A cross-sectional study was performed to characterize mHealth platforms developed for depression or anxiety that were available for commercial, research, or both purposes. The mHealth platforms included apps, websites, and web-based software. A 2-pronged approach was used in this study. This included systematically searching relevant literature that described mHealth platforms as interventions developed for research and simultaneously searching 2 major app stores systematically to identify commercially available platforms. The protocol was registered on PROSPERO (International Prospective Register of Systematic Reviews; CRD42020193956) after preliminary searches were conducted. The piloting of inclusion and exclusion criteria and formal screening were conducted after the official PROSPERO registration.

Descriptive statistical analyses were performed for all the variables. The categorical variables were presented as absolutes and relative frequencies. Statistical analyses were performed using Microsoft Excel (version 16.43).

RCTs are clinical evidence and a prerequisite for regulatory approval. However, not all platforms had RCTs. Additional regulatory information—RWE and regulatory approval—was searched. In the secondary search mentioned in Table 1, a subset of apps had a credible source that offered some form of empirical evidence. In the app store searches, 3.4% (6/179) of the apps included scientific evidence claims, whereas 2.8% (5/179) of the apps were affiliated with health care experts and reputable institutions.

None of the apps in both literature and app store searches were reported as medical devices. In the literature search, 44.4% (75/169) and 45% (76/169) of articles presented evidence from RCTs and RCT protocols (Table 7), respectively, whereas the remaining 10.7% (18/169) did not provide evidence from RCTs. Further analysis of the clinical and regulatory evidence—through the search for additional RCTs, RWE, and regulatory approvals—demonstrated that 41.4% (70/169) of platforms presented evidence from RCTs, 0.6% (1/169; Wysa) presented preliminary RWE, 1.2% (2/169; myStrength and Moodgym) presented evidence from RCTs and RWE, and 0.6% (1/169; Deprexis) presented evidence from RCTs and RWE and had obtained regulatory approval.

From the app store search, a subset of apps was pinpointed to be based on substantial validation studies to assess effectiveness. In particular, RWE can serve as a key gateway and vitally important indicator for supporting health care decisions, as well as impact and value toward patient care and the broader health care ecosystem. For example, 1.1% (2/179) of the apps provided evidence from each of the following sources: RCTs (Sanvello and Moodmission); RWE (Talkspace and Wysa); RCTs, RWE, and regulatory approval when applicable (Flow and Sooma).

This review provides an in-depth analysis of the current mHealth platforms developed for anxiety or depression available for research and commercial purposes. Through a novel combined approach—a 2-pronged systematic search across the research literature and marketplace—we identified 169 and 179 mHealth platforms, respectively. Most platforms developed for research purposes were designed for depression (116/169, 68.6%), whereas the platforms made available in the app stores were mainly developed for anxiety (Android: 58/179, 32.4%; iOS: 27/179, 15.1%). We identified that mHealth characteristics such as platform name, aspects of mental health addressed, purpose, type of intervention, type of technology, and level of evidence varied largely across these platforms. In addition, we highlighted a significant lack of adoption of the regulatory framework and standards governing these mHealth platforms. An overview of the existing platforms for anxiety and depression available for research, commercial use, or both, as reported in this review, allowed us to identify gaps in the literature and marketplace and subsequently provide guidance for recommendations for future frameworks of mHealth development and regulations.

Among the abundance of mHealth options currently available for anxiety or depression, relatively few have been developed for youths and emerging adults [167,168]. This is surprising, as these age groups are among the most prevalent users of mobile technologies as well as the groups that are most in need of mental health interventions. In particular, typical engagement with standard care for mental health has low appeal to young people and can be antagonistic to their developmental needs [168]. The low motivation inherent in emotional disorders; the need for agency, anonymity, and privacy; the greater sensitivity to stigma and discrimination associated with mental health; and the development of self-monitoring among young people can be addressed by mHealth approaches that provide personalized and timely treatment [169,170]. Future developers of mHealth platforms for anxiety or depression should consider developing them with this targeted group in mind.

The main purpose of most of the mHealth platforms for anxiety or depression identified in this review focused on treatment. The treatment used features such as module-based sessions and the teaching of coping and management strategies, which were based on the principles and guidelines of CBT, problem-solving therapy, and psychoeducation. Previous research has revealed 3 primary areas for enhancing the effectiveness of the treatment function of mHealth [167]. First, mHealth needs to be complemented by human interactions, especially with a medical physician or health care worker, to increase compliance, adherence, dropout [2,171,172], and efficacy in treatment [173-175]. However, findings from this review highlighted that few mHealth platforms included human interactions. To be effective, the digital framework for human interaction needs to consider individuals’ desire for anonymity and privacy. Second, initial engagement and sustained use of mHealth platforms for young people require some form of social support from peers, schools, or mental health professionals. Although the role of social support and the associated outcomes are similar to that found with the standard of care, the distinct feature of digital social support that mHealth platforms provide is anonymity and accessibility. The opportunity to share similar lived experiences of anxiety or depression with one’s peers and being accountable to an authority (eg, health care professional) for the purpose of monitoring or assessment of their symptoms has the potential to increase their engagement and adherence to the mHealth platforms. Notwithstanding the support function of mHealth platforms, few platforms included this in their design feature. Finally, an mHealth platform needs to be easy to navigate, relatable, engaging, and esthetically appealing in providing education for mental health users—another common function of mHealth platforms. Previous studies have found that the boring and repetitive nature of modules and lessons presents a barrier to greater engagement with the education feature of an mHealth platform [10].

mHealth can be deployed via one or a combination of mobile technologies. An increasing trend, which is also observed in health care, is to reach the individuals by communication technologies that are already in use for the seamless incorporation of the intervention into their daily life. Our literature search reported web-based platforms (90/169, 53.3%), mobile apps (57/169, 33.7%), and messaging (12/169, 7.1%) as the most common modalities deployed, which is consistent with findings in other studies [176]. In our literature search, 13% (22/169) of the platforms used ≥1 type of technology.

The app stores search identified discrepancies in the number of apps available in each app store. The higher number of apps in the Google Play Store (140/179, 78.2%) than in the Apple App Store (61/179, 34.1%) can be attributed to a preference for the Android operating system over iOS in Asia (82.7% vs 16.5%) and Singapore (65.2% vs 33.9%), as recorded in June 2020 [177]. Although the infrequent download count data from iOS disables the comparison in this category in our review, the literature suggests that in the United States, the iOS preference is reversed (41.7% vs 58.2% for Android and iOS, respectively) [177], and iOS apps show a higher number of downloads of behavioral health mobile apps [178]. Interestingly, a recent study collecting location data has shown that the platform type had a measurable effect on the retention of passive data collection because of the increased battery consumption of the app for the iOS system. In addition, the operating systems differ in their battery-saving modes and memory space–saving strategies, for example, app offloading to a cloud, which may have further implications. In the same study, no differences were detected in active data collection retention, suggesting that the operating system had no effect on active user interactions with the mHealth platform [179].

The deployment technology type and its specific technical considerations (eg, operating system) can make or break an mHealth platform. A digital intervention design should carefully consider the context of how each technology is already being used, as well as its technical and behavioral limitations. An mHealth platform can be deployed either via one technology type only or in a hybrid model to reap the benefits of each technology type to maximize usability with sustained functionality toward the highest efficiency of the intervention.

Of note, only 4.7% (8/169) of the mobile apps identified in the literature search were available in app stores. This can be potentially related to the geographical constraints of the app store search compared with the unconstrained geography of the literature search. In addition, the literature searches covered articles from inception to 2020, whereas the app store search included only the apps present in the marketplace in June 2020. It is not unusual that after the gathering of evidence, the apps become unavailable in the marketplace [178,180,181]. The lack of commercial translation of research published in literature can be attributed to multiple factors: funding timelines (eg, relatively short grant duration and high staff rotation may not enable long term app management) and downstream research objectives (basic or fundamental research, tool development, and proof-of-concept to support potential future commercial and large-scale deployment) [182].

From the perspective of the marketplace, 1.7% (3/179) of apps in the app store search were described in a peer-reviewed article identified in our search. This can be potentially attributed to a business strategy that favors quick-in and quick-out rather than laborious and costly app evaluation without prevailing guidelines or incentives—monetary, regulatory, or otherwise—in doing so. Although consumers make the initial decision of which app to download based on easy-to-judge attributes such as graphics, price, and rating [178,183], it is the credibility and trustworthiness that drive engagement with mental health apps [184-186]. With this realization, it is important for commercial app developers to generate scientific evidence that validates their apps to substantiate the scientific descriptions, which will collectively enhance the app’s appeal. In a study of the app descriptions of the top-ranked apps, it was previously found that apps use scientific jargon and are not supported by principles backed by peer review and validation studies, or they do not align with established findings in the scientific literature [187]. Instead of objectively informing about the evidence, the function of the app description is to appeal to consumers and drive the number of app downloads, independent of the platform’s objective quality [187]. Accordingly, app user ratings do not indicate clinical utility or quality [188,189].

Clinical evidence, in the form of RCTs, reported in the publications found in the literature search (151/169, 89.4%) vastly exceeded that for apps found in the app store (2/179, 1.1%). Regulatory approvals (0.6% vs 1.1%) and RWE (0.6% vs 1.1%) were more balanced between platforms from the literature and app stores. The relatively high frequency of RCTs but low numbers in RWE and regulatory approvals for platforms featured in the literature search seem to indicate a pre-eminently investigative intent of those platforms. In addition, 33.7% (57/169) of the platforms were available as mobile apps, further hinting at treatment commercialization as a secondary possibility. Maintenance and update of the website platforms were, in some cases, not frequent, indicating ad hoc use.

