To explore the natural patterns of mobile app use for depressive symptoms among primary care patients, this study employed a qualitative design using individual semistructured phone interviews. Ethics approval for this study was granted by the Human Research Ethics Committee of The University of Melbourne (ID: 1543648.9).

Participants were recruited from the Target-D trial (ACTRN12616000537459). This is a large Australian study of a clinical prediction tool to triage and target depression treatment, in which patients with depressive symptoms were directly recruited from primary care practices across metropolitan Melbourne, Victoria [25]. On completion of baseline data collection, participants were stratified into mild, moderate, and severe symptom groups and randomly allocated to receive usual care or a symptom severity-matched treatment as follows: internet-based self-help, guided internet-based cognitive behavioral therapy, or nurse-led collaborative care (for the mild, moderate, and severe groups, respectively).

Participants were eligible for the Target-D trial if they were aged 18-65 years, had current depressive symptoms (assessed during the screening process as scoring 2 or more on the two item version of the Patient Health Questionnaire, PHQ-2), had access to the internet, and had sufficient English language comprehension to provide informed consent. Participants were excluded if they were taking an antipsychotic, had recently started or changed their antidepressant, were currently receiving psychological treatment, or were currently using an internet-based cognitive behavioral therapy program.

The only additional eligibility criterion applied to this study was that participants had completed the 3-month intervention phase of a randomized controlled trial to avoid prompting those in the usual care group to change their approach to mental health care.

Purposive sampling was undertaken from randomly generated lists of Target-D participants, 30 at a time, stratified by depressive symptom severity assessed using the PHQ-9 [26] to ensure a mix of age and gender in the final sample. The interviewer was blinded to whether participants were allocated to the control or intervention groups.

Potential participants were sent an email invitation to take part in the study and followed up via phone call within 2 weeks to establish interest, obtain informed consent, and arrange a suitable interview time. Of the 60 participants who were emailed invitations, 26 were contactable by phone and 16 took part in the study, as seen in Figure 1. Of those who did not participate, 3 did not consent (one did not provide a reason, one was in hospital, and one was uninterested in the study) and 7 initially consented to the interview; however, they were lost to follow-up thereafter.

Semistructured phone interviews were conducted with 16 participants between May and October 2017 by a female general practice registrar who had no prior contact with the participants. This was conducted using an iterative approach, where data collection occurred incrementally and simultaneously alongside data analysis.

Interviews began by asking participants about their current depression management strategies to establish if mobile apps were used. If any mobile apps were mentioned, this was explored in further detail including type, purpose, duration, referral source (ie, how participants became aware of the app), perceived challenges, benefits, and recommendations. If no apps were mentioned, questions focused on attitudes to using mobile apps for managing depressive symptoms. The semistructured interview discussion guide can be viewed in Multimedia Appendix 1.

All participants consented to audio recording of the interviews, which ranged from 10 to 32 minutes. Recordings were subsequently transcribed verbatim by an independent professional transcription service and verified by AP. Field notes were created during and after the interviews and data were stored confidentially on protected university servers. Participants were offered an Aus $20 gift voucher as an honorarium.

To identify the natural patterns of app use, inductive analysis was employed utilizing NVivo 11 Pro software (QSR International, Melbourne, Australia) to organize and identify core themes and subthemes in accordance with Braun and Clarke’s 6 phases of thematic analysis [27]. This consisted of several stages including familiarization, transcription, generating initial codes, searching for themes, reviewing themes, and then defining and deciding on meaningful themes.

An iterative approach was adopted with the initial coding performed by a general practice registrar following the completion of the first set of 5 interviews. Subsequent interviews were then conducted incrementally with data analysis occurring simultaneously, and new codes were added from the dataset to NVivo Pro 11. SF and JG reviewed the transcripts and codes during this process until consensus was reached by all members of the research team in regards to the final themes. Data saturation was achieved (and data collection ceased) when no new themes emerged, which was determined by the point in analysis where no new codes were able to be created in NVivo 11 Pro that provided additional value to the identified themes [28,29].

The age of the participants ranged from 20 to 58 years. Overall, 58% (9/16) were female and the majority had achieved at least a tertiary level of education. Three quarters of participants had PHQ-9 scores indicating moderate to moderately severe depression. On average, participants completed the interview 269 days after receiving their Target-D treatment recommendation (range 99-369 days). Further demographic information is provided in Table 1, whereas Table 2 details app use and participation in Target-D, including whether Web-based tools were recommended.

Overall, 11 out of 16 participants reported using at least one mobile app to manage their depression with 15 apps mentioned in total (Table 3). All apps were free to download and included, but were not limited to, mental health-specific apps. Awareness of the apps came from numerous sources including self-directed research, inbuilt on phone, and recommendations from friends, organizations, and health professionals. Participants could be divided into 2 groups, users and nonusers. Among users, the following 4 main patterns of app use were identified: skill acquisition, social connectedness, inquisitive trial, and safety netting, as seen in Figure 2.

