informed factors. The review is guided by two questions:

The scope of this study centers on adults who engage with DMHTs for self-care and well-being rather than clinical treatment. We consider tools intended for general mental health support rather than those targeted to a specific diagnosed condition. Consistent with this focus, we include primary studies that examine users’ experiences, intentions, and engagement with tools and exclude papers whose primary aim is to test treatment efficacy for diagnosed disorders or to evaluate clinical outcomes.

Although prior reviews have examined determinants of engagement with DMHTs [15,49], and TPB has been reviewed in relation to mental health help-seeking [41,50], to the best of our knowledge, none of the reviews have integrated a TPB-aligned synthesis of psychological determinants alongside a parallel analysis of design and access features specific to initiation and engagement with DMHTs. This dual-lens approach allows us to distinguish attitudinal and normative targets for communication from engineering, service, and policy levers that shape real-world opportunity and effort.

By mapping evidence to TPB constructs and, in parallel, to concrete product and context features, this review aims to present findings on factors that shape first use and continuation. This will inform future development of DMHTs by translating qualitative insights into actionable priorities for product design and delivery, including usability, privacy assurances, and barriers related to cost or connectivity. For policymakers, this study outlines implications for raising awareness, reducing stigma, and ensuring the resources and infrastructure needed to facilitate equitable access and engagement.

This review was designed and reported in accordance with established methodological guidance for evidence syntheses of qualitative and mixed methods research, incorporating Population, Intervention, Comparator, and Outcome (PICO) for search formulation, PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses; Checklist 1) for study screening and reporting, and quality assessment with diverse studies (QuADS) for quality appraisal [51-53]. The review focused specifically on primary research reporting end users’ experiences with DMH services and tools, enabling the synthesis of barriers and facilitators to initiation and ongoing use.

In June 2025, 9 databases were systematically searched, including PubMed/MEDLINE, Embase, PsycINFO, Scopus, Cochrane Central Register of Controlled Trials, CINAHL Complete, Web of Science Core Collection, ACM Digital Library, and Google Scholar. Search strategies combined database subject vocabulary with free text terms for digital mental health services and tools, user experience and engagement, barriers and facilitators, and adult populations. Before screening commenced, a research librarian reviewed the database coverage and draft strategy. The consultation followed the Peer Review of Electronic Search Strategies (PRESS) guideline to enhance sensitivity and precision; minor refinements were made to improve the strategy [54]. Search strings were drafted in PubMed/MEDLINE and translated to the syntax of each database; full database-specific strings are provided in Multimedia Appendix 1. Gray literature searching followed established recommendations; the first 200 relevance-ranked Google Scholar results were screened for eligibility [55].

The PICO framework was used as the conceptual starting point for formulating the review question in a health context and drew on 2 extensions that better matched this topic: Population, Intervention, Comparator, Outcome, and Context (PICOC) to make the contextual boundary explicit and Population, Intervention, Comparator, Outcome, and Study design (PICOS) to prespecify eligible study designs [51,56,57]. Operational definitions were as follows. The population of interest was adults without a recorded diagnosis of a mental health condition at the time of participation. The intervention of interest was digital mental health services and tools, including websites and mobile apps that deliver psychoeducation, self-help content, or other user-directed support. Outcomes were users’ experiences of digital services and tools, including perceived barriers and facilitators, acceptability, intention to use, engagement, and adherence. The context was early-stage help seeking and use outside clinical service delivery; services or tools designed for clinical service delivery were excluded, as clinical use of digital mental health support is outside this review’s scope. Eligible study designs were those that collected primary data directly from end users. Secondary analyses, review studies, protocols, editorials, and narrative opinion pieces were not eligible. Although review articles were excluded from the synthesis, their reference lists were screened to identify additional eligible primary studies through backward citation chasing.

Records identified through the database searches were imported into Covidence, a web-based collaboration platform for deduplication and screening. Study selection followed the PRISMA recommendations, with a PRISMA flow diagram presenting the document identification, screening, eligibility assessment, and inclusion [52]. No limits were applied regarding geography, race, gender, culture, or publication year at selection; the only limit applied during the search was the English language.

