This qualitative study was underpinned by an end-user development framework [24] and based on semistructured interviews as part of a broader blended treatment research project at Orygen Digital, Orygen’s digital mental health arm. End-user development has traditionally been used in the human-computer interaction and software development fields to anticipate and flexibly address the needs of the intended user [24]. With the advent of digital intervention in the psychological field over the last decade, end-user development is now widely used in psychological research as an effective framework to gather and incorporate relevant feedback into the design and implementation of an intervention in a cost-effective and timely way [24].

In this study, participants had the unique experience of completing a previous 18-month to 2-year face-to-face engagement period with the Early Psychosis Prevention & Intervention Centre, followed by participation in a long-term digital mental health intervention known as Horyzons. The Horyzons randomized controlled trial [3] randomly allocated participants, who were from 16 to 27 years of age, to either an intervention (n=85; access to the Horyzons platform in addition to treatment as usual) or a control condition (n=85; treatment as usual). The Horyzons platform was a multicomponent digital platform that included evidence-based online therapy (Figure 1), peer moderation, clinical moderation, and an interactive online social network (Figure 2). Young people assigned to the intervention were allocated a key moderator who supported the participant in completing relevant therapy modules, participating in group discussions, interacting on the social network, and engaging in individual web-based conversations with clinical and peer moderators. The platform is discussed in length in the Horyzons trial protocol paper [3]. The Horyzons randomized controlled trial was registered in the Australian New Zealand Clinical Trials Registry (ACTRN12614000009617).

Young people were eligible to participate in this study if they (1) were previous clients of the Early Psychosis Prevention & Intervention Centre in Melbourne, Australia and (2) had been allocated to the intervention arm of the Horyzons randomized controlled trial.

Young people who met the criteria (n=10) were randomly contacted via phone call or text and invited to participate in a semistructured qualitative interview exploring their perspectives on the design and implementation of a blended model of face-to-face and digital mental health treatment. All 10 consented and were interviewed for the study.

Ethics approval was obtained from the Melbourne Health Research and Ethics Committee in 2018 (HREC/13/MH/164). Data were collected from October 2018 to March 2019. Participants specified their preferred interview locations, and as such, interviews took place in a variety of locations which included the Orygen clinic, libraries, cafes, and participants’ homes. A semistructured interview guide was designed to explore participant perspectives of blended models of face-to-face and digital treatment for first-episode psychosis. The interview questions focused on perspectives on blended models of care, perspectives on how to deliver blended models of care, and perspectives on functions to include in a blended model of care. Author LV had previously established relationships with all of the participants as a research assistant on the Horyzons randomized controlled trial. Before the interview commenced, participants were provided with a plain language consent form and an information statement detailing the study, and they were given the opportunity to ask questions about the study. All interviews were audio-recorded and transcribed verbatim. All participants were reimbursed Aus $20 (approximately US $13.91) for their participation.

Given the end-user development framework and exploratory nature of the study, thematic analysis was considered the most appropriate method of data analysis [25]. Author LV completed the analysis and was supervised in this process by senior author SB. SB and LV had regular face-to-face meetings in which codes, themes, and LV’s thematic interpretations were interrogated thoroughly by SB to ensure rigor. Any disparity between authors regarding the analysis was debated until a resolution was reached. There were no a priori themes; all themes were derived directly from the data during the analysis process. Familiarization with the data was achieved by reading and rereading participant transcripts. Initial codes were assigned to the transcript to signify meaning, and codes that were similar within and between transcripts were noted. Aligned with a thematic analysis framework, the codes within and between transcripts were grouped into preliminary themes and reviewed in relation to all other themes. Some themes were superordinate—that is, the theme represented a whole category—while other themes were subordinate to the larger themes—these became subthemes [25]. In accordance with Morse’s [26] recommendations to maintain rigor, thick and rich descriptions of the themes were written up, these rich descriptions were then further debated between authors LV and SB.

Participants (n=10) aged 19 to 28 (mean 23.4, SD 2.62) years were included in the study; 70% (7/10) were female, and 30% (3/10) were male. Female participants were all cisgender. Of the male participants, 2 were cisgender and 1 was transgender (Table 1).

Three themes emerged relating to young people’s perspectives on the design and implementation of a blended model of care in first-episode psychosis treatment: (1) blended features, (2) cautions, and (3) therapeutic alliance (Figure 3).

This qualitative study used an end-user design approach to explore young people’s perspectives on blended models of face-to-face and digital care in first-episode psychosis treatment. We found that perspectives could be grouped into three overarching themes and that these themes can be practically applied to the design and implementation of emerging blended models of mental health care.

We found that young people were very enthusiastic about the prospect of blended models of mental health care, in so far as it was used to enhance their experience of traditional face-to-face treatment but not to replace it overall. Aspects of blended treatment that could enhance clinical care were readily identified by young people as increasing accessibility, continuity, and consolidation; accessing posttherapy support; strengthening the relationship between young person and clinician; and tracking personal data that could be used to better inform clinical decision making. Overall, young people communicated that the digital experience could not embody the communication characteristics of face-to-face treatment. Tone, body language, and facial expressions were identified as important aspects of communication that young people felt would be lacking in a purely digital space. Previous research [28], however, has suggested that the benefit of online anonymity is particularly beneficial to some participant groups, and this benefit can work to counterbalance the loss of nonverbal communication. Furthermore, the potential that young people may underreport symptoms on mood tracking tools was identified. Concern centered on fears of clinicians and services responding to changes in clinical states in unwanted ways, such as changes to medications when reporting low mood. This has been reported in other studies [29,30], reflecting the importance of considering the role of, and response to, personal clinical information collected in daily life from users of clinical services.

