E-mental health research is expanding around the world [1]. This area of mental health research and intervention relies on digital technologies to deliver complementary care, support, and information [2]. Over the past ten years, digital mental health interventions have appeared at an unprecedented rate, largely in the form of mobile apps, social media, chatbots, and virtual reality [3]. Since the beginning of the COVID-19 pandemic, the digital health world has expanded at an unprecedented rate [4], and its potential to improve access to care has never been greater [5]. However, some important challenges remain in the field. Several issues raised in recent years have not been resolved: there is some distrust in the field that is not served by the lack of empirical validation of its benefit [6-9]; it raises privacy and data security concerns [9,10]; it presents commercial issues (eg, financial interest, user access, advertising) [11,12]; and the solutions often lack usability and show low user engagement [13,14].

The promise of digital technology still far outweighs the reality of its use. This is particularly evident in the field of digital mental health, in which designs must survive successive waves of adoption: phases of preuse, first use, and sustained use [15]. A study of 93 mobile mental health apps showed that the overall user retention is very low, with a 15-day retention rate of 3.9% and a 30-day retention rate of 3.3% [16]. Another study with 77 participants in two randomized controlled trials demonstrated how difficult it is to motivate people to begin using an e-mental health solution [17]. Any digital health trial will see a significant proportion of users drop out or cease using the app before completion. Eysenbach [18] calls this phenomenon “the law of attrition.” Data on the health app market are scarce but do converge on two findings: the majority of health apps are downloaded fewer than 5,000 times, and 46% of apps have less than one monthly active user [19]. While usage is only one indicator of engagement [20], these statistics are consistent with adoption issues commonly reported among users, such as a lack of awareness of the app or lack of time and motivation to use one [21,22]. This is concerning because the use of these apps may not be associated with a significant decrease in mental disorders if they are not used for the intended period of time [14]. The average cost of developing a mobile health app is US $425,000 [19]—a cost-benefit ratio too high and unsustainable in the long run if we do not change how they are developed.

Research has shown that most people are willing to adopt and use some form of new technology in the interest of improving their mental health [23]. So why the low utilization rates? Given the already significant barriers to adoption that users face (eg, privacy concerns, commercial issues), we seek to underscore the importance of user-centric design approaches for the development of e-mental health solutions, of which a solid notion is lacking in the digital mental health design sphere [14]. Based on the existing literature [13], we hypothesize that the lack of adoption of digital mental health solutions could be largely due to a lack of attention to user perspective in the design of these technologies, or at the very least, a lack of understanding of design approaches that accommodate user perspectives. In the field of mental health, there are very few examples of involving real people with mental disorders in the development and design of mobile apps intended for them [14]. The most common development approaches seem primarily researcher- and expert-driven, top-down in style, and to rely mainly on a bilateral partnership between clinicians and engineers [24]. This is not adapted to the challenges of contemporary digital culture that places the user at the center of these platforms by empowering them [25].

In this section, we recall some fundamental principles of design culture and explain how they can help actors better account for the needs of users and integrate their perspective early on in the e-mental health design process.

The main objective of this literature review was to understand how human-centered design is considered in e-mental health intervention research. The following research questions were considered:

Through our efforts, we seek to open the discussion on the place of human-centered design methods in e-mental health research.

To answer the 3 questions, an exploratory mapping review was conducted by researchers from the fields of design and mental health. The aim of mapping reviews is to map out and categorize existing literature on a particular topic to identify gaps in knowledge or opportunities for further research [49]. It focuses less on findings and more on activities related to the findings, such as the quantity and quality of the literature [49,50]. To streamline the process and identify a relevant sample of articles for this interdisciplinary exploratory review, a search was conducted among journals in mental health whose explicit scope covers technology. The following journals were identified: JMIR Mental Health, Frontiers in Psychiatry, Internet Interventions, and the Journal of Technology in Behavioral Science. Articles published between 2015 and 2020 were examined using the search terms “design” and “design*” to narrow down the results.

