Interventions are intended for a particular audience who, by interacting with the interventions, hopefully, achieve positive changes that are explained by the interventions’ mechanisms of action [10-13]. To contrast and discuss the paradigms of health care and entertainment, we focus on 4 interrelated themes that are necessary, but not sufficient, for the development of new products (Table 2). We acknowledge that other aspects are also necessary, including privacy [59,60] and integration with existing infrastructure [61]. However, we consciously focus on the 4, as they are illustrative of the differences between and strengths of the 2 paradigms and serve to solve the tension between them.

Intervention development begins with the target audience’s needs [62]. In mental health, there are 2 main ways in which user needs are conceptualized: a diagnostic approach and transdiagnostic approach [63]. The former frames the need through diagnostic taxonomies, and the latter focuses on a particular symptom or health-related behavior. However, both approaches frame the intervention in terms of a health indication, and it is the health care paradigm that provides the initiative; goal; and, consequently, methods for evaluating the intervention’s success.

Health indications, such as ADHD, schizophrenia, and depression, create the initial frame for the target audience. However, the actual target market is a subpopulation of this total addressable market influenced by several implicit factors. Users’ age, gender, education, personality, and prior experiences with mental health services and the severity of their symptoms influence user engagement [64]. Knowing these background factors provides an opportunity to target audiences that favor digital interventions and to extend the intervention reach to new and underserved audiences through inclusive and participatory design [65]. Although our discussion emphasizes the end user, the intervention’s target audience may also include other stakeholders, such as therapists in case of interventions that blend web-based and face-to-face therapies [66].

Besides the factors influencing DMHI use in general, some aspects influence user engagement with gDMHIs specifically. This reflects the considerable differences in users’ game preferences [67]. Appreciating the divergence in player motivations goes back to Bartle [68], who noticed that the players of multiuser dungeons (an obsolete multiplayer game genre) engage with the games for divergent reasons and in different ways. On the basis of “a long, heated discussion,” he identified the following 4 motivations to play: achievers play to pursue game-related goals, explorers seek to understand the fictional game world, socializers focus on role playing and interacting with others, and killers compete for dominance through conflict. Bartle’s [68] insight was that players engage in an apparently similar activity for widely different reasons and use the affordances of the game to serve their preferences. Since his original work, player motivations have been explored further. A review of 12 player typologies found that there is surprising conceptual uniformity in the categorizations, concluding with 5 primary player motivations, adding immersion¸ which is concentration on the story, fantasy, and narrative, to Bartle’s original 4 [69]. In addition to affecting how games are played, game preferences influence which games are played. Using survey data, Mandryk and Birk [70] classified players into 4 categories: those who appreciate single-player games, those who appreciate first-person shooter or action games, those who appreciate casual games, and those who appreciate most genres. Moreover, the researchers found that their categorization was not associated with depression, indicating that people with depression appreciate various genres. Thus, the game genre decision partly defines who finds the genre attractive and engaging.

We advocate an integrative approach to the target audience. The health indication provides a rationale for the intervention and limits the target audience. In addition to the factors influencing engagement with DMHIs, the chosen game genre influences who are drawn to the game-based intervention, continue to engage with it, and hence gain a positive impact from the interaction. Understanding the target audience allows developers to identify and cater to the variance in the users, their needs, play styles, and preferences and to remove frictions that prevent players from enjoying the intervention as much as they expect to.

The user’s motivation to interact with the gDMHI is to gain certain benefits and enjoy their time while doing so. According to concepts proposed by Siriaraya et al [40], the “therapeutic value” of the intervention is a function of its “game value”: the more experientially pleasurable the intervention is, the more likely the player is to continue to interact with it. These 2 concepts are related through the behavioral change that the interaction invites. The capability, opportunity, motivation, and behavior (COM-B) system describes 3 antecedents for behavioral change [12]. To build the user’s capabilities, opportunities, and motivation, the intervention can, for instance, educate, persuade, train, and model new behaviors, which we call the mechanisms of action. Behavioral change can lead to symptom change [13], although often after a significant delay of days and weeks, which should motivate attention to the player experience.