Tracking mood, together with the environment and related keywords, was identified as a useful tool to guide positive decision making. Ecological momentary assessment is a method of recording momentary experiences using a mobile device in the context of daily life [31]. There has been recent interest in the use of information collected in daily life for clinical purposes. For example, one research group [32,33] explored the integration of ecological momentary assessment with standard psychological therapy for people with voice-hearing experiences. People tracked their voice-hearing experiences (and the context in which they occurred) over a week, allowing a clinician to analyze the patterns of occurrence. This was used to provide a formulation of the voices, which informed the selection of relevant intervention strategies. In our study, the use of tracking tools or ecological momentary assessment to identify mood patterns that could be subsequently shared with a clinician to the benefit of the young person's therapy was identified as potentially useful. The incorporation of a digital tracking tool could support young people working collaboratively with their clinicians in identifying patterns of symptoms, environments, and relationships with other determinants to inform treatment decisions and enable momentary interventions. This form of collaboration between young people and clinicians is aligned with autonomous and strength-based practice [34]. Importantly, in line with themes from our study, findings of qualitative research have suggested that this approach may fast-track treatment processes by enhancing the personalization of therapy, which would enable more individualized and targeted intervention strategies to be produced more efficiently based on the information gathered [35]. Additionally, an ongoing connection with a clinician, who has access to personal insights about the young person through their data, has the potential to enhance the therapeutic alliance.

The therapeutic alliance refers to the goal-directed, collaborative relationship between client and clinician and is the strongest predictor of client mental health outcomes [36,37]. According to Bordin [38], the alliance comprises three core conceptual elements which include bond, tasks, and goals [36]. One of the strongest findings identified in this study was the perception that a blended model of treatment had the potential to enhance the therapeutic relationship. Integrating a digital platform with the face-to-face experience was consistently endorsed as a means to strengthen the relationship between the young person and their clinician by providing greater consistency (ie, accessible sessions and more regular contact), collaboration (ie, joint decision making), accountability (ie, a sense of commitment to complete tasks on time), and the opportunity to consolidate information. This supports previous findings from Lederman et al [37], in which the experience of technology-mediated mental health therapy was examined and that, like Horyzons, included therapy pathways, a social network, clinical moderation, and peer moderation; Meridian was designed to support carers of young people experiencing mental illness. Lederman et al [37] found that the digital system could create an alliance-like experience for the users. Their research identified that it was likely that the combination of peer support workers, clinical moderators, and different core functions of the platform came together to “facilitate the formation of a state that mirrors therapeutic alliance [37].” The finding that a therapeutic relationship is possible beyond that of only the client–clinician experience is valuable knowledge to the digital mental health field, particularly because the therapeutic alliance is the strongest predictor of mental health outcomes [39].

In addition to the emerging possibility that the therapeutic alliance can develop beyond that of only the client–clinician relationship to incorporate a blended system (resembling client–technology–clinician), the notion of a digital therapeutic alliance between the user and the technology itself (ie, client–technology) is also a burgeoning area of research in the mental health field [27,37,40]. Among other technologies, chatbots are an avenue of such interest. Vaidyam et al [27] identified chatbots as an “effective” and “enjoyable” technology in the mental health setting. Chatbots were similarly acknowledged in this study, as a helpful venting tool and a plausible solution for a mental health service to provide sustainable 24-hour access. The concept of a digital therapeutic alliance between young people and technology or the digital platform itself is a notable avenue for further research.

For many young people who are native users of digital technology within their everyday lives [15] and are interested in receiving care via blended modalities, there is an important opportunity to assist them in strengthening connections with their therapist and offering ways to improve engagement with therapies. As the therapeutic alliance is considered one of the most important predictors of treatment efficacy regardless of the therapeutic approach [39], there is great potential in capitalizing on the ways to enhance this alliance through digital technologies or with digital technologies themselves [40]. The therapeutic alliance that develops through blended treatment has the potential to enhance engagement with both face-to-face and digital therapies through increasing therapeutic intensity and personalization, and by simultaneously addressing poor engagement and treatment continuity which are key issues in the digital health field [13,17].

Future research is needed to investigate the efficacy of blended models of care by evaluating its impact on the therapeutic alliance, engagement, and treatment effects. Additionally, the formation of a digital therapeutic alliance between a user and a digital platform or other technology is an emerging area of research worthy of further exploration.

Author LV had established relationships with all participants due to a prior role as a research assistant on the Horyzons randomized controlled trial and also completed all interviews in this study. While we found that the existing relationship could be conceptualized as a strength that possibly aided in creating a more open dialogue between interviewer and interviewees, a previous relationship generates an opportunity for participant bias as interviewees may, both consciously and unconsciously, provide answers that they believe will please the interviewer as opposed to what they may genuinely feel [41].

Young people were very enthusiastic about the prospect of blended models in first-episode psychosis mental health treatment. Aspects of blended treatment that could enhance clinical care were readily identified by young people as increasing accessibility, continuity, and consolidation; accessing posttherapy support; strengthening the relationship between young person and clinician; and tracking personal data that could be used to better inform clinical decision making. Tone, body language, and facial expressions, however, were identified as important aspects of communication that young people felt may be lacking in a purely digital space. Furthermore, the potential of underreporting symptoms through mood tracking was identified. Blended care was identified as an avenue through which the experience of mental health treatment could be enhanced. Future research is needed to investigate the efficacy of blended models of care by evaluating its impact on the therapeutic alliance, clinical and social outcomes, cost-effectiveness, and engagement.