Given the interdisciplinary nature of this work, extensive discussions were conducted among the coauthors to agree on a common understanding of the concept of design. The following inclusion criteria were defined: articles reporting original research on the development of a digital technology in mental health and those addressing the concept of design (at least one explicit use of the term design) in connection with at least one core element of a design activity.

The use of the term “design” in research (eg, “study design”) or in its common sense was excluded.

The third author (MD) screened all titles and abstracts for potential articles. Then, the second (SB) and third (MD) authors independently assessed the full text of the articles for eligibility. There was an initial level of agreement of 80.7% (42/52) between the two authors (SB and MD), which is usually considered acceptable in the literature [51]. When there was discrepancy, the first author (SV) made the final decision.

For each article selected, data regarding the design approaches and the four core elements of a design activity (ie, the object, the context, the actor, and the process) were extracted. This included the type of solution created, the design approaches reported, the setting in which the project took place, and the type of actors involved throughout the design process. The design process was examined according to the steps defined in the UK Design Council’s framework for innovation [36]. The design methods reported in the articles were used to define the steps addressed in the development of the digital solutions. The analysis process was conducted jointly by the second (SB) and third (MD) authors.

Of the 1035 articles initially found, 51 full-text articles were assessed for eligibility. Of these, 30 studies met the inclusion criteria. The articles came from JMIR Mental Health (22/30, 73%), Frontiers in Psychiatry (4/30, 13%), Internet Interventions (2/20, 7%), and the Journal of Technology in Behavioral Science (2/30, 7%). Multimedia Appendix 1 presents the characteristics of the included studies, indicated from left to right: the specific research domain (eg, depression and anxiety, psychosis, well-being, etc), the synthesized naming of the adopted approach, and the 4 core elements of a design activity (object, context, process, and actors). Process is presented according to the 4 steps in the Double Diamond (coding each actor type with a number across the steps). Finally, we reported whether the study indicated that the process was iterative or not.

To develop the digital solutions, 22 studies mentioned using human-centered design approaches or specific design methods. Various design approaches were reported, and there were many variations in the names given to these approaches. After several rounds of discussions between all authors, different approaches were classified under the 3 common names used in design studies, as listed in Table 1: participatory design (11/27, 41%), co-design (6/27, 22%), and user-centered design (5/27, 19%). Under the term “participatory design,” generally named as such in the studies, we considered alternative names such as “user involvement” [52]. Under the term “user-centered design,” we included other names like “person-based approach” or “person-centered approach.” Other studies reported specific design methods (5/27, 19%) that did not correspond to these 3 common names. Those methods are not common in the design studies field, except for the UK Design Council’s Double Diamond method. Among the studies included in this review, 8 (27%) did not refer to any human-centered design approach, so they are not listed in Table 1 [53-60]. Five studies reported more than one approach, mixing 2 common approaches or 1 common approach with 1 specific design method, and these studies are demarcated with a superscript in Table 1. Only 16 studies provided a definition of the reported approach(es), either by referring to other studies (15/16, 94%) or by offering their own definition (1/16, 6%). This means that about half (14/30, 47%) of the studies did not cite or provide a definition for their chosen approach, include references, or mention the theoretical underpinnings of the design approach. Although it was the second most reported approach, co-design was never defined in the 6 studies that mentioned it.

Although the 30 studies selected addressed the concept of design and reported a variety of human-centered design approaches, very few explicitly mentioned the presence of designers on their teams. Regardless of the step of the process, only 8 studies mentioned designers, representing about 27% of all included studies. For the 22 studies that did not mention them (74%), we do not know whether it is because no designer was involved or because the presence of designers was not considered important enough to be reported.