Screening proceeded in 2 stages against the prespecified eligibility criteria (Table 1). At the title and abstract stage, the first authors (NC) screened records to remove clearly ineligible studies. Articles judged potentially relevant were retrieved in full. At the full text stage, the first and second authors (NC and MKL) independently assessed each article for final eligibility. Studies were excluded at full text if they included ineligible populations such as participants younger than 18 years or clinical cohorts with diagnosed mental health conditions (marked as “Ineligible participant population” in the PRISMA diagram); if outcomes did not report users’ initial access to digital mental health care or users’ experiences while interacting with the digital service or tool (“Irrelevant outcomes”); if the intervention was not a digital mental health service or tool, including studies limited to traditional in-person services or general mental health services without specific findings on digital initiation or use (“Ineligible intervention”); or if the context was clinical service delivery, including digital tools used within formal treatment or case management (“Ineligible context”). Any disagreements between reviewers were resolved through discussion, and when consensus could not be achieved, a third reviewer (CG) aided in the resolution.

Methodological quality was appraised using the QuADS instrument, which is suitable for reviews synthesizing qualitative, quantitative, and mixed methods evidence [53]. This tool has been used in previous systematic review studies that assessed qualitative data in the field of mobile health tools [58]. QuADS comprises 13 criteria scored on a 4-point scale from 0 (not reported) to 3 (explicit and detailed), yielding a total score between 0 and 39. For each included study, we calculated a percentage score by dividing the obtained total by the maximum possible score. Consistent with prior reviews, we classified studies as poor (<25%), moderate (25%‐49.9%), good (50%‐74.9%), or high quality (≥75%). Item-level QuADS scores and study-level classifications are provided in Multimedia Appendix 2.

Structured extraction spreadsheets were used to capture study characteristics (design, setting, and country), participant characteristics, details of the digital mental health tool, and all author interpretations and participant quotations relevant to initiation, use, engagement, and continuation of tool use. Extraction was conducted in NVivo (version 15; Lumivero) to support coding and traceability of evidence to source texts, with a companion spreadsheet to organize study identifiers, raw excerpts, code labels, and evolving theme definitions (Multimedia Appendix 3).

An abductive thematic analysis was undertaken. The initial coding framework was deductively informed by the TPB to organize evidence under attitude, subjective norms, and PBC [59]. During the review, practical constraints such as the technical and logistical access–related factors emerged frequently and suggested shaping real-world use of digital tools [60,61]. With the abductive approach, data aligned with TPB were collected and sensitively analyzed while remaining open to patterns that fell outside the TPB framework. This approach is appropriate for this study because our aim was 2-fold: to examine internal factors that inform an individual’s intention to seek help through digital tools (for which TPB provides a well-established lens) and to identify practical conditions that shape tool uptake and continued use in everyday settings, which are not fully encompassed by TPB alone.

The first author (NC) coded all included evidence and wrote analytic memos throughout. The coding and theme development followed the 6 phases described by Braun and Clarke [62], including familiarization, initial coding, searching for themes, reviewing themes, defining and naming themes, and producing the synthesis. To enhance validity and reliability within a single-coder workflow, which is common and defensible in qualitative evidence syntheses, we used safeguards, including maintaining an explicit codebook with definitions and inclusion and exclusion rules, linking every code to verbatim evidence to ensure auditability, and holding regular consensus meetings, in which the second and third authors (ML and CG) reviewed exemplar extracts for each code and theme and provided feedback, leading to clarification and refinement of the coding frame [15,63,64].

For classification of determinants, we defined a factor as a barrier or a facilitator when an included paper explicitly labeled it as such or when author interpretations or participant quotations clearly described the factor as hindering or enabling initiation, use, engagement, or continuation. For example, “Another hindrance to access to DMH interventions cited is the high cost of DMH apps” [60] and “Participants also stated that it was too expensive as follows: I’m not sure I would pay $10 a month for it. when life was crazy” [65]. In both instances, costs are identified as barriers.

To aid transparency, we report simple counts of how often a code or theme appeared across the included studies. These counts describe reporting prevalence rather than importance, magnitude, or causal influence. Consistent with guidance on qualitative synthesis, salience was judged on the quality and richness of supporting evidence, the coherence of the pattern across contexts, and the contribution of a theme to explaining variation, not on frequency alone [62,66].

A total of 1668 records were identified through comprehensive searches across 9 databases, including PubMed (n=664), Embase (n=153), PsycINFO (n=168), Scopus (n=216), Cochrane Central Register of Controlled Trials (n=16), CINAHL Complete (n=8), Web of Science (n=112), ACM Digital Library (n=181), and Google Scholar (n=150; see Table 2). The search strategy incorporated key terms related to DMHTs, early-stage mental health support, and adult populations. Boolean operators and truncation were used to optimize retrieval, with minor adaptations made to fit the indexing structures of individual databases.