A large proportion of the mHealth platforms from the app store were unregulated, clinically untested, and unevaluated for real-world effectiveness (173/179, 96.7%), indicating a potential need for business models that support investment in robust clinical or real-world validation. Importantly, although 1.7% (3/179) of the apps found in the app store search in this review could claim to have scientific evidence or regulatory approval, 15.1% (27/179) and 10.1% (18/179) of the mHealth platforms for anxiety or depression extracted from the Google Play Store and Apple App Store, respectively, were included in the Medical category. Under this category (excluding telehealth), 8.4% (15/179) of the platforms affirmed that health professionals were involved in the design of the apps. From the app stores, 8.9% (16/179) of the platforms included a disclaimer notifying that the app was not intended to be a replacement for treatment or any sort of medical intervention. Another 12.3% (22/179) of the apps mentioned information sources used to design the app, such as validated tests (eg, Patient Health Questionnaire-9 for the assessment of depression and Generalized Anxiety Disorder-7 for the assessment of anxiety), international organization guidelines (eg, World Health Organization), validated therapies (eg, CBT), or peer-reviewed articles. These numbers reflect the need to develop more actionable and meaningful app store categorizations for medical apps. In addition, increased app store categorization oversight and strategies to enable or incentivize owners to implement changes, as well as clearer accountability guidelines pertaining to store content quality, may be needed. For example, oversight could be extended to app store owners to assist in enhancing specificity regarding the medical classification of apps. In addition, greater user education in the selection of apps may be a downstream consideration for health authorities. Consumers may need to be able to discern real-value apps by analyzing their claims, the involvement of health professionals, regulatory status, and scientific evidence. Apps that have not been supported by established validation guidelines adversely affect the field and users in multiple ways. For example, they may add a burden to already strained health care systems. In addition, they may create opportunity costs when the technological promise is not realized, resulting in mHealth fatigue of the clinicians, patients, and consumers, thus delaying the implementation of evidence-based, efficient mHealth platforms.

Overall, a need for enhanced regulation, especially in the Medical category of the stores, indicates that authorities may be able to positively affect the mHealth landscape toward requirements that include threshold validation and other applicable compliance guidelines. Several attempts have been made in that direction from various government organizations and professional societies. For example, the Institute of Electrical and Electronics Engineers—an engineering professional society—developed the Institute of Electrical and Electronics Engineers 1752 standards in 2017, which define specifications for standardized representations for mHealth data and metadata for cardiovascular, respiratory, and metabolic measures [190]. Health Level Seven International—an international mHealth professional workgroup—released the Consumer Mobile Health Application Functional Framework in 2018 with the primary goal of providing a standard whereby a mobile app’s characteristics can be assessed [191]. In 2020, the American Psychiatric Association—a professional organization of psychiatrists—developed the American Psychiatric Association App Advisor, an initiative to identify an app evaluation model that was specific to mental health to be used by clinicians and their patients [192]. Most recently, in February 2021, a joint unit formed by the National Health Service England and the Department of Health and Social Care—released the Digital Technology Assessment Criteria with the aim to set the expectation of digital health technologies for potential developers [193]. In addition, other digital health and medicine professional societies, such as the Digital Medicine Society and Digital Therapeutics Alliance, also have initiatives to develop frameworks, standards, and resources for the different stakeholders of digital health interventions [194,195]. Furthermore, the International Organization for Standardization Technical Specification, in collaboration with the European Committee for Standardization, is currently developing a health app quality label ISO/TS 82304-2 Health software–Part 2: Health and wellness apps–Quality and reliability [196,197]. Although the continued development of these regulatory frameworks is critical, to have a real-world impact on the quality of mHealth platforms and apps, they must also be widely adopted by developers and requested by regulators, marketplace owners, and target populations alike. This will ultimately result in the development of high-quality mHealth platforms with stringent safety and security measures required in health care.