Five participants had never used an app for their mental health. Perspectives were varied with 3 participants expressing a positive view, one who felt apps were unhelpful, and one who was ambivalent. Among these participants who felt positively toward apps but did not use them, access was the most common barrier. The chaotic nature of the app marketplace and lack of knowledge regarding suitable apps could impede app selection. Additionally, a deficiency in technological competency, perhaps correlating with age, could hinder the use of apps.

The perceived discordance between technology and mental health treatment was also a reason for nonuse because some participants did not view technology as a natural response to seeking treatment for their mental health. One nonuser reported concerns about excessive dependency on apps, focusing on the possibility of addiction and social disengagement.

Another viewpoint was that although apps could be beneficial for mental health to an extent, it cannot act as a replacement for social support and human connection. Consistent with the views of many app users, an app can be considered as a supportive tool for depression management but not as a treatment in itself.

A few participants who did not report current app use expressed an inclination to use apps if their situation deteriorated similar to the safety netting behaviors reported by some app users. This suggests that in contrast to the statements above, these participants perceived apps as an effective option for mental health treatment. Their nonuse of apps could be considered in part a reflection of the fact that they did not consider themselves to be currently in need of treatment.

Health care providers had a role in facilitating the initiation of mobile app use for depressive symptoms. In this study, 4 app users engaged with an app following a direct recommendation from a health care provider, which included a general practitioner, psychologist, midwife, and psychiatrist. These were all meditation or relaxation-type apps, and participants demonstrated a willingness to try these apps on the basis of professional suggestion and generally persisted with use if they found it helpful.

However, other app users perceived their app use for mental health independent of professional health care and did not involve their health professionals. There was a tendency among these participants to view professional therapy as separate to their own personal self-directed strategies.

Regardless of whether or not they were recommended apps, participants demonstrated a sense of trust in guidance from professionals with the perception that these recommendations had a strong basis in experience or knowledge.

The role of health care providers in ongoing app use was unclear because no app users reported involving their health care providers following the initial recommendation.

This is the first qualitative study exploring the patterns of mobile app use for depression in a primary care setting, revealing that many participants are already utilizing mobile apps as a way of improving or maintaining their mental health. Prior mental health app research has primarily focused on the prevalence of use and acceptability and attitudes toward apps [12,13,22,30,31] and this, therefore, provides a unique insight into natural app use behavior and the potential role of mobile apps for managing depressive symptoms in the community.

Four main patterns of use were identified, which were not mutually exclusive; some participants indicated different patterns of use of the same app over time. A minority of participants did not report using any app, which is consistent with previous reports of the prevalence of app use in primary care populations [23]. Among those who did report active app use, the most common patterns were using apps to develop or reinforce a particular skill or to enhance social connections.

Importantly, the results of this study suggest that primary care patients use a wide range of apps to manage their depressive symptoms and not only those that are specifically designed for depression management. Interestingly, however, the apps used did encourage engagement in activities that have evidence for efficacy in combating depression, such as acquiring skills through exercise [32] and mindfulness [33] and social connectedness [34]. This suggests that when looking to improve their mental health, primary care patients gravitate naturally toward evidence-based management strategies and should be empowered to engage in techniques they perceive as helpful. This is supported by a study of evidence-based mental health app recommendations suggesting that apps focusing on behavioral activation can empower users to learn beneficial skills and better engage with therapy [35]. Furthermore, self-efficacious behavior changes were demonstrated in a survey where 90% of mental health app users reported increased motivation, desire to set goals, confidence, and intention to be emotionally healthy from their app use [36].

Few app users reported negative effects of apps, although some concerns about addiction were raised by nonusers. That said, the relationship between social networking tools and depression is complex with research suggesting a dual relationship influenced by numerous mediating and aggravating factors. There is a strong implication that the nature of social media use may have more impact on depression symptoms than frequency of use and that the personal perception of Web-based interactions influences mental health; it may be positive if it increases social support or conversely and negative if it increases social comparison and rumination [37]. How patients use and perceive networking tools for social connectedness may be more important than the frequency or type of app used.

Participants did not consider apps as a standalone replacement for traditional professional management but rather as a tool to support and enhance treatment, a viewpoint consistent with clinicians’ beliefs in a qualitative study regarding apps for severe mental health issues [38]. The value in professional mental health treatment was strongly recognized, but participants utilized apps to support activities that coincided with their treatment beliefs, often when accessibility barriers hindered traditional mental health care.

It is also important to acknowledge the role of health care providers in app use for depression. One third of app users (4/11) directly engaged with their app following a health professional’s suggestion compared with 10% of app users in a broader primary care sample in the United States [23]. The higher proportion of health professional recommendations in this study may be a reflection of a number of factors. First, the app marketplace has grown significantly since 2013, as have the popularity and acceptability of apps generally. It is possible that recognition of apps as a genuine option for mental health care has likewise grown. Second, in 2015, a review of Australia’s mental health system resulted in a shift in policy direction toward digital solutions to improving care. As a result, general practitioners have been provided with greater guidance on what mental health apps are available and how to incorporate them into treatment plans [39]. The high level of trust in general practitioners expressed by many participants in this study suggests that they can have a strong influence on initiating app use for depression symptoms. However, it is important to note that the influence of professional support in continuing app use was not determined in this study. This may be due to the perceived discordancy between health care and app use because 70% of respondents in a primary care survey reported that it was not at all or only a little important for physicians to know about their health app use [23]. However, there is some evidence that the therapeutic relationship between care providers and patients has an impact on sustaining app use and enhancing mental health treatment in primary care [22] and this should be an area for future investigation.

In light of the findings above, we suggest that general practitioners consider a holistic approach when supporting patients to engage in mobile technology to improve their mental health. The complexity of depression, which presents with differing physical, psychological, and social issues in every case warrants an approach that is personalized to each individual’s circumstances and needs. A social networking app, for example, may not have evidence for improving depression but for some patients, it may be a more acceptable way of building social connection than more traditional interventions [40]. Thus, considering a person-centered approach and tailoring app recommendations accordingly may provide a more effective approach to integrating apps into depression management. There is some evidence to support this approach, for example, a pilot study of IntelliCare, a modifiable skills-based suite of mental health apps each focusing on a specific task or activity, demonstrated improvements in depression and anxiety scores among app users [41].

It is important to be aware that a wide range of apps, not simply those that are specific to mental health, may be relevant to depression care. Understanding personal preferences and goals may help general practitioners assess what app, if any, would be suitable for an individual and guide appropriate recommendations, especially as the ever-expanding and chaotic nature of the app marketplace remains a significant challenge [14,18]. Ultimately, the end is more important than the means and if patients can view an app as a way of improving or maintaining their mental health, it can reasonably be considered as a part of their mental health care.

It should be noted that, in exploring patterns of app use, one pattern identified was that of nonuse. By employing a personalized approach focusing on treatment goals rather than apps themselves, general practitioners may find that some nonusers can transition into the “inquisitive trial” category as a starting point. This is particularly likely when the main reason for nonuse is lack of knowledge about what app to use or how to use it. However, general practitioners must also remain cognizant that apps will not address unmet needs for all of the 65% of Australians currently not accessing mental health treatment. They may more usefully be construed as one referral option available to general practitioners within a matrix of other options of varying modalities, intensity, and treatment focus.

There are some limitations to this study. First, the findings are based on a small number of general practice patients recruited in Melbourne, Australia, and whether these patterns of use are evident outside this group is unclear. Second, as participants were all under 65 years of age and were recruited from a trial which includes Web-based interventions, this could have led to selection bias because the sample involves a relatively younger and more technology-inclined population. Hence, generalization to all primary care patients would not be feasible.

Further, a substantial proportion of the sample reported depressive symptoms toward the severe end of the spectrum. Guidelines would suggest that this group should be recommended more formal treatment than self-help via an app. However, as a study of the natural use of apps, we see that the inclusion of these participants is a strength rather than a limitation. By understanding patients’ existing management strategies (including apps) and treatment preferences, general practitioners may be better placed to engage patients in more intensive mental health interventions when appropriate, particularly given that apps were viewed as a complement to rather than substitution for formal treatment.

In addition, owing to the cross-sectional nature of this study, the patterns of use identified relate only to the point at which interviews were conducted and do not assess the impact of changes in depressive symptoms over time. We previously investigated use of mental health websites over a 9-year period and found the primary care patients were more likely to report the use of these websites when their depressive symptom severity increased [42]. Given that “safety netters” and some nonusers reported they would consider using an app in the face of worsening mental health, a longitudinal study of whether these intentions are translated into actions would add to our understanding of the patterns of app use.

Finally, this study focused on the perspectives of app users and not the effectiveness of the apps themselves. It is possible that the app use reported here resulted in improved mental health; it is also possible that it made no difference or worsened symptoms. More research is required to explore the potential benefits and harms of app use. For instance, although there is evidence that social media apps may facilitate social connectedness [40], little is known about the effects of dependency on mobile apps in depression. It is important for future research to evaluate the patterns of app use that may exacerbate depression as well as those that facilitate its management.

The rise of mobile technology heralds a new era in community mental health treatment. There is growing interest from clinicians, consumers, and developers in harnessing the potential for mobile apps to address some of the barriers to traditional mental health care. The results of this study suggest that like any mental health intervention, apps are used only to the extent that they align with patient preferences and treatment goals. Further, although individual apps have limited evidence for effectiveness, patients naturally gravitate toward apps that encourage behaviors known to improve depression. General practitioners may capitalize on this by considering apps as another option in their referral toolbox, a new way for patients to access and engage with old interventions and not a new intervention altogether.