Abstract
The therapeutic alliance is a well-established predictor of psychotherapy outcomes. With the advancement of digitization, guided internet-based cognitive behavioral therapy (ICBT) has demonstrated effectiveness. However, the conceptual framework of the digital therapeutic alliance (DTA) within ICBT remains conceptually underdeveloped. Data were collected from a guided ICBT program in China. We conducted virtual individual semi-structured in-depth interviews among a purposive sample of intervention recipients (young sexual minority men, n = 20), and all implementers (counselors and supervisors, n = 6). Data were analyzed using qualitative template analysis. A DTA framework was developed with 4 key components: roles, interactions, characteristics, and influencing factors. Clients typically initiate the counseling process through online materials and assignments, treating the digital platform as a “safe space,” “invisible friend,” “warm harbor,” or “repository of knowledge and tools.” Counselors serve as guides and provide professional support asynchronously. High frequency of interaction and emotional connection via phone calls strengthened DTA. “Being client-driven,” “fostering an equal relationship,” and “being difficult to establish but easy to terminate” were identified as characteristics. Influencing factors included client-related aspects (eg, motivation), counselor-related elements (eg, response timeliness), and platform-related considerations (eg, trust, confidentiality). This study develops a DTA conceptual framework, elaborating on the roles, interactions, characteristics, and influencing factors among clients, counselors, and digital platforms in guided ICBT. Findings inform strategies to enhance engagement, counselor training, and platform design in digital psychosocial interventions.
Keywords
Introduction
Therapeutic alliance (TA) is a fundamental element in psychotherapy, encapsulating the collaborative and trusting relationship that develops between counselors and clients. 1 A stronger alliance has consistently been linked to better engagement, adherence and clinical outcome across both traditional psycho-dynamic therapies and more contemporary psychotherapies, including cognitive behavioral therapy (CBT).2 -4 CBT has been endorsed by the World Health Organization as an effective psychotherapeutic approach to enhance mental health, 5 and better TA in CBT has been demonstrated to facilitate greater decreases in depression and anxiety.6 -9 In face-to-face psychotherapy, TA is most commonly conceptualized through Bordin’s pan-theoretical tripartite model, which distinguishes goal, task, and bond as core domains of the counselor–client alliance.1,10 Studies have further detailed the roles of both counselors and clients within this framework and identified important factors such as counselor empathy, client engagement and mutual responsiveness for developing and maintaining alliance.10,11
Although face-to-face counseling has important benefits, it can be constrained by access barriers (eg, travel and scheduling demands) and affordability, whereas online delivery has been proposed as a way to expand access and convenience and potentially reduce costs.12 -15 Asynchronous guided Internet-based CBT (ICBT) is a form of psychological intervention that delivers structured ICBT content online, where clients receive guidance from mental health professional without real-time interaction and communication typically occurs via secure messaging, allowing participants to engage with the material at their own pace. 16 Empirical studies indicate that a TA can be established in ICBT. However, both alliance levels and their associations with clinical outcomes (eg, symptom change) vary across studies, and high alliance ratings do not always translate into stronger symptom improvement.17,18 This variability highlights the challenges of applying and comparing traditional TA measures in ICBT, where alliance processes may vary with intervention configuration (eg, type, intensity of guidance, asynchronous communication), population characteristics, and study design.
In this context, the notion of a “digital therapeutic alliance” (DTA) has been proposed to capture relational processes that involve not only clients and counselors but also digital tools and platforms. 19 DTA may overlap with traditional TA while also incorporating technology-related elements (eg, trust in the digital system and perceived responsiveness), which may shape alliance formation in digital care.20 -23 At the same time, DTA research remains conceptually heterogeneous and often modality-general. Some authors argue that DTA should be defined as a broader relational system that explicitly incorporates human–computer interaction and users’ relationships with digital tools alongside clinician relationships, 19 whereas evidence from text-based digital psychotherapy shows that most empirical studies still operationalize alliance primarily as a client–therapist construct using traditional instruments, with technology treated largely as the delivery medium. 24 Recent syntheses further highlight substantial variability in DTA definitions and measures and ongoing uncertainty about whether DTA represents a distinct construct beyond traditional TA, usability, and engagement. 25 Qualitative work also foregrounds alliance-relevant perceptions directed toward the digital programme itself (eg, trust in the programme, perceived interaction quality, and feeling considered), suggesting that technology-facing elements may be salient in fully internet-based care. 26 Collectively, these findings suggest that DTA cannot be assumed to be uniform across digital configurations. Qualitative inquiry is especially useful here because it can unpack how alliance is produced through client activity, counselor input, and platform structures in asynchronous, text-based care, beyond what standard alliance measures capture. Critically, however, most prior work does not specify how DTA operates in guided, asynchronous, text-based ICBT, where alliance is co-produced within a client–counselor–platform triad.