After removing 362 duplicates (357 through Covidence [Veritas Health Innovation] and 5 manually), 1306 records remained for title and abstract screening. Of these, 1212 records were excluded for not meeting the eligibility criteria, primarily due to lack of relevance to digital tools targeting adults in early-stage mental health support.

The full texts of 94 studies were assessed, resulting in the exclusion of 72 studies. Reasons included participants aged 18 years and younger (n=10), absence of relevant outcomes on factors influencing initial access to DMHTs (n=6), use of ineligible interventions such as traditional or in-person mental health services (n=23), general digital health services without a mental health focus (n=19), studies centered on clinical interventions (n=12), unavailable full-text (n=1), and ineligible publication types such as theses (n=1).

Ultimately, 22 [28,60,61,65,67-84] studies met the inclusion criteria and were included in the final synthesis (Figure 1).

The 22 included studies were conducted across 12 countries, reflecting a broad range of cultural and ethnic contexts, with the majority based in the United States. Sample sizes ranged from 10 to 5556 participants, with ages spanning from 18 to 107 years. Most studies used a qualitative design, primarily using semistructured interviews to collect participants’ perspectives. The DMHTs examined included web-based programs, mobile apps, telehealth services, and chatbots. The most commonly targeted concerns were general well-being, followed by depression and anxiety.

Detailed information on each reviewed study is presented in Multimedia Appendix 4, and a summary of study characteristics and key findings is provided in Table 3.

The quality and risk of bias of all included studies were appraised using the QuADS tool (see Multimedia Appendix 2). The methodological quality of the studies ranged from 54% to 87%. Of the total sample, 12 [60,68-71,73,74,76,79-81,84] studies were rated as high quality (77%‐87%), and 10 [28,61,65,67,72,75,77,78,82,84] studies were rated as good quality (54%‐74%). No study was rated as moderate or poor (ie, below 50%).

To address the review aim of exploring factors influencing individuals’ initiation of DMHTs, a thematic analysis was conducted, guided by the TPB and supported by inductive insights. This abductive analysis resulted in 5 major themes, including attitudes and individual’s evaluation, subjective norms and social expectation, perceived behavioral control, content design and framing, and technical and logistical access.

Each of the 5 themes is introduced with a brief conceptual definition, followed by the corresponding codes and illustrative excerpts, along with the number of studies in which each code was reported. The first 3 themes align directly with TPB constructs, capturing individuals’ evaluations of DMHTs, the influence of social expectations, and perceptions of control over access and usage (see Table 4). The latter 2 themes emerged inductively, highlighting practical and contextual factors shaping users’ experiences with DMHTs, influencing both initiation and sustained engagement.

Attitudinal and individual evaluations influence on initiation were identified in 9 [28,60,68-72,74,77] of the 22 included studies. Three key patterns emerged, including negative attitudes toward general mental health, skepticism toward digital health tools, and belief-based rejection of mental health support.

Four [28,60,69,72] studies reported that participants expressed negative attitudes toward seeking professional mental health support more broadly. These attitudes often stemmed from framing help-seeking as a sign of personal weakness. For example, one study noted, “People perceive seeking help for a mental health problem as a sign of weakness rather than a strength” [60]. Another participant shared, “When presented with the idea of professional help, they immediately say no or something like ‘why do you think I’m like that?’” [69].

In 3 [28,70,71] studies, participants demonstrated predetermined negative attitudes specifically toward DMHTs. These views were often rooted in low expectations of usefulness toward digital tools for managing mental health. One study found, “Many participants had low expectations for how useful the intervention was going to be” [74], while another described “strong negative attitudes toward apps (ie, apps aren’t helpful, are harmful)” [77].

Moreover, 3 [68,74,77] highlighted how deeply held personal beliefs shaped negative attitudes toward professional mental health support. Some participants doubted the legitimacy or effectiveness of preventive approaches, believing that mental health conditions like depression could not be prevented or treated, or seeking alternative coping strategies. As noted in one study, “Many participants believed that depression cannot be prevented” [70], and others expressed a preference for alternatives such as “praying” or “alcohol” [28].

Negative attitudes toward professional mental health support, often rooted in perceptions of weakness, functioned as barriers. Similarly, skepticism about the usefulness of digital tools limited engagement, with low expectations of effectiveness reinforcing avoidance. In addition, belief-based rejection, such as the view that depression cannot be prevented or the preference for alternative coping strategies, further restricted willingness to seek professional support. These findings suggest that when individuals hold unfavorable beliefs or doubts toward mental health care, whether directed at professionals, digital tools, or the concept of treatment itself, these evaluations act as barriers to initiation.

Social expectations influencing initiation were identified in 17 [28,60,61,65,67,69-74,76,77,79-81,84] of the 22 included studies, grouped into 3 subthemes, including stigma, cultural beliefs, and perceived social support.

The theme of PBC was identified in 14 [28,60,65,67,69-71,73,76-78,80,81,84] studies and comprised 4 subthemes, including autonomy and self-determination, awareness, perceived fit, and knowledge.

Content design and framing factors that were identified in 15 [28,60,61,65,68,70,71,73-76,78-80,83] out of the 22 included studies as factors shaped users’ perceptions of tool usability and emotional connection. Participants’ feedback highlighted 4 subthemes, content delivery format, linguistic style, embedded features, and human contact options (Table 5).

Technical and logistical access emerged as a major factor influencing individuals’ initiation and engagement with DMHTs, with relevant barriers and facilitators reported in 20 [28,60,61,65,67-71,73-83] out of the 22 included studies. This theme was further categorized into 6 subthemes, including ease of use and navigation, technical difficulties, cost and financial accessibility, transportation and geography, time flexibility, and privacy and confidentiality.

This review synthesized evidence from 22 qualitative and mixed methods studies on DMHTs. Findings clustered into 2 domains aligned with the review questions (Figure 1). The first comprised the TPB: attitudes (individual evaluation), subjective norms, and PBC. The second encompassed design and access features, including cost, time flexibility, connectivity, content design, and privacy (Figure 2). These domains are interdependent: design and access conditions frequently shape the same pathways specified by the TPB, especially PBC. Across studies, the same feature often functioned as either a barrier or a facilitator, depending on users’ contexts, resources, and goals, underscoring the need for context-sensitive implementation.

Across the included studies, factors aligned with the TPB were consistently reported as important influences on the initiation and engagement with DMHTs. In this review, elements classified under subjective norms were discussed most often, followed by PBC and attitudes. This pattern mirrors findings from prior work on mental health help seeking that applied TPB, where subjective norms were the most frequently reported significant predictor, with attitudes and PBC also showing consistent associations [41].

Across the included studies, features of DMHTs related to design and access shaped both the decision to initiate use and the capacity to persist. The most consistent patterns concerned costs, time flexibility and immediacy, geographic and logistical accessibility, content design and framing, and privacy and confidentiality

This review synthesized evidence on DMHTs to identify common factors that shape users’ engagement and experience. Translating these findings into practice suggests a 2-track strategy. First, strengthen willingness to seek support by addressing attitudes, social expectations, and PBC as outlined in the TPB. Second, design and deliver services that reduce effort at every step while building trust through clear communication and dependable safeguards.

Policy has a central role in this agenda. Reducing stigma and fear around help-seeking, raising awareness of available services, including digital options, and improving equitable access are critical. This includes addressing data costs, device constraints, and connectivity, which can suppress uptake among people with fewer resources. Policies provide broader efforts that normalize help-seeking among family members and the general public can build supportive norms and produce cumulative benefits over time [50,67]. Schools are another important setting. Integrating age-appropriate mental health knowledge and help-seeking skills into the curriculum can foster open attitudes among young people and improve early access to support [103]. Structural conditions, such as data costs, unstable internet access, and limited device availability, constrain use of DMHTs, particularly in low-resource settings [60]. Without deliberate mitigation, DMHTs can unintentionally widen inequities. The Digital Health Equity Framework highlights how poverty and restricted digital access can contribute to poorer health outcomes [104]. Equity considerations should therefore guide policy, commissioning, and product development to ensure digital health benefits are realized across diverse communities.

Developers can use these findings to guide product choices that support engagement. Usability needs to be tested and improved for the intended audience, including simple navigation, stable performance, and timely technical assistance. Content should be tailored and presented in a warm and supportive manner that respects diverse experiences and preferences. Trust requires more than privacy policies, and users could benefit from plain language explanations of what data are collected, how they are protected, and who can access them. Co-design with intended user groups, including people from culturally diverse communities and rural areas, can help ensure that features align with real needs and contexts [78].

Service providers can apply these insights when matching clients to tools. A key finding across the studies is that perceived fit strongly influences whether people start and continue with self-care DMHTs. Fit includes personal relevance of goals, cultural and linguistic appropriateness, and practical alignment with daily routines and constraints. Providers should assess these elements during intake and review and support users to select tools with content and features that feel suitable and relatable.

Several limitations should be considered when interpreting the findings. First, although many included studies involved participants from diverse ethnic backgrounds, most primary studies did not analyze experiences separately by cultural group. As a result, it limited the conclusions this study could draw specific to particular cultural contexts. Future reviews should prioritize designs and search strategies that enable analysis by culture, language, and migration status and should report subgroup findings in a clear and consistent manner.

Second, most of the studies reviewed were conducted in resource-limited countries that have strong and stable economies and greater access to health care and education [105]. In this review, 82% of the included studies came from such settings. Our understanding of experiences in lower-resource contexts remains limited. Factors linked to availability and access, such as connectivity, data costs, device access, and local service integration, are likely to vary across settings and may shape both uptake and sustained use in different ways. More research from low- and middle-income countries and from rural or remote communities is needed to test how these contextual factors influence engagement.

Third, there was marked inconsistency in how studies measured satisfaction, engagement, and continued use. This problem has been observed in earlier reviews of digital mental health interventions [15,106]. Inconsistent definitions and instruments reduce comparability across studies, limit the ability to detect patterns, and make it difficult to combine results. They can also bias conclusions when similar terms refer to different constructs or when short follow-up periods are used to infer sustained engagement.

This review focused primarily on user experience; most included studies used qualitative designs, such as semistructured interviews and face-to-face group discussions. These methods are well suited to explore perceptions and context, but they limit statistical generalization. The heterogeneity of designs, samples, and analytic approaches further reduced the scope for quantitative synthesis. Future quantitative studies that estimate the statistical impact of the association between specific factors and engagement with DMHTs would add value for policy and practice.

A further limitation of this review is that the included studies were primarily based on one-off interviews and survey data, with little evidence from other data collection approaches, such as ecological momentary assessment, diary methods, or broader longitudinal designs. Although these approaches were not explicitly excluded by the eligibility criteria, studies using them may have been underrepresented in the available literature within this field. The relative absence of these methods means that the present synthesis is more reflective of retrospective and cross-sectional accounts of user experience and may therefore be less able to capture fluctuations in engagement, moment-to-moment barriers to use, and how attitudes or behaviors change over time. As engagement with DMHTs is often dynamic and shaped by everyday circumstances, future research would benefit from incorporating data collection approaches that are able to capture user experiences as they unfold in context. Such approaches may provide a richer understanding of when, how, and why barriers and facilitators emerge across the course of use.

Finally, stigma emerged in participants’ accounts as a factor shaping both general help seeking and engagement with DMHTs. Prior work suggests that not all stigma constructs operate equally; for example, negative help-seeking attitudes and personal stigma show stronger and more consistent associations with reduced support seeking than some other stigma types [2]. In this review, we considered attitudes and general stigma, but differential effects of distinct stigma dimensions (eg, self-stigma, perceived public stigma, and anticipated stigma) on DMHT uptake and continuation were not examined. Future research should investigate how specific stigma types, and their interaction with cultural and contextual factors, influence adoption and ongoing use.

This review shows that engagement with DMHTs is not driven by design alone or by beliefs in isolation. Rather, attitudes, social expectations, and perceived behavioral control interact with the practical realities of cost, connectivity, device access, time flexibility, content design, and privacy. When design and access conditions support feasibility and trust, they amplify the same TPB pathways that foster intention and continued use. When they do not, they undermine them, particularly by eroding perceived control.

For policy and practice, a dual approach is suggested: (1) strengthen willingness to seek support by addressing attitudes and norms and by reducing stigma and (2) provide support to the user journey through reliable connectivity, affordable data and devices, usable interfaces, culturally and linguistically appropriate content, and plain-language data protections.

The evidence base remains constrained by concentration in high-income settings, limited cultural subgroup analyses, and inconsistent engagement measures. Future work should include studies in lower-resource and rural contexts, adopt common, theory-informed metrics with appropriate timeframes, and examine how distinct stigma dimensions influence both adoption and continuation. Advancing along these lines will enable DMHTs to deliver equitable, trustworthy, and sustained benefits at scale.