Mere interaction with the intervention, however, is not sufficient, which Yardley et al [38] highlighted through the concepts of microlevel and macrolevel engagement. For instance, an exergame delivers its effects when the user trains with it, exhibiting microlevel engagement. This may turn into macrolevel engagement with the broader goals of the intervention when the intervention teaches the user healthy exercise routines and encourages the transfer of new behavior outside the immediate interaction. Thus, unlike commercial entertainment games, whose success is often closely tied to microlevel engagement, gDMHIs are driven by 2-tier engagement, and a successful intervention becomes obsolete once the player has internalized the intended change, which is a considerable digression from the entertainment paradigm. Subjectively, microlevel engagement may be experienced as a reduced effort to interact with the intervention. Subjectively, microlevel engagement may be experienced as a reduced effort to interact with the intervention [71]. This is also what the player may expect from a game-based approach: when the intervention is introduced through the notions of “game” and “play,” they allude to interactions that are more pleasurable than those with solutions more strongly associated with a functional approach.

To design for microlevel engagement, the concept needs a clear definition and metrics. Here, we turn to Perski et al [72], who provide a useful definition: “(1) the extent (eg, amount, frequency, duration, depth) of use and (2) a subjective experience characterized by attention, interest and affect.” The 2 sides represent complementary aspects of engagement, objective and subjective aspects, which require different approaches in their evaluation [73,74]. The subjective criteria may be pursued through qualitative research, such as user interviews, think-aloud protocols, and focus groups that provide rich and detailed qualitative data on engagement [75], or swiftly through questionnaires [76]. Subjectively, high levels of engagement with gDMHIs may be experienced as immersion, where the player becomes unaware of the mediating technology, perceiving the interaction as unmediated [77]. The phenomenon is also known as presence, a sense of being “in the game.” It has many similarities to the notion of flow, where the subject is absorbed in the task and becomes unaware of their surroundings [74,78,79]. This pleasurable mental state is the experiential value players seek from the game and manifests as a sense of “want to” rather than “have to” engage with the game [80].

To complement the evaluation of a subjective sense of engagement, game analytics can provide objective indicators of engagement. Among the strongest and most valuable tools in commercial entertainment game development, game analytics provides proof, also in business terms, that the target audience enjoys spending time with the game [81]. User analytics offers insights into microlevel engagement by measuring the time spent with the intervention, accompanied by metrics of retention, that is, players returning to the intervention.

Entertainment games are a medium focused on building engagement that is experienced as immersion. This provides the developers with a user-centered goal that can be evaluated both subjectively and objectively. The health care paradigm complements the approach by ensuring that evidence-based mechanisms of action are in place so that user interaction contributes to the expected health benefits.

The effectiveness of gDMHIs is indicated by behavioral and symptom changes [10,13]. Reaching a meaningful reduction in symptom scores indicates a clear benefit expected by clinicians and the players and payers of the intervention alike. To complement the symptom- or disorder-focused approach, the intervention may have other, equally meaningful benefits that occur in parallel or in addition to the symptom change. As the World Health Organization (WHO) [97] states in its constitution, “health is a state of complete physical, mental and social well-being and not merely the absence of disease or infirmity.” In subjective terms, symptom alleviation may be related to improved quality of life [98]. To illuminate the complexity of the concept, a qualitative synthesis of the experiences of people with mental disorders found that good quality of life was associated with feelings of well-being, positive self-perception, control, autonomy, a sense of belonging, participation in meaningful activities, and a positive future view [99]. In other words, an improved quality of life is not perceived merely as the absence of symptoms.