To date, there is limited qualitative work that develops a context-specific framework of DTA in such guided, asynchronous, text-based ICBT configurations, particularly under conditions of stigma and among marginalized populations such as young sexual minority men, for whom anonymity and asynchronous text-based support may simultaneously facilitate access and alter relational processes. The roles and characteristics of counselors, clients, and the digital platforms that are utilized, as well as the factors influencing DTA development, remain insufficiently specified in guided, asynchronous ICBT configurations. This gap necessitates further exploration to understand how the DTA is developed and maintained in digital contexts. Currently, there are ongoing concerns about the potential challenges in forming a robust therapeutic alliance in ICBT due to the remote and asynchronous nature of counselor support. The absence of immediate feedback, reduced non-verbal cues, and delayed counselor responses might weaken emotional connections, and hinder the establishment of trust and rapport.27,28
Thus, this study aimed to develop a conceptual framework of DTA within the context of guided ICBT. We focus on 4 research questions: (1) What are the roles of clients, counselors, and the digital platform in guided ICBT? (2) How is DTA established in guided ICBT? (3) What are the characteristics of DTA in guided ICBT?, and (4) What influences variability DTA in guided ICBT? By addressing these elements, this research seeks to enhance the understanding of DTA and inform best practices in digital mental health interventions.
Methods
Study Design and Participants
This qualitative study was conducted after completion of the parent RCT. Qualitative interviews were used to reflect participants’ experience of how DTA was established. Participants were recruited from the young sexual minority clients and counselors participating in a guided ICBT trial study program in China, which has been described elsewhere. 29 Overall, the program combined 10 self-paced online modules with asynchronous counselor guidance delivered via a web-based platform and 2 brief scheduled calls. Specifically, the treatment (a Chinese culturally adapted affirmative ICBT, called Yisitang) was delivered to the intervention participants (n = 60), and 5 counselors provided asynchronous therapeutic support with 1 supervisor overseeing the whole counseling process. In addition to 10 sessions of ICBT, clients received 2 virtual phone calls from a counselor, which included a welcome call before initiating the treatment to explain the structure of the study and briefly discuss the client’s motivations and goals for participating in the study, and a midterm check-in call after finishing 2 sessions to discuss the progress through the study sessions.
Purposive sampling was used to enroll participants from those clients, maximizing variation in theoretically important demographics (eg, age, education level, marital status, employment status) and response to treatment (large or little change in depression, anxiety, and sexual behavior), ICBT session attendance, and qualitative changes in mental health systems and unprotected sexual behaviors over the course of the intervention. For providers, we used an exhaustive sample of all implementers involved in the program.
Inclusion criteria for clients were enrollment in the parent intervention arm, initiation of guided ICBT (eg, welcome call), and ability to complete a Chinese online interview with verbal e-consent including audio-recording. Exclusion criteria were no initiation, inability to complete the interview, or declining recording. Thematic saturation was assessed iteratively during data collection and template refinement through team meetings. For clients, saturation was reached by the 18th interview, after which no substantively new themes emerged, and the remaining interviews confirmed and enriched the thematic structure. Overall, 20 clients, 5 counselors, and 1 supervisor participated in this study.
Table 1 describes the sociodemographic characteristics of participants. The 20 clients were young sexual minority men, aged between 21 and 31. Most were single, with at least a bachelor’s degree, and session attendance ranged from 2 to 10. Employment and identity disclosure status varied across the sample. The 6 counselors (C001-006, including the supervisor) were aged 30 to 47. Most held doctoral degrees (n = 3), with others holding a master’s or bachelor’s degree. Professional roles included professor, nurse, doctor, and LGBTQ-affirming CBO staff. Reported years of professional experience ranged from 1 to 12 years, across both part-time and full-time positions.
Sociodemographic Characteristics of Participants and Counselors.
P means participant.
Partially means disclosing their sexual identity to either their partners, parents, or friends.
Data Collection and Researchers’ Characteristics
Data were collected through online individual in-depth interviews from June to November 2022. A semi-structured interview guide was developed to cover key domains (eg, experiences with guided ICBT vs in-person counseling; perceived roles of the client, counselor, and platform; client–platform relationship; and the triad interactions). The full interview guide is provided in Supplemental File 1. The guide was not a standardized or validated instrument, however, 2 pilot interviews prior to the formal interviews were conducted by the research team to ensure the quality of the interviews and question approaches.
Each interview lasted between 50 and 120 min. All interviews were audio-recorded and transcