Mental health problems experienced by the working population are a global public health issue. Worldwide, more than 210 million people, representing 70% of those affected by common mental health disorders (eg, anxiety and mood disorders) are employed [1]. Several risk factors, including working conditions, workplace culture, and the nature of work, have been linked to occupational mental health [2-4]. Public health emergencies, such as the COVID-19 pandemic, are linked to specific stressors, including the threat of infection, social distancing measures, stigma, and job insecurity, which considerably increase the prevalence of mental health problems in the working population [5].

Occupational eMental health (OeMH) interventions significantly improve mental health in work settings [6]. OeMH interventions use information and communication technology, including internet- and web-based services, mobile apps, and wearable technologies, to deliver knowledge and services such as psychoeducation, workplace health promotion, psychological and medical treatment, and return to work assistance to employees [7,8]. OeMH interventions have the potential to be more available, accessible, and scalable than traditional interventions [9,10], especially in public health emergencies, leading to physical-distancing policies to contain the spread of threatening conditions such as COVID-19.

However, implementing OeMH interventions is a complex process characterized by unique challenges involving adherence to new and crude regulatory frameworks, interoperability and compatibility with existing systems and procedures, threats to employees, organizational privacy and security, and associated costs [11]. Newly introduced working arrangements in response to public health emergencies, such as the COVID-19 pandemic, could also compound existing implementation challenges and persist after the pandemic ends. Carefully developing and planning implementation strategies, which can be defined as a method or technique used to enhance the adoption, execution plan, and sustainability of an intervention [12], is therefore essential to guarantee the sustainable uptake of OeMH interventions by employers and employees.

Nonetheless, it is difficult to establish a best practice for the implementation of OeMH interventions. Implementation strategies are often inadequately documented and seldom evaluated and published [12,13], especially in comparison with studies on the effectiveness of interventions. Even when reported, implementation strategies have been discussed within a general context, and researchers have called for more tailored implementation strategies that focus on specific contexts [14], for instance health care [15]. Context encompasses the environment, broad setting, and circumstances (eg, systems and structures) in which an intervention is implemented and its associated characteristics [16]. It is a key component of several widely adopted implementation frameworks, as evident in the Consolidated Framework for Implementation Research (CFIR) [17]. Currently, those implementing new OeMH interventions are likely insufficiently informed about the procedure, strengths, and weaknesses of poorly documented implementation strategies, or uninformed about many potentially useful facilitators in this context. Furthermore, replicating positive results from similar implementations or overcoming barriers encountered in similar contexts would be challenging to achieve [18,19].

Therefore, a compilation of possible implementation strategies for OeMH interventions is critical to fostering improvements in their uptake and can serve as a reference for identifying and overcoming likely barriers and informing the future development of best practices. The objective of this scoping review was to identify implementation strategies relevant to the uptake of OeMH interventions that target employees and detail the associated barriers and facilitation measures. This scoping review would achieve these objectives by mapping the existing literature on the implementation of OeMH interventions and identifying gaps for future research. This work was conducted under the EMPOWER (European Platform to Promote Well-being and Health in the Workplace) project, funded by the European Commission, which investigates the impact of an eMental health platform aimed at preventing common mental health problems and reducing psychological distress in the workplace [20]. It is also one of the series of review papers on different aspects of the knowledge base related to the development of the EMPOWER platform.

A scoping review was conducted to identify implementation strategies relevant to the implementation of OeMH interventions and to describe related barriers and associated facilitation measures. The scoping review is an established method for assessing and mapping the extent of evidence to address and inform practice in a topic area [21-24]. The review proceeded through five stages as developed by Arksey and O’Malley [23], extended by Levac et al [22], and further modified by Westphaln et al [25] to accommodate a team-based approach: (1) identifying the research question; (2) identifying relevant studies; (3) selecting studies; (4) charting the data; and (5) collating, summarizing, and reporting the results. Accordingly, this scoping review provides an overview of the existing evidence without a formal assessment of the methodological quality. It is conducted and reported in accordance with the widely adopted PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) [26] to help ensure a high level of methodological rigor and reporting quality.

Electronic bibliographic databases, including MEDLINE, Scopus, CINAHL Complete, PsycINFO, and Web of Science Core Collection, were searched to find eligible peer-reviewed and gray literature. Search terms were based on concepts related to mental health, digital tools, the workplace, and implementation strategies (Multimedia Appendix 1). The MEDLINE search strategy (Multimedia Appendix 1) was adapted for other databases using relevant syntax and keywords in consultation with all coauthors who are also experienced researchers in the area. Hand searching of the reference lists of included articles was also completed for further relevant literature not identified during the search of databases. Members of the EMPOWER Consortium (ie, mental health researchers, clinicians, and experts focusing on well-being in the workplace) were also requested to suggest potentially eligible references via email.

Publications were eligible for inclusion if they described implementation strategies (ie, according to Proctor et al [12]) or related barriers or facilitation measures relevant to the uptake of OeMH interventions targeting employees. For example, all other eligibility criteria being met, approaches with the following characteristics would be considered: aim to introduce and encourage continued use of an intervention; prescribe actions in support of the intervention (eg, adaptations, fiscal strategies, and testing); and ensure that interventions can deliver intended benefits to the relevant organization over time, for instance, creating routine organizational policies or best practices. OeMH interventions are broadly defined here as mental health information and services delivered by information and communication technologies to employees [7,8]. This definition is consistent with the definition of eHealth [27,28], as well as the broader term digital health [27]. Studies with employed participants aged ≥18 years, that were written in English. and published between January 2010 and May 2021 were considered. Primary research studies, systematic reviews, books, and gray literature (eg, conference proceedings, theses, government documents, and professional publications) were considered. Gray literature, such as commentaries, letters to editors, and editorials, were excluded.

A total of 10 researchers (AO-T, AR, CdM, CT, CMvdFC, DM-K, KS, MdM, MTP, and RMB), including psychologists, health scientists, and health economists, were involved in screening. To ensure consistency across researchers, they attended a web-based training workshop to practice the skills needed to reliably execute screening using the web and an app-based service Rayyan, Qatar Computing Research Institute [29]. A training set of 100 publications was screened by all workshop attendees. Screening decisions (ie, include, maybe, or exclude) were reviewed and discussed to clarify any misunderstandings and identify difficulties using Rayyan QCRI. Instructions not to use the natural language processing–, artificial intelligence (AI)-, and machine learning–based features offered in Rayyan QCRI as well as tips to overcome minor usability shortcomings were given. Screeners were randomly assigned a screening set, and a screener performed a second screening of 20% of titles, abstracts, and full texts, and 100% of the publications that received a maybe screening decision. All screenings were conducted independently to reduce the likelihood of reviewer bias [30] and inconsistencies in screening decisions were resolved in reconciliation meetings.

In all, 5 researchers (AR, CdM, CT, MdM, and RMB), including psychologists and health scientists, of the 10 (50%) screeners, were involved in data extraction and attended a web-based training workshop focused on developing consistency across researchers by practicing the skills needed to reliably execute data extraction using a web-based data extraction form. The form was reviewed and improved for clarity regarding the questions asked, user friendliness, and efficiency of data entry. For instance, it was clarified that single-component implementation strategies were to be extracted, and any bundling of strategies (ie, multifaceted strategies) in publications to address a goal were to be noted. Each researcher was randomly assigned an equal number of included records, and a researcher reviewed the extracted data for all the included publications.

A descriptive synthesis was performed, where identified implementation strategies, barriers, and facilitators were collated and later summarized. The synthesis was conducted by 3 (CT, MdM, and RMB, ie, psychologists and health scientists) of the 6 (50%) researchers involved in data extraction and guided by the RE-AIM (reach, effectiveness, adoption, implementation, and maintenance) framework [31-33] and the CFIR [17] and further informed by the Expert Recommendations for Implementing Change [34]. RE-AIM and CFIR were chosen as they are widely used frameworks in implementation research (IR) [35] and were deemed by the authors to be the most comprehensive of the recently reviewed implementation frameworks [35] and most applicable to our objectives. The RE-AIM was originally developed as a framework for reporting findings regarding health promotion and disease management interventions in various settings. RE-AIM is used here to highlight essential strategy components with respect to its five steps: reach—the number of people who are willing to participate in a given initiative; effectiveness—the impact of an intervention on important outcomes (eg, individualistic and economic); adoption—the number of people or organizations who are willing to initiate and deliver an intervention; implementation—fidelity of delivery for the intervention including adaptations, costs, and consistency of delivery; and maintenance—sustained delivery and effects of an intervention after the associated initiative has ended. The CFIR unifies implementation theories to help build a robust implementation knowledge base across a wide range of studies, settings, contexts, and processes. The CFIR was used to provide a comprehensive view of multiple implementation contexts in which factors that might influence intervention implementation and effectiveness could be well detailed. Both frameworks determined the data for extraction: key publication characteristics, strategy definitions, key strategy implementation tasks, implementation processes, barriers and facilitators to strategy implementation, and any other data that holistically captured the complex and multilevel nature of strategy implementation were considered for data collection. Further synthesis of the identified barriers and facilitators produced recommendations for each relevant CFIR construct to improve the implementation of OeMH interventions.

The 31 included publications comprised 28 journal articles [36-45,47,48,50-57,59-63], 2 book chapters [8,64], and a doctoral dissertation [46] (Tables 1 and 2). Most (25/31, 81%) [8,36-47,49-59] of the included articles were published between 2015 and 2021 and were mainly primary studies (23/31, 74%) [36,39,41-51,53,54,56-58,60-63,65]. Of the 31 included publications on OeMH interventions, 14 (45%) focused on their implementation [8,37,38,40,41,49-51,53,54,62-64] and 17 (55%) did not focus on their implementation but had results or noted implications related to their implementation [36,39,42-47,52,55-61,65].

A total of 24 interventions were reported in 27 studies [36-46,49,51-56,58-65] (Table 3). These interventions were largely web-based (n=16, 67%%) [41-46,48,49,51,53-55,60-65] and most aimed to improve (n=19, 79%) [36,37,40,42-46,49,51-56,58-63] and, to a lesser extent, educate users about mental health problems. Most interventions have focused on stress-related disorders and symptoms (n=17, 71%) [8,36-47,49-51,53-63,65], but a wide range of mental health problems (eg, burnout, anxiety disorders, and substance-related disorders) have also been covered to some extent. Where reported [36,38,41,43-46,49,51-53,55,56,58-65], these interventions (n=19, 79%) largely targeted employees in professional occupations (eg, teachers and physicians). Most of these interventions were made available in specific countries, mainly in Europe (n=15, 63%%) [36,38,39,41,42,45,49,51,53-56,59,61-65], except for one that was available internationally (n=1, 4%) [64]. Standardized information about these 24 interventions, including year of launch, language, number of employees and employers interested in and who adopted the app, organizational size, and internal policies, was not clearly reported where relevant and could not be accurately extracted in detail.

The 31 included publications revealed 98 implementation strategies used when implementing OeMH interventions, 114 barriers, and 131 facilitators. The findings support observations [12,13] that the reporting of implementation strategies used for eHealth interventions is largely incomplete, nonsystematic, and unstructured. Nonetheless, the findings provide valuable insights into what is known and where knowledge gaps lie in the area.

The OeMH knowledge base does not provide definitive answers regarding the implementation strategies to adopt and when and how it is most effective and efficient to adopt them. For example, the efficacy and cost-effectiveness of using innovative methods such as web-based targeted advertising compared with traditional methods (eg, posters) to increase reach is unclear [66,67], despite the former’s success in being more time-efficient [67] and effective at recruiting hard-to-reach populations [67,68]. Those responsible for implementation must use their judgment about which of the provided strategies would be most appropriate for their circumstances. These findings support the notion that the implementation of eHealth technology (eg, eMental health [eMH] interventions) is often narrowly seen as a postdevelopment activity rather than being a crucial part of the development process [69]. Nonetheless, this could be partly a consequence of many included studies not specifically or comprehensively investigating implementation and therefore not reporting other details regarding implementation. Alternatively, publication restrictions [70] (eg, strict word limits) and the multidisciplinary nature of digital health research [71] may prioritize other study information over details regarding implementation when reporting on digital health interventions.

Similar to findings related to medical devices [72], the findings here also suggest that usability [73] appears to be the main design consideration in the evaluation of OeMH interventions, with little consideration given to other critical elements of the user experience. Findings regarding the CFIR inner setting domain highlight the need for researchers to articulate potential facilitators, including those that may have failed in one implementation context, as they might work in other contexts. Existing research [74,75] addresses many of the barriers (eg, associated with symptoms associated with mental health problems and limited digital literacy skills) categorized under the CFIR characteristics of individual domains and could provide an easy opportunity to improve implementation if given more consideration during the planning phase. Factors external to the organization (eg, external policies, partners, and competition) are known to greatly hinder or support the successful implementation of technology [76-78] but have been largely undocumented or overlooked by the included publications.

Similar to recent eMH reviews focusing on college students [79] and user engagement [80], this review also reported an increase since 2015 in eMH intervention studies meeting broad inclusion criteria. Recent reviews [79,80] also found that the eMH interventions described in the included studies were primarily web-based despite the added benefits of mobile apps that are coded for a specific mobile operating system such as iOS and Android (eg, faster, functionality-rich, and offline access) [81]. This is perhaps because web-based interventions likely cost less to develop and could be accessed via more devices if they were developed in a responsive way [81]. Emerging technologies, including AI, were considered in our search strategy, but were not used by the OeMH interventions described in the included studies. Nonetheless, this knowledge area is expected to increasingly feature the use of emerging technologies in the near future as the focus extends beyond nascent explorations of their applications for mental health and investigates the optimization of their implementation as well [82].

Based on the findings of the scoping review, four practical recommendations could be considered to avoid and mitigate the identified barriers and improve the implementation of OeMH interventions:

Future IR should continue with the broad aim of understanding what, why, and how OeMH interventions work under real-world conditions, and how to improve their implementation. The findings do not support the prioritization of any one aim over others. However, findings show that IR principles [83] such as the importance of context (eg, industry, size, and policies) and the people using the research need more attention for OeMH interventions. For example, surprisingly few findings were relevant to CFIR contextual domains (eg, outer setting), which speak to governmental regulations similar to COVID-19–related policies that have a strong influence on working arrangements. In addition, the general lack of detailed, systematic, and standardized reporting on proposed digital health interventions (eg, CONSORT-EHEALTH [Consolidated Standards of Reporting Trials of Electronic and Mobile Health Applications and Online Telehealth]—Expanded CONSORT figure) [32,84-86] and the implementation strategies used to achieve these outcomes (eg, Standards for Reporting Implementation Studies) [19,70] need to be remedied for IR to be properly used in this area. Reporting could benefit from subscribing to technology-centric frameworks (eg, the mobile health evidence reporting and assessment checklist [71] and the integrated technology implementation model [76-78]) that are more comprehensive in capturing key technology implementation factors (eg, accreditation, regulation, technology vendors, individual adoption factors, and interfacing systems). This should allow future studies to replicate and develop theories based on assessments of the implementation strategies used. In addition, any encountered or anticipated barriers and corresponding remedies that might be useful in avoiding these barriers or reducing their negative impact on implementation should also be reported. The development of an OeMH implementation checklist that includes comprehensive reporting guidelines and other prompts to ensure consistency and completeness when implementing these interventions would be beneficial. Future IR should also focus on investigating a wider range of common implementation outcomes (eg, cost-effectiveness and sustainability) [87] facilitated by implementation strategies for OeMH interventions that also target more common mental health problems in the workplace (eg, anxiety, substance use, and addiction). Issues regarding lack of digital access and digital inequity are an ongoing challenge [88], although not prominently featured in the results, and should be considered to avoid OeMH interventions contributing to any disparities. This study should also investigate how implementation strategies for OeMH interventions could benefit from emerging technologies. For instance, AI can use usage data to complement existing methods to better identify people who are at a high risk of mental health problems, support health decision-making, and offer resources that meet users’ individual health needs [89]. This could have a profound positive impact on implementation through improvements in the effectiveness and maintenance of interventions.

Search results were limited to publications in English, and a publication date restriction was imposed from 2010 onwards; however, given the broad search strategy, it is not anticipated that many, if any, potentially eligible publications were missed as a result. The term eMental was coined in 2002 [90], merely 8 years before this review’s year restriction, and a recent review of 50 publications about OeMH interventions [6] included 11 publications that were published before 2010 and none were eligible for inclusion in this study. In addition, despite our exhaustive search strategy, 6 publications from 2010 to 2015 compared with 25 from the subsequent 5-year period were eventually included. Incomplete reporting also made it challenging to detail strategies (eg, their effectiveness), barriers, facilitators, and contextual data (eg, industry, organizational size, and employee level) from the included publications and to synthesize these data later. Nonetheless, all researchers involved in data extraction completed the training specifically for this review, followed the same thorough approach, and the extracted data were reviewed at least once by a second researcher. Interrater reliability was not calculated, and reasons for disagreement in screening decisions were not reported, which might have affected the reproducibility of this study [91]. However, this does not compromise the consistency and accuracy of the screening. Moreover, two 2-hour workshops were conducted with training sessions, and reconciliation meetings were consequently held when there were inconsistencies in screening decisions.

Although multiple implementation strategies can legitimately contribute to multiple RE-AIM domains, adopting a framework with more specificity could potentially be useful for the identification of more targeted strategies. Common implementation models (eg, RE-AIM and CFIR) predate the current development of eHealth, and concerns about their inability to fully capture the complexities of eHealth implementation have been raised [69] and persist [92] despite some recent updates [32,86] and clarifications [33]. Nonetheless, these generic frameworks are useful for guiding data extraction and as tools for making valuable comparisons with other types of interventions. Similar to other scoping reviews, this review reports on the nature and features of the literature on the topic of focus and does not attempt to present a view regarding the appropriateness of the used methods and the strength or quality of evidence. Similarly, the provision of more detailed recommendations would have been premature and potentially misleading, as this was unsupported by the data collected. Further research is needed to determine valid facilitators and how they should be used in the process of OeMH development and delivery on a case-by-case basis while considering contextual factors such as industry, organizational size, employee level, and internal and external policies. Nevertheless, these recommendations could still be particularly relevant for OeMH interventions in comparison with similar interventions in different contexts. Consequently, readers should be mindful that the review cannot determine whether the included studies provide robust or generalizable findings.

This scoping review represents one of the first steps in a research agenda aimed at improving the implementation of OeMH interventions by systematically selecting, shaping, evaluating, and reporting implementation strategies. It has identified 98 implementation strategies, 114 barriers, and 131 facilitation measures related to the implementation of these interventions. A synthesis of these findings offers 19 recommendations that provide initial guidance on how to improve the implementation of OeMH interventions. This scoping review also highlighted the need to combine common implementation models (eg, RE-AIM and CFIR) with more technology-centric frameworks (eg, integrated technology implementation model and the mobile health evidence reporting and assessment checklist) to fully capture the complexities of eHealth implementation. Despite yielding less detailed insight than hoped, owing to incomplete reporting and the adoption of incomprehensive frameworks by the included publications, this scoping review’s findings can still be critically leveraged by discerning decision-makers to improve the reach, effectiveness, adoption, implementation, and maintenance of OeMH interventions.