Expanding from the symptom-based approach, the impact of gDMHIs may be considered through the self-determination theory [100]. It describes the following 3 universal human needs that contribute to intrinsic motivation, self-regulation, and well-being: competence, autonomy, and relatedness. Digital games have a unique possibility to contribute to the 3 needs by fostering a sense of competence by presenting the players with increasingly difficult puzzles as their skill increases, creating a sense of connection with other players enjoying similar activities, and supporting player autonomy by offering choices and various ways to play. As opposed to mediums that are more passively consumed, games are characterized by a sense of agency [101]: the player does not spectate the protagonist but becomes them, controls them, and sees the fiction through them. By answering to the players’ needs, game-based interventions can have positive effects across the cognitive, social, and emotional spheres [56].

However, whether the interaction serves the underlying human needs depends on both the game design and the players’ mindset. Games that serve the players’ needs are associated with continued motivation to play [102,103] and affective well-being [55]. When the players use the game to fulfill their psychological needs, playing can be viewed as adaptive and healthy [104], whereas using the game as an avoidance behavior is associated with mental distress and problematic gaming [105]. Thus, creating designs that answer to the players’ deeper psychological needs and encourage healthy playing styles is vital.

User interaction with the gDMHI produces a vast amount of data, including behavioral biomarkers [106], which may contribute toward understanding the psychological needs, mechanisms of action, and mental well-being beyond the disorders. Players’ interaction with the game mechanics and puzzles provides indications of their motor and cognitive performances, which are particularly useful in cognitive training interventions. Players’ affective states and changes in their social interactions may be derived from their text-based outputs and interactions with other players, which contributes to therapies that use reflective approaches. However, the use of behavioral biomarkers as outcome measurements is only growing. Recently, they have been used, for instance, in the assessment of social anxiety [107], mood disorders [108], and mild cognitive impairment [109].

To summarize, gDMHIs are expected to contribute to a significant reduction in mental health symptoms, which is preceded by player engagement and behavior change [13]. Thus, the overall effects of the intervention may extend beyond the symptom change that the digital medium affords to accurately measure. Leveraging in-game data provides corroboratory insights into why and to what extent the interventions are effective, including in terms of broader psychological needs. This contributes to understanding how the interaction impacts a user’s behavior, which is, hopefully, generalized in the broader context of the user’s life [110], exhibiting macrolevel engagement.

New product development is a gradual, evolving process [10,11]. It can be described in phases, bearing in mind that the research, design, and development efforts, particularly in a novel domain, are often characterized by exploration, iteration, looping back to earlier phases, and redesign based on feedback. Therefore, the development phases should be considered nonlinear.

We summarized gDMHI development into 4 phases (Table 4). The design begins with research into the target audience and their needs, which are performed through a high-level concept that outlines the intervention. The intervention, as well as the service concept more broadly, is advanced iteratively during development. The evaluation of the intervention likewise gradually grows more comprehensive, culminating in a randomized controlled trial (RCT) that is valued as a particularly strong proof in health care. Should the results prove favorable, the solution may be implemented, and its development may continue in live production.

Throughout the 4 phases, the focus of the development team fluctuates between the paradigms of health care and entertainment: designing an effective health care intervention and an engaging video game experience [111], pursuing instrumental and experiential outcomes [39], or, in other words, creating therapeutic value and game value [40]. The 2 complementary perspectives intertwine through the development process.

The concept provides an overview of what the intervention is, which is intimately related to the following question: who is it for? Understanding the target audience and their contextual needs and preferences allows defining the characteristics of the intervention [10,11,112]. Interviews, questionnaires, focus groups, desk research, and ethnography can be used to create a rich understanding of the target audience’s preferences, media used, and target audience's thoughts about and frictions with the existing solutions and services. However, the tools are secondary to the aim of shifting one’s perspective through empathy, which is a cornerstone of design [111]. In fact, several development frameworks, models, and philosophies focus on users. They include participatory design [113]; service design [111]; persuasive design [114]; and user-centered design [115], which is also common in gDMHI design. A study exploring 20 development processes found that 50% of them adopted a user-centered approach in the concept phase, and the rest invited user participation in the development phase [116]. Accentuating the importance of users reflects the threats of overly focusing on the solution instead of the need it addresses, which may lead to users rejecting the intervention, and necessitating costly changes later in the development [4,117]. Starting from the target audience, as well as other necessary stakeholders, facilitates the definition of the intervention objectives, technology, and game genre [51].

