Meeting Young Adults’ Social Support Needs across the Health Behavior Change Journey: Implications for Digital Mental Health Tools

2.1. Young Adults and Digital Mental Health Tools

Young adults are the age group with the highest prevalence of mental health conditions. In 2019, nearly 30% of young adults in the United States (US) experienced a mental illness, relative to 20% of adults aged 26 or older [97]. This high rate of mental health symptoms among young adults is exacerbated by the fact that they have the lowest rates of mental health treatment seeking. In contrast to the 46% of US-based adults age 26 and older who sought in-person care for their psychological or psychiatric symptoms in the past year, less than 39% of those aged 18-25 did so [98]. Structural barriers, such as financial hardships, are often pointed to as reasons for the low use of in-person mental health care among young adults; however, attitudinal barriers (e.g., low perceived need for treatment, preferences for self-management, etc.) are over four times more prevalent than structural or financial barriers [68]. This suggests that efforts to increase access to traditional care alone will be insufficient to address the treatment gap within the young adult age group.

New technologies offer tremendous promise to deliver effective interventions to individuals experiencing mental health symptoms in formats that are accessible and appealing [5,21,93,105]. For instance, digital technologies to support mental health can be used in targeted ways that require short periods of engagement, increasing convenience [8,67], they can be used in private, and they have the ability to provide on-demand care when it is needed [6,92]. Digital technologies have been useful for treating and supporting the self-management of myriad mental health concerns including depression, anxiety, bipolar disorder, schizophrenia, substance use disorders, and suicidal ideation [4,17,32,33,53,59,63,112]. Young adults are heavy users of digital tools, 96% of US-based young adults own a smartphone, and they are the age group most likely to access the internet only through their smartphone [77]. Therefore, smartphone-delivered mental health interventions are a well-suited delivery medium [40,53,101]. DMH tools include apps or services designed to ameliorate mental health symptoms [4,32,33,59]. Most often, they work by promoting or facilitating a skill that targets a key cognitive or behavioral mechanism (e.g., ruminative thinking or cognitive distortions). Importantly, DMH tools offer a new intervention system that increases the accessibility of mental health treatment in a form that is effective and novel, democratizing the mental healthcare system [35,93,105,110].

However, despite many clinical trials demonstrating their effectiveness in reducing symptoms [5,32,33,59], existing mobile DMH tools have high dropout rates, low adherence, and generally low usage [9,60,104]. One recent analysis found that, in real-world settings, only 4% of individuals who have installed a DMH app open it daily [9]. While existing mobile DMH tools are effective in controlled conditions, their effectiveness in real-world settings remains limited. This research-to-practice gap highlights that, even among research studies – which typically include more highly motivated users than are found in real-world use – engagement can be low [104]. This underscores the importance of designing tools that address the needs of less motivated and less proactive potential users. Increasing rates of sustained engagement is one of the great design challenges for DMH tools. Recent work has suggested that social interactions can provide an important avenue to increase engagement with DMH tools [15]; however, many questions remain about the forms of social interaction that are most appealing and helpful, particularly for those who are initiating their behavioral change journey.

2.2. Stages of Change

The transtheoretical model (TTM), or the stages of change model, is frequently used to understand an individuals’ readiness to engage in health behavior change, and is often used as a framework for intervention in smoking [22,88,89,113], alcohol, and other substance use disorders [50]. The stages of change model has also been used to understand an individuals’ readiness for change in the context of depression and affective disorders [57], and is linked to psychotherapy outcomes, such that individuals who were in later stages of change prior to initiating psychotherapy had improved psychotherapy outcomes once they engaged in mental health treatment [50,73].

The stages of change model [84] stipulates that health behavior change happens over time and that an individual’s stage of change is a reflection of their progress across six stages. The first stage, precontemplation, is generally conceptualized as the period of time before an individual is considering any health behavior change. In this stage, individuals may not be aware of negative consequences of any maladaptive or otherwise unhelpful behaviors or practices, or they may be aware of these negative consequences of their behaviors but are nonetheless not considering health behavior change. Contemplation is the second stage of change and can be characterized by an individual weighing the benefits and drawbacks of health behavior change. In the contemplation stage, an individual may be aware of the negative consequences of certain behaviors, but they may not have decided to take any action to shift their behaviors. In this stage, individuals may not yet know the specific change behaviors to take, but may understand the negative consequences involved in not changing. Therefore, the cost of changing behavior is often weighed or considered against the potential gains of engaging in health behavior changes. During the preparation stage, individuals have resolved to imminently engage in health behavior change. They are likely to have engaged in preparatory work such as information gathering, researching ways of engaging in health behavior changes, talking to loved ones about changes, or otherwise generating a plan to engage in new health behavior changes. The next stage, the action stage, includes individuals who are engaging in change behaviors. They have overcome the cost of new behaviors and begun to modify their routines even if these changes continue to have high costs associated with them. The fifth stage is the maintenance stage, in which the new health behavior changes become more routine and the high cost of initiating a behavior lessens, instead shifting to the prevention of lapses and relapses. Finally, the termination stage is one where an individual no longer needs to expend energy or effort to stave off lapses or relapses, but has integrated the change behavior as a matter of routine [84]. Within this model, individuals can move in and out of each stage of change, depending on individual circumstances, the health behavior change in question, and the support available to the individual.

The stages of change model has previously been applied to better understand how best to meet the needs of users of a digital health service. In particular, a study by de Vries and colleagues [106] used the stages of change model as a framework to tailor motivational text messages developed by Amazon Mechanical Turk workers, with the aim of promoting exercise adherence. After coding crowdsourced motivational messages based on how well they represented a particular stage of change, the researchers then tested how motivating each message was if it was delivered to an individual based on their self-described stage of change. de Vries and colleagues found that there was benefit to tailoring messages to an individual’s stage of change, especially when individuals are in the latter stages of change. Our present study expands on this work by beginning to explore the needs and preferences non-treatment seeking young adults have on their mental wellness journeys which can then be addressed by tailoring specific tools and services.

The stages of change model allows for understanding the processes, struggles, and motivational factors that an individual goes through when confronted with the prospect of health behavior change. However, additional models can further build on the TTM by specifying how behavioral change can be supported through an individual’s interactions with technologies. In particular, two of these models include the Lived Informatics Model of Personal Informatics [31] and the Technology-Mediated Reflection Model [13]. Like the TTM, these models describe ways individuals move through various phases; however, unlike the TTM, which describes overall trajectories of a health behavior change journey, these models describe interactions with a particular type of technology: self-tracking.

The Lived Informatics Model of Personal Informatics (LIMPI) [31], describes the processes individuals go through as they consider the prospect of tracking personal behaviors, and the cyclical or non-linear processes they may encounter as part of the self-tracking experience. The LIMPI is an expansion of Li et al.’s [58] Stage-based Model of Personal Informatics, which is closely related to the TTM in that it outlines consecutive stages that individuals move between as they interact with personal tracking technologies for the purposes of behavior change [31]. The LIMPI expands on this model by describing a broader range of rationales for personal tracking beyond behavior change (e.g., curiosity and interest in quantifying personal behaviors, or simply to create a digital record of a behavior). The LIMPI, therefore is a useful model for understanding how and why many individuals begin, continue, and stop using personal tracking. The LIMPI describes the entry-point to personal tracking, the deciding phase, which the authors describe as akin to the precontemplation and contemplation stages in the TTM. This is followed by the selecting phase in which an individual decides what tracker to use. The next phase, tracking and acting, comprises three simultaneous activities including collecting personal data, integration of tracking into daily life, and reflecting on personal data. At any point in this phase, an individual can have a lapse and re-enter the deciding or selecting phase. In this way, the LIMPI mirrors the TTM in that it describes non-linear phases of personal tracking activities, but its narrower focus on personal tracking differentiates it from the TTM.

Another, related, model of personal informatics is the Technology-Mediated Reflection Model (TMRM) [13]. The TMRM expands on the LIMPI [31] by expanding upon the ways that users in the tracking and acting phase of the LIMPI enter and maintain the process of personal informatics reflection. In this model, personal health tracking technologies can help with facilitated reflection, a transitory state in which a tracking technology’s features and data summaries match a user’s (a) conceptual needs (i.e., the extent to which a user’s interpretation of data summaries matches their perceptions of their experiences), and (b) their temporal needs (i.e., how often a user wants to engage in reflection, and how frequent and long this reflection takes). As users’ tracking needs change over time, this model highlights ways in which users can adjust technology settings, or other features to create strong alignment between conceptual and temporal attributes of the tracking experience to help facilitate health behavior change. The TMRM not only helps illustrate ways in which technology-facilitated reflection can be used to bring awareness and motivation to an individual while on their health behavior change journey, but also highlights ways in which misalignment between a tool and a user’s needs can create a dissonant experience and jeopardize the critical state of facilitated reflection. The TMRM focuses on an area of a behavior change journey only once an individual has identified a desire for health behavior change, whereas the TTM describes the process of behavior change beginning before an individual is even interested in making a health behavior change (e.g., precontemplation) and spanning to a point at which behavior change has been well integrated and no longer requires effort to maintain (e.g., maintenance). While self-tracking data and facilitated reflection can be an important aspect of symptom self-management, prompting movement across stages of change, the present study is focused on the full trajectory of young adults’ mental wellness journeys.

In this paper, we explore the utility of the stages of change model as a lens to understand the kinds of social supports young adults with depression find helpful for symptom self-management.

2.3. Social Support in Digital Mental Health Tools

Social support has successfully been incorporated into a number of DMH tools, with studies suggesting it can help to sustain engagement [19,54,74]. This reflects the fact that social support helps bolster a user’s intrinsic motivation, helps normalize struggles, and provides access to compassionate others [19,54]. One study that included a systematic search of mental health apps on the Google Play store, in combination with app use data from a panel of mental health app users, found that DMH tools that incorporated peer support had longer use times and greater open rates [9]. In online mental health interventions and forums focused on helping users manage mental health symptoms, social support played a key role in promoting skills use and intervention effectiveness [3,18,24,26,28,29,54,74,102].

To categorize different types of social support in this work, we draw on the concept of “directedness,” as introduced in the HCI literature [20], which distinguishes between communications aimed at specific individuals from those aimed at a broad or public audience. The need to distinguish directed and undirected forms of support increased alongside the growth of online communication formats that give individuals routine opportunities to use ubiquitous technologies to reach audiences on a mass scale through personal technologies, and to act as senders, receivers, or both [69]. Social media sites, for example, allow for directing messages at a public audience (e.g., via a newsfeed) or at specific individuals (e.g., through one-to-one messaging). Past work in the arena of self-disclosure has found that individuals disclose differently through directed and undirected channels, with greater openness and wellbeing benefit typically being associated with directed disclosures within close dyads [11,20]. On the other hand, indirect online disclosure can also allow for candor when key elements are present, such as anonymity, a lack of pre-existing social ties, or similarity to the audience [49,85]. In addition, disclosing within many undirected forums creates a record of individual experiences that can be accessed by others.

In the context of mental health concerns, in particular, many individuals living with invisible stigmatized chronic illnesses such as mental health disorders are hesitant to disclose in directed ways, such as by seeking help from close others or specific peers [23,91]. Undirected engagement (e.g., consuming or producing knowledge from, and for, general audiences [26]) is therefore common, and includes behaviors such as online information seeking and “lurking” on peer-to-peer forums [100], as well as disclosing within these forums. Thus, there are highly varied ways that individuals may seek out input and help from others when dealing with mental health concerns, ranging from relatively direct (e.g., intentionally interacting with specific individuals [20,90]) to relatively undirected [10].

Undirected support is common among individuals engaging in health behavior changes. One study examining online health information seeking among adults across the lifespan found that, overall, approximately 40% of participants searched for health information at a minimum of once per week [27]. Reasons for seeking out the experiences of others online include: to identify alternative sources of information (relative to conventional sources of health information), to seek personal recommendations or advice, and to find similar others [27]. One study of Finnish 18-year-old men who had been “called up” for civil service, examined online information-seeking practices in the context of physical activity changes and found that 60% of their sample sought information about physical activity or exercise in the weeks leading up to the study. Information seeking behavior was more likely from participants in the latter (i.e., maintenance) stage of change relative to participants in the earlier (i.e., pre-contemplation) stage of change [38], suggesting important interactions between the kinds of social support that are important for health behavior change and an individual’s position in the change process.

The utility of undirected online social interactions also differs by medium, reflecting each medium’s perceived capacity to facilitate social support [25]. Especially when coping with stigmatized conditions, aspects of the platform and medium play an important role in whether an individual chooses to disclose issues or seek support [91]. One study that leveraged publicly available data from an alcohol use disorder support forum to examine the different ways people use technology platforms when engaging in behavioral change [25]. They found that posts on a public-facing online forum and posts on an online journal, that could be set to be either public-facing or limited to certain contacts, were both far more likely to contain informational support than nurturant support or validation. In contrast posts directly to others’ profiles (akin to a direct message) were more likely to contain nurturant support than informational support. Journals and Forums provide convenient methods for documenting personal experiences with health behavior change and thus may lend themselves well to non-conventional and idiosyncratic information that may be useful when individuals are engaging in behavioral change.

Directed forms of support can also be accessed online to support behavioral change. Directed forms of computer-mediated communication have previously been linked to increased feelings of closeness with others and reduced feelings of loneliness [20]. Directed support also incorporates low-effort interactions such as “likes” or brief non-specific reactions to other users’ shared information or self-disclosures [2]. These low-effort interactions can be delivered in real time or asynchronously and are generally considered reinforcing for users engaged in health behavior change [107]. This type of support is often generalized in that it does not specifically reference the “liked” element of a message or post, but instead provides users with a sense of encouragement. One study [107] examined the role of “likes” and comments in retaining users in an online health community and support network for both patients experiencing health conditions as well as caregivers caring for an individual with a health condition. The authors found that while both comments and “likes” are important for engagement and retention, the value of “likes” changes depending on users’ levels of engagement; encouraging text-based comments were associated with better retention than “likes” when there was ample engagement on a post, but when engagement was sparse, “likes” were more useful than text-based comments for retention. Across studies, even when brief encouragement was provided in the form of a short text response, these low-cost and low-effort interactions served to motivate users to continue engaging in health behavior change or in online communities [1,56,107]. These results highlight the critical importance of building different levels of social support into a digital platform, as both high- and low-effort directed engagement is important for the long-term success of health behavior change and continued engagement in an online community.

When using digital technologies, directed support can further be distinguished according to a number of media affordances that shape the experience of receiving support from others. Past work has highlighted the importance of affordances that include anonymity, visibility (i.e., the number of nonverbal cues, for example “leaner” media such as SMS vs. “richer” media such as video chat), and similarity of users when directly communicating with others [12,46,56,87], and these may moderate the type of social support solicited and how it is received. Together, these studies suggest that direct personal relationships are important for helping users connect and obtain social support for their health behavior change.

2.4. Summary

This study focuses on the ways young adults – the adult age group with the greatest mental health burden and lowest level of treatment seeking [97] – engage with social support across their behavioral change journeys as they move toward improved mental wellness. In particular, we explore, via the stages of change model, the changing social needs of young adults with depression as they move through their behavioral change journeys. We also detail the role technology and different forms of communication play in supporting users at different stages of behavioral change.

3.1. Asynchronous Remote Community (ARC)

We used an Asynchronous Remote Community (ARC) method for data collection. The ARC methodology consists of using an asynchronous elicitation workshop format in which participants respond to researcher- and participant- posted prompts and replies, much like an asynchronous digital conversation. Given that we were primarily interested in understanding the needs and preferences of young adults who were not interested in face-to-face interventions, face-to-face elicitation workshops would not have been an ideal setting as they may have presented barriers for participants who are just beginning their health behavior change journeys and may not yet feel comfortable sharing their personal experiences if they are identifiable to other participants. Moreover, individual interviews were not chosen as the primary modality of data collection because we wanted to ensure participants had the ability to respond to the ideas and experiences of similar others. As a result, we used an ARC methodology which has been used to understand the needs of specific userbases in previous work [e.g. 41,42,53,57], and which can increase engagement of underrepresented populations by decreasing barriers related to in-laboratory studies (e.g., transportation, child-care, time off work) [37]. Additionally, the group format was chosen to facilitate conversation between young adults with shared experiences. These types of remote elicitation workshops have been commonly used within HCI to facilitate conversation, elicit several perspectives simultaneously, and provide the environment for participants to build on each other’s creative contributions [61].

3.2. Participants

Mental Health America (MHA), the United States’ oldest and largest patient advocacy organization, began offering mental health screening tools online in 2014, including the widely used depression screener the PHQ-9. Visitors to their site completing online screeners have increased year-over-year, totaling over one million users completing one of 10 online screeners annually. Among those visitors who completed screeners and reported their age, 1/3 are aged 18-25 and more than 2/3rds have clinically meaningful levels of depression (PHQ-9>=10). Visitors to MHA’s screener site are generally representative of the ethnic and racial diversity of the United States with an overrepresentation of particularly vulnerable groups such as LGBTQ+ individuals. Crucially, the great majority of the young adults with a PHQ-9>=10 who complete the depression screening have not had and do not want formal mental health treatment, such as pharmacotherapy or psychotherapy [72]. Recruitment ads were hosted on screening portions of MHA’s website as the goal of the project was to elicit design principles for digital tools that could be administered by mental health advocacy organizations such as MHA to support non-treatment seeking young adults with depression and reach young adults with depressive symptoms who were uninterested in seeking traditional forms of mental health care.

Participants were included if they voluntarily completed a digital depression screening tool (PHQ-9) and, at the time of screening, were experiencing elevated symptoms of depression (defined by a score of 10 or greater on the PHQ-9). Given that we were primarily interested in the needs and preferences of non-treatment seeking young adults, participants were precluded from participating if they endorsed any psychotherapy or antidepressant medication use or if they were actively seeking mental health treatment. Participants were also excluded if they required immediate psychiatric or medical intervention (e.g., suicidal ideation with a plan and intent to act, psychosis, bipolar disorder, or other serious mental illness for which participation would be contraindicated); if they were located outside of the United States; or were unable to read and write in English.

Participants were 29 self-identified non-treatment seeking young adults between the ages of 18 and 25 (Table 1). Mean age was 21.4 years (SD=2.5). Attempts were made to include diverse participants with demographics similar to those of other MHA site visitors between the ages of 18-25 [72]. Participants identified primarily as female (72%; 21/29), 10.3% (3/29) identified as male, and five participants elected not to report gender. Participants self-identified race as White (44.8%; 13/29); 17.2% (5/29) as Asian; 10.3% (3/29) as more than one race, 3.4%; (1/29) as Black or African American; 3.4% (1/29) as American Indian or Alaskan Native; and 20% (6/29) elected not to report race. A minority of our sample identified as Hispanic or Latinx (6.9%; 2/29).

3.3. Data Analysis

We employed a stepped thematic analysis [16] using transcripts from the ARC study. Participants generated a corpus of 49,226 words. Transcripts were open coded at the level of paragraphs or complete participant responses. Four coders, each trained in thematic analysis procedures [16], open coded overlapping data slices amounting to ½ of the full corpus for each coder such that each coder’s slice of data overlapped with at least two other coders’ slices of data. Coders then met to discuss their results and generate the initial codebook. The codebook was then iteratively revised over three rounds of coding to focus the codebook on codes that were most prevalent and central to answering the research questions, to exclude less relevant codes, to consolidate overlapping codes, and to clarify definitions to better encompass the data. In each subsequent round of coding, coders reviewed and applied codes to 1/4 of the dataset. Coders met and discussed discrepancies and resolved them through a consensus process. After three rounds of coding, coding consistency was reached and the final codebook developed. At that point, each coder was familiar with all the data and was randomly assigned ¼ of the data to code using the final codebook. Following the coding of the full corpus, coders met to organize higher order thematic codes. Codes were applied to the data using a qualitative data analysis software, Dedoose [115].

Importantly, as part of this thematic analysis, the authors coded participants statements as being representative of a particular stage of change based on literature that describes the core features of each stage. Participants, themselves, were not categorized as being exclusively in one stage or another. Thus, some participants may have made statements that were representative of multiple stages of change. Additionally, the authors applied codes representing each stage of change aside from the pre-contemplation and termination stages (contemplation, preparation, action, maintenance). Because inclusion criteria for this study required that participants experience at least some distress or impairment related to their depressive symptoms and be actively seeking self-management strategies for their depressive symptoms, they would not be in the pre-contemplation or termination stages.

4.1. Types of Support: Directed and Undirected Support

This section addresses RQ1 by detailing the types of online support young adults found helpful as they engaged in symptom self-management. The directness of support played an important role in supporting participants’ health behavior change. Participants highlighted the importance of obtaining social support by way of more undirected means, where they engaged with others’ public content, or shared their own content broadly without knowledge of who would be accessing it (e.g., through social media use such as Twitter, Instagram, and TikTok). Undirected support provided social connection without centering on any party, but rather on a shared activity or topic. For example, P25 highlighted that certain social media accounts that deploy public health campaigns, rather than engaging directly with participants, normalizing mental health challenges and effective coping were very useful,

“It’s been really helpful to make my instagram home screen be filled with both aesthetic and happy things while also making sure to have a lot of realistic things such as posts by [the mental health advocacy group] and other people who are open about their live[s].” (P25)

While Instagram and social media were sources of ambivalence for many young adults, some like P25 found ways to curate a feed that served to provide them with support. Participants highlighted that controlling the directness of the support and matching it to their needs across the health behavior change process was important for symptom self-management.

Other participants highlighted the important role directed support played. Direct modes of obtaining social support included face-to-face talking or spending time with friends, telephone calls, texting or online messaging, and email. Direct social connection with others provided participants with a powerful coping tool. Participants discussed ways they elicited direct social support from others using technology. For example, participants highlighted the ways they use technologies to maintain and develop their social connections,

“I enjoy texting people when I need them. Although it has to be specific people I text. People telling me its ok to be sad makes me feel like I’m not making excuses to be sad or depressed. It makes my feelings feel validated. It also makes me feel as if I can get through this and it’s not just something that’s staying.” (P22)

Although users reported a variety of modes of eliciting direct support from others, texting was the most commonly used medium. P22 highlights the intentionality involved in eliciting support from specific others. Participants have a clear sense for who will be the most supportive person(s) to text and often are able to selectively seek out those particular supportive people when they are needed.

Participants described the ways that different media affordances moderated how and when they chose to both give and receive directed support. Media characteristics such as anonymity, visibility (e.g., ranging from “leaner” media such as SMS, which convey few nonverbal cues, to “richer” media such as video chat), and the similarity of others online were important considerations that helped participants feel more comfortable with directed support. For example, when discussing the hesitancy to seek in-person therapy, several participants noted that one of the key barriers was the loss of anonymity,

“I would not seek help from a professional over the phone or online/apps unless it was anonymous (such as something like this [a reference to the anonymous nature of the ARC elicitation workshop]). I am only able to open up if I feel detached from what I’m saying and that I can hide behind it in anonymity.” (P1)

Anonymity, as described by P1, allows young adults to express their feelings and emotions without fear of stigmatization or suffering wounds to ones’ sense of identity. Relatedly, other participants discussed how control over their online visibility by using media with different levels of nonverbal cues (e.g., low: text-based conversations, high: audio- or video-based conversations) impacted their comfort expressing different ideas online. For example, participants reported that they would sometimes visibly display their emotion (e.g., crying), when discussing important ideas or difficult experiences. When using richer media, that display of strong emotion might de-center the ideas or experiences being expressed, leading to a preference for text-based communication. On the other hand, a video or audio-based conversation sometimes enabled participants to connect with others in a way that required less planning or cognitive effort, as described by P13,

“I have started to move away from texting long-form conversations though--I find it much easier to, say, catch up with an old friend, via voice call.” (P13)

Participants provide and receive directed social support using many technologies and media with different characteristics. The choices of which tool they use at what time is made with the consideration of the content and ideas being expressed and the cognitive and emotional load they are experiencing.

4.2. Need for Support

In this section we examine RQ2 by addressing the differential social support needs as a function of stage of change. In Table 2, we aggregate and summarize identified social needs as they related to particular stages of change.

4.3. Challenges to Obtaining Support

Young adults face many challenges to managing their mental health symptoms, in this section we specifically address RQ3 by exploring the barriers young adults encounter when attempting to gain social support.

4.4. Summary

These findings illustrate that while young adults with depression experience unique behavioral change journeys, it is useful to examine the process of behavioral change in the service of mental wellness through a series of localized stages of change. Young adults’ social support needs were dynamic depending on their readiness to engage in health behavior changes. Our findings reveal a need for informational support and validation early in the change process (i.e., the contemplation stage). As participants began to overcome the barriers to initiating a new behavioral change, and entered into the preparation stage, planning support and decisional reinforcement became important needs along with motivational support and fortifications to one’s sense of self-efficacy. Many young adults in the preparation stage also found validation in the form of stories or the experiences of similar others to be extremely useful. Once participants initiated some behavioral change and entered the action stage and beyond, there was a need to link up with healthy coping partners who could help provide accountability. Relatedly, after initiating behavioral changes, trusted others were helpful for bridging gaps in motivation, and reducing barriers to re-initiating health behavior change if lapses occurred. Our findings also highlight several significant barriers to behavioral change such as invalidation and stigma, which can serve to interrupt, delay, or reverse one’s forward movement through a behavioral change journey. Participants also highlighted the challenges to technology mediated DMH tools such as perceived negative psychological effects of social media. Beyond these dynamic needs, our workshop revealed that young adults wanted support that spanned a full spectrum of directness and was accessed via media with different characteristics depending on the circumstances and needs of the individual. Taken together, our results suggest the need to consider an individuals’ behavioral change journey and the needs across these journeys when designing tools that aim to support young adults pursuing mental wellness.

5.1. Conceptual Contributions

In the section that follows, we outline the unique contributions our findings add to the literature on mental wellness and behavioral change journeys.

5.2. Design Implications: Facilitating Behavioral Change Journeys

Understanding behavioral change not as a single decision or action, but as a journey in which a user has varying needs depending on their individual experiences is necessary for developing DMH tools that can flexibly meet users where they are and to engage them across their recovery. Social support is one of the most critical components of mental wellbeing and must meet the dynamic needs and preferences of users as they engage in their own symptom self-management. Considering that, social support needs change, as do the preferred modalities of that support, it is crucial that the next generation of DMH tools be designed to facilitate behavioral change while minimizing barriers. In this section, we explore methods of incorporating optimal forms of social support that meet users’ changing needs, as well as leveraging multiple media characteristics to appeal to the broad preferences of young adults. Finally, we discuss design principles that address users’ ambivalence to DMH tools.

5.3. Limitations

A number of limitations are noted. First, while the data collected reflect heterogeneous opinions, ideas, and experiences, we note that our sample of 29 young adults may not be representative of the larger young adult population with depression. Future studies must replicate these findings before generalizing to other populations, conditions, or health journeys. Furthermore, all participants were recruited from the same source: MHA. While MHA is a large community organization that serves a diverse constituency, future formative user-centered design studies may wish to recruit from a variety of online sources. Another limitation of the current work is that participants’ stage of change was not assessed at baseline. In addition, we coded behaviors and reflections that suggest particular stages as they emerge in each comment, and thus a single participant’s data may contribute to understanding multiple stages of change, at different points. This coding approach reflects the fact that behavioral change is non-linear, and often involves moving back and forth through the stages [84]; however, a consequence is that we do not present an overall breakdown of how many users are in each stage of change.

We also acknowledge the need for the mental health community to pursue two parallel activities that address the state of mental health care delivery in the service of increasing access to care. First, we note that it is important to improve self-management tools for those who prefer to self-manage outside of traditional or existing care structures. Second, it is necessary to address and reduce barriers to accessing formal care. In this paper, we have focused on addressing the former, but we acknowledge that work in this domain is incomplete unless the latter challenges are also addressed.

5.4. Conclusions

Our findings suggest that young adults on mental wellness journeys have dynamic and diverse social needs, partially as a function of their readiness for change. It is incumbent on designers of DMH tools to meet users where they are in their behavioral change journeys, which may include incorporating dynamic social elements to support behavioral change. Specifically, our results suggest that young adults may require DMH tools that offer control over the directness of support as well as different media affordances to facilitate user comfort with social interactions. The mental health community must also consider threats to DMH tool uptake and adoption such as ambivalence about technology and its ability to support mental wellness journeys.