DMH apps developed for both underrepresented and majority populations have been hindered by a lack of real-world effectiveness data and a paucity of implementation trials [28]. The lack of timely effectiveness data prohibits consumers and medical providers from making informed choices, reducing trust and uptake. Implementation science offers frameworks and methodologies applicable to both increasing sustained use of new technologies and more rapidly evaluating digital products [29]. Drawing on implementation science and human-computer interaction, the Accelerated Creation to Sustainment model [30] offers an alternative approach to traditional clinical evaluation methods that can take over 10 years to move an intervention from conceptualization to implementation [31]. This approach proposes an iterative process of design and evaluation across three phases (create, trial, sustain) and may be facilitated by real-time engagement and outcomes data gathered by the device itself. This approach highlights the need to utilize user-centered design [32,33] to support both the development of a product and an associated implementation and sustainment strategy. In the trial phase, products and implementation strategies are concurrently tested in “real-world” settings utilizing Optimization, Effectiveness, and Implementation studies that explore effectiveness and implementation outcomes. Outcomes such as fidelity, adoption, uptake, and cost are particularly important to track to ensure new technologies are acceptable to a diverse population of consumers. Finally, in the sustainment phase, ongoing passive or low-effort data collection informs continued optimization as the research team transfers responsibilities to clinical or organizational staff.

As the field moves toward generating a more robust and easily interpretable evidence base for specific apps, providers and consumers must be empowered with alternative strategies to make informed decisions on which apps to utilize in clinical practice. The American Psychiatric Association’s app evaluation framework offers a guiding framework for providers and consumers to self-evaluate apps [34]. In principle, this framework offers a much-needed solution to a complex problem, but it also places a potentially high burden on providers to evaluate digital therapies, which the framework itself notes is “not what psychiatrists and mental health clinicians are classically trained to do [provide]” [35]. Given these challenges, it is unclear how provider evaluation of mental health apps will practically fit into the workflow for clinicians.

Moreover, given that a substantial proportion of mental health services is rendered through primary care [36], additional steps need to be taken to help provide primary care providers with the necessary information to make informed decisions about app selection for their patients. It is therefore critical to train both primary care and mental health professionals in digital technologies to support patient care. Frameworks for digital competencies in mental health have primarily focused on telehealth and do not fully address integration of broader products including the prescription of digital therapeutics or integration of apps to support treatment [37,38]. There must be a synergistic balance between providers being empowered to effectively choose and evaluate digital products on an individual basis for their patients, and more transparent and accessible information on individual products. The balance may be facilitated by the inclusion of a digital specialist or “digital navigator” [39]. Finally, broader provider and stakeholder education is critical as lack of understanding or knowledge of digital approaches is a key barrier to DMH uptake and adoption [40].

Digital approaches offer the opportunity to develop adaptable interventions that provide tailored content based on individual characteristics, such as race, gender, sexual identities, family structure, and language preferences [41]. The pandemic has further highlighted the potential for external stressors to exacerbate previously managed chronic mental health concerns and the need to flexibly ramp up care. We can leverage novel clinical trial methodologies (eg, the Multiphase Optimization Strategy [42] and sequential, multiple assignment, randomized trials (SMART) [43]) to develop adaptive DMH interventions that adjust the type or dosage of the intervention based on patient characteristics. SMART trials facilitate this development by randomizing patients to different treatment options across time based on predetermined decision points, allowing for an assessment of effectiveness for different interventions at each stage. These novel designs allow for the evaluation of the tailoring variables and intervention components in the same trial. Moreover, they support the development of decision rules for assigning treatments based on patient characteristics and response to interventions, rather than a priori decisions. This can result in shorter and more targeted interventions.

Considerable work is required to ensure the availability of DMH products that fit a wide range of user needs and preferences. Efforts have been hampered by the underrepresentation of BIPOC researchers and developers, calling for fundamental shifts to address structural barriers in academics and industry. In order for digital technologies to recognize their full potential to mitigate widening behavioral health disparities exacerbated by the COVID-19 pandemic and ongoing cultural-based traumas, they must be designed with those who are both at most risk for mental health concerns and who face the greatest barriers to health care engagement. At risk of oversimplification, this fundamentally calls for us to follow the core principle of designing for the end user [32]. This includes both the type and way we utilize technology to deliver care and the intervention content we deliver. Human- or user-centered design approaches that ask the question, “What are the barriers to receiving effective behavioral health care and how can technology help overcome them?” must be used to develop strategies that mitigate disparities. However, such approaches must engage consumers whose voices are traditionally underrepresented in care settings. Culturally salient recruitment strategies [44], which include addressing community mistrust, participant resource constraints, and potential risks (eg, community stigma), could help researchers and developers more effectively integrate the input of those consumers most likely to benefit from their products [45].

BIPOC underrepresentation in the product development process has potentially reinforced structural inequalities in our health care system by limiting the availability of products that are culturally inclusive and effective [8]. The lack of culturally appropriate tailoring, stakeholder input, and broader community-partnered implementation plans limits the effectiveness of DMH tools, leads to decreased uptake, and may contribute to lower rates of DMH utilization in BIPOC populations despite equal or higher preferences for using DMH apps compared to White peers [46,47]. To increase potential effectiveness and uptake, cultural tailoring must go beyond language translation to incorporate cultural values, norms, and references [48]. This tailoring offers the opportunity to capitalize on community or cultural resiliency factors offering opportunities to increase efficacy of interventions in addition to improved engagement. Finally, digital technologies offer the opportunity to develop intervention approaches that are not as inherently embedded in cultural paradigms or at minimum are language agnostic (eg, attentional bias modification) [49].

As we work toward the development and rapid real-world evaluation of DMH approaches for diverse populations, there is a strong need for a sustainable ecosystem to increase consumer and provider confidence, and empower users and providers to effectively choose appropriate and effective DMH interventions based on their specific needs and context. At an aspirational level, there are several approaches that could help increase this confidence.

Across all DMH products, there must be increased industry pressure to self-monitor, conduct rigorous research, and disseminate results in a manner that is interpretable by both consumers and providers. A potential short-term step toward this end is clear reporting of digital products released under the COVID-19–related FDA guidance to ensure company accountability to consumers and facilitate uptake of new low-risk products. On the open app marketplace, the inclusion of clear, easily interpretable statements regarding potential efficacy and risks, including data security, must be encouraged. A review of 73 popular apps found that none of these evaluations or membership in app libraries were noted in their app store descriptions [27], suggesting a potential lack of awareness of these schemes outside academic communities or that accreditation could benefit companies. App stores could therefore provide a more standardized way to include reporting of clinical testing results [50]. It is recognized that these recommendations introduce burden to companies and, without enforcement-related contingencies, are unlikely to be fully effective. Nevertheless, continued dialogue among stakeholders, including patient and advocacy groups, providers, and DMH companies, will help advance the field.

The lack of reliable, easily accessible efficacy and safety data impedes clinician and consumer ability to make effective health care choices. Digital formularies that provide a curated list of clinically supported DMH products could be an important mechanism for enabling dissemination of products including apps into clinical care [51]. Formularies could integrate into electronic health records and associated patient portals, allowing both patients and providers to know which products are supported by their health care organization. Indeed, some steps have been taken in this regard, and lessons learned from these efforts will be important to refine the approach [52,53]. Similarly, the development of independently curated app libraries [54,55] takes a clear step toward increasing transparency around the safety and potential efficacy of apps outside of a formalized health system. However, attempts are hindered by the vast quantity and rapidly changing nature of products on the market. Conscious of the challenges of keeping up with this massively expanding and growing field, the FDA’s Digital Health Precertification Pilot Program [56] shifts regulations toward the app developers and manufactures themselves, allowing for products to be developed efficiently through rapid iteration more appropriate for digital products, therefore reducing the need for burdensome hurdles to be cleared with frequent software changes, etc. For this to be successful, there must be an understanding between regulators and developers of a consensus framework for product evaluation. In addition, there must be adequate incentives for engagement in any regulatory programs, as prior voluntary certification attempts have failed [57].