It is a common misconception that user research would directly inform the design. Users can describe their experiences with existing solutions, but developers are responsible for integrating the understanding into a model that considers the mechanisms of action, technological possibilities, stakeholder dependencies, commercialization, and regulations. Often, the digital concept aims to challenge the dominant, nonscalable, one-to-one, and in-person model of mental health service delivery [118]. The concept can also include considerations of the business model: how the intervention is priced, purchased, and compensated [119]. In contrast to the entertainment industry, which is dominated by the direct-to-consumer model, where the player chooses, uses, and pays for the game, health care characteristically includes more complicated commercial models. Aitken and Nass [32] described 4 such commercial models: direct-to-consumer model, where the user pays for the service; device-like reimbursement model, where the cost of the intervention is covered by the medical benefit plan when the intervention is prescribed by a physician; drug-like reimbursement model, which relates the intervention to pharmacy benefit; and value-based contracting, where the intervention is provided for an organization and paid per the benefits achieved. Thus, the preliminary documented concept is a user-centered account of whom the team is designing for, a science-based account of how the intervention contributes to the user’s health, and a description of the intended commercial model.

During development, the team gradually turns the concept into an actual intervention. The phase is iterative: it oscillates between design and testing, which seeks to ensure that the team is progressing in the right direction [10,112,115]. During development, the team can make use of methods from entertainment, which allow the team to focus specifically on the intervention features, and methods from health care, which focus on establishing and implementing the intervention in the complex health care ecosystem.

The success of games is tied to the experience they create [46]. There are various perspectives on game experience [120]; however, in general, it is built of layers, as the user interacts alone or with other players with the game system, its look and feel, and the underlying systems. The interaction leads to an experience that the player may find immersive, motivating, and pleasurable. Entertainment game design has a wealth of methodologies that can be used to hone the game-user interaction [121-123], which acts as a focus point for the interdisciplinary efforts in visual and audio design, software development, and storytelling [44].

The business-to-consumer digital entertainment industry has established distribution platforms that allow developers to focus on game development. Health care, in contrast, currently lacks such global, or even national, distribution platforms. Therefore, it is important to design for feasibility and consider the implementation of the intervention early [11]. If the intervention is introduced into an organizational context, it needs to fit the existing practices, improve current processes, and integrate with the existing software solutions [124]. In fact, many interventions are expected not only to offer end user benefits but also to be easy to adopt and contribute to organizational efficiency. The Non-adoption and Abandonment of technologies, and the challenges to Scale-up, Spread, and Sustainability of such technologies (NASSS) framework structures the complex health care environment in which the intervention may be implemented [125]. It highlights the interrelations among the health condition, technology, value proposition, stakeholders adopting the solution, and their organizational and wider context over time. In conclusion, the development process gradually builds not only the digital intervention, in the narrow sense of the word, but also the related services and the operating models that sustain it.

The evaluation of the intervention grows gradually more comprehensive. In development, the intervention is tested for its playability and game experience, ensuring that the game-based intervention is found sufficiently favorable by the players in the target audience [46] and is meaningful for other stakeholders. As the development progresses, there is increasing interest in evaluating the intervention’s health-related effectiveness. A feasibility study, a higher-order concept than playtesting, focuses on the following question: can the intervention work? [126]. Feasibility studies allow for evaluating whether the intervention can attract the appropriate target audience, understanding how acceptable the participants consider the intervention and its procedures, and understanding how the intervention is to manage in practice and contribute to the development of data collection methodologies. However, a feasibility study provides only initial indications of the intervention’s effectiveness. Pilot studies can create stronger, although inconclusive, evidence of intervention effectiveness [126]. Should the intervention pass the preliminary evaluations, an RCT can be conducted. It provides high-quality evidence on the effectiveness and safety of the intervention for regulators, clinical stakeholders, and the scientific community. However, because conducting an RCT is demanding, expensive, and slow, progressing to them without ensuring the overall feasibility of the intervention is not advised [112].