Looking at the few studies mentioning designers in their teams (8/30, 27%), it is interesting to note that some designers were explicitly present for all steps but mostly just the first three: discover (4/8, 50%), define (3/8, 38%), and develop (7/8, 88%). Only 1 was explicitly present for deliver. It is also interesting to note that 3 of the 4 studies that included designers in the discover step also included them in the define and develop steps, reflecting their ongoing involvement in the process. These 3 studies represented a small proportion of the studies that reported using participatory design and co-design. Overall, designers were clearly more involved in the develop step (7/30, 23%) but much less involved here than software development companies (Table 3). The latter were exclusively present in this step (14/30, 47%). End users were the most present participants at each step of the design process (Table 3).

In this initial exploratory research study, we investigated how design is considered in e-mental health research. Our results show that there have been attempts to integrate human-centered design methods into the development of e-mental health solutions, but they are still rare and rely very little on designers or design research. Most reported design approaches such as user-centered design, participatory design, and co-design are well known and documented in the design research literature, but most of the included studies did not rely on them. Almost half of the included studies did not bring or report any existing definition of the design approach they used. Moreover, it was not possible to link the use of an approach to its influence on the main core elements (steps conducted through the process or actors involved) and vice versa. The impact of each chosen approach on the whole process is not clear, nor is the reason behind the selection of a particular approach. This indicates that there is a lack of shareable knowledge on how design approaches are understood, and by extension, applied in the mental health field. This suggests that human-centered design methods are not fully integrated in e-mental health and that reported design approaches are still primarily used from the outside without a deep understanding of the design culture that is needed to fully leverage their power.

There has been very little research conducted on human-centered design methods in e-mental health and on how to guide the design of e-mental interventions. Thabrew et al [82] highlight the importance of active collaboration using co-design jointly between researchers, designers, developers, and users to develop more engaging and useful interventions. The results from this literature review show that such collaboration among all these stakeholders remains limited throughout the design process. While most design approaches reported were consistent with human-centered methods stemming from the design discipline, the choice and combination of the approaches varied greatly across studies. Orlowski et al [83] claim that the e-mental health development process must prioritize empathy and understanding over innovation, as proposed in participatory design and design thinking approaches. Torous et al [14] highlight the poor usability of mental health apps and the lack of user-centric design. Aryana et al [84] attempt to identify the key principles of the design process relevant to mobile mental health. Among the 6 principles identified, they mention “high quality user experience,” which is closely related to user-centered design, and an “empathic design process,” which is closely related to participatory design and co-design, and conclude that there are few examples of the implementation of several of these design principles in real-world products. This was also the case for the identified research projects. Bakker et al [85] note that design principles that have led to the huge success of many physical health and social networking apps have not been utilized in the mental health apps field. These findings are all consistent with our study and show that human-centered design methods are largely underutilized and neglected when their impact could be very important, especially on user engagement.

This exploratory review offers significant insights into how design is considered in e-mental health. We consider it to show a fairly representative sample of the type of design-related research currently being conducted on the development of digital technology in mental health. We do not think that additional studies would significantly change our main conclusions. However, this study does not meet the criteria for a systematic review and has a few limitations. First, when analyzing the core elements of design activity, we could only rely on the information reported in the articles, which was fairly heterogeneous. We had to conduct several rounds of interdisciplinary discussions among ourselves (the authors) to ensure its best interpretation. Second, to analyze the design process described in each study, we chose the Double Diamond framework, which is a global reference, but other frameworks could also be used and might yield additional results. Third, in all the studies selected, it was difficult to understand how end users influenced the design, especially in participatory approaches. User involvement can be informative, consultative, or fully collaborative [86]. Orlowski et al [46] have already concluded that it is difficult to track ongoing user participation and clearly determine the contribution of participatory design to the effectiveness of designed interventions.

Bridging the gap between design and e-mental health is our next research agenda. We are currently developing a health intervention research framework called Design For e-Mental Health [87]. This framework refers to the broad range of human-centered design creative strategies that define the structure, function, and form of a digital mental health with a high quality of experience in terms of user experience, scientific validity, privacy, and viability.