Although we took a novel approach of combining both a systematic literature search and app store search, some mHealth platforms may have been excluded. For example, although we had no geographical filters in our literature search, our commercially available app search was based on regional availability in Singapore’s Google Play Store and Apple App Store. Although this may be a region-specific search, this review also provides new insight into the field as it serves as an early resource detailing available mHealth apps and platforms in South-East Asia. As Singapore is a globally recognized regional information communications technology hub with a well-established and efficient health care innovation ecosystem, this review provides the basis for a review of commercially available mHealth platforms available in surrounding countries and worldwide, alongside actionable recommendations. In addition, we restricted our inclusion criteria to mHealth platforms in English. It is possible that some apps may be published exclusively in other countries’ stores (eg, United States, United Kingdom, or Canadian stores) and in other languages that were ultimately excluded from our search and this review. Furthermore, in the academic setting, the goal of the research team may be to assess the efficacy of the mHealth intervention they developed. After proven efficacious, commercialization may be the next step in the pipeline, in which case a name may be given to the platform. As such, this review does not include nameless mHealth platforms that were developed solely for research purposes. This decision was made so that we could identify the app crossover between the 2 searches. In addition, we used exclusion criteria such as mHealth platforms developed for specific target populations (eg, comorbidities), secondary mental health problems related to a physical condition (eg, pregnancy), and platforms targeting health behaviors (eg, anxiety for smoking cessation) and severe mental illnesses. These criteria narrowed down the included platforms considerably. Overall, the possibility of mHealth platform exclusion may have introduced bias into our analysis. However, by including a comprehensive search of both the literature and multiple app stores, we aimed to minimize the possibility of mHealth platform exclusion in this review. Finally, as previously mentioned, we carefully selected search terms that could be used in both the literature and app store searches to ensure that these searches could be conducted as homogeneously as possible despite the different methodological requirements for each. As a result, some common key terms (eg, treatment and prevention) and phrases for the clinical conditions (eg, mood, sadness, affective disorder, phobias, and CBT) that would be included in a traditional literature database search were excluded. Therefore, it is plausible that some mHealth platforms described in the literature may have been excluded from this review. However, this review aimed to provide an overview of current mHealth platforms developed specifically for general nonspecific anxiety or depression; thus, we do not expect many potentially excluded platforms from this review. If more specific anxiety and depression conditions such as mood disorders or phobias are of interest, the keywords should be developed accordingly.

Furthermore, this review does not include a quality assessment of the mHealth platforms. We intended to use the Mobile App Rating Scale [198], as indicated in our PROSPERO registration (CRD42020193956). However, it was beyond the scope of this project to download all the included mHealth platforms, many of which required payment for installation or were not accessible (ie, from the literature). Therefore, we used the information detailed in the descriptions provided by the app stores or in the literature. As a result, some data were not available, and we could not avail of the Mobile App Rating Scale for the intended purpose. We attempted to use an alternative mHealth quality assessment tool; however, we did not find one that would be suitable for these purposes. Nouri et al [199] has previously identified substantial heterogeneity in the assessment criteria for mHealth apps in different studies. As valid and reliable frameworks to assess the quality of mHealth platforms are developed, future reviews, such as ours, should incorporate them.

It is estimated that 264 million people have depression and 284 million people have anxiety worldwide [200]. The COVID-19 pandemic may exacerbate these numbers [201]. Although effective treatments exist, many health systems worldwide are underresourced and struggling to respond to the burden of anxiety, depression, and other mental health disorders effectively and efficiently [201]. As a result, mHealth platforms continue to gain popularity and are emerging as a feasible option for improving the quality of and overcoming access barriers to mental health support. However, the aforementioned concerns of lack of regulation and low levels of evidence are limiting factors for the current mHealth platforms developed for anxiety and depression to become successful solutions. A leap forward would involve venturing into digital therapeutics (DTx), which is widely regarded as the next step for mHealth. DTx offers effective, safe, highly personalized, objective, and cost-saving health care where barriers between information silos are overcome [202], all of which are ideal for addressing mental health conditions such as anxiety and depression. DTx require clinical and regulatory oversight and demonstrated RWE and are further differentiated by the delivery of a clinical-grade intervention (beyond tracking) to treat or manage a disease [198]. In other words, DTx interfere with the patient to modify their state, often through behavioral alteration. They can operate as a standalone treatment (eg, treating a mental disorder through gamification) or adjunct to other therapies to enhance them (eg, improving drug adherence). With the expectation to grow dramatically over the next few years, DTx can potentially affect users, organizations, and the health care industry in different ways: new treatment options, improved care pathways, incorporation of new standards of care, improved patient and population health outcomes, similar coverage to existing therapies, international product quality standards, and acceptance as an independent class of medicine, among others [203]. DTx is a favorable next step in mHealth solutions for anxiety and depression.

This study analyzed the characteristics of mHealth platforms for anxiety or depression that were available for research, commercial use, or both. A systematic search of the current literature and 2 popular app stores identified 169 and 179 mHealth platforms, respectively. mHealth platforms varied in their targeted condition, targeted group, purpose, intervention type, and technology type. In addition, most platforms from popular app stores lacked clinical and RWE, and a small number of platforms present in the literature were available commercially. Future efforts should focus on further accessing the quality—utility, safety, and effectiveness—of the existing platforms and ensuring the adoption by developers, from both commercial and academic sectors alike, of a reporting guideline for their platform description, as well as a regulatory framework to facilitate the development, validation, and deployment of mHealth platforms for anxiety or depression.