It is important to evaluate the intervention comprehensively beyond the summative results: formative evaluation can provide insights into how and why the intervention is or is not successful in its context [127]. Qualitative research can contribute to understanding the intervention’s impact and player experience [128,129], and evaluating engagement creates insights into the intervention’s feasibility, acceptability, attractiveness, and dose-response relationship. However, the evaluation of engagement is often lacking, and there is substantial heterogeneity in how it is reported [26,73]. One approach is reporting the following 6 factors: the number of users, their profiles, the number of modules in the intervention, the number of times they are accessed, the percentage of users receiving a therapeutic dose, and the achieved clinical change [26]. Economic evaluation is not to be overlooked either, as it often forms the rationale for adopting a digital approach. Economic evaluation can be performed to understand, for instance, the cost-effectiveness of the intervention and to guide strategic decision-making and investments [130].

We encourage a progressive and comprehensive evaluation of the intervention, that is, considering the subjective user experience and objective engagement alongside the clinical metrics. Complementing summative evaluation with formative and economic evaluations allows for an understanding of the intervention feasibility and whether the benefits justify the costs, which facilitates translating the research to complex real-world environments should the intervention prove both effective and engaging.

When the intervention works in controlled research settings, it may be implemented in more complex, real-life settings. Following the principle of path dependence, the decisions made in the earlier stages regarding, for instance, the health indication, mechanisms of action, game genre, and technical implementation continue to exercise their influence in the production phase. Building on the established foundation, the attention turns to marketing, service production, partnerships, and continuing development.

Implementing the intervention often encounters challenges: the research-to-practice gap [27]. Common barriers include user preference for face-to-face therapies, the perceived complexity of the digital intervention, limited research evidence regarding the intervention, low user engagement, the costs of the intervention, and practitioners’ resistance. To overcome these challenges and achieve real-world impacts, developers should prepare for them using numerous models [124,125,131]. The Reach, Efficacy, Adoption, Implementation, and Maintenance (RE-AIM) model provides an overview of the components necessary for creating an impact on a population level: the intervention needs to reach a broad audience, be effective, be adopted in numerous settings, be implemented as intended, and be maintained over time [131]. Thus, besides the need for the intervention to be proven effective, the implementation of the intervention demands marketing and organizational support for dissemination. Another useful perspective is the consolidated framework for implementation research, which consists of 5 domains [124]. Intervention implementation is dependent on the intervention characteristics, relationship of the intervention with external stakeholders, organization’s internal capabilities to produce the service, individuals within the organization delivering the intervention, and implementation processes. Outside the research setting, the development team needs to consider ways to deliver the intervention to users en masse, which demands changes in organizational structures and service management.

Should the intervention reach a sufficiently considerable population, the developers may begin to view the software as a continuous service. Here, methodologies already in use in game development may prove useful. In contrast to the earlier, linear development process that considered the launch of the game the conclusion to the development, the prevailing game as a service model focuses attention also on postlaunch development [132]. This approach, also called live operations, relies on using business metrics and user feedback to guide the continuous development efforts and add new content [133]. In the future, there is potential in considering gDMHIs as a continuous, evolving service rather than unaltered products [134]. However, it is still unclear to what degree the digital intervention may be developed while retaining the effects and evidence gathered in the previous stages, and there is little research on gDMHI live operations. This reflects the nascent industry that is still waiting for successful interventions to attract a large number of users to fund the continuous research and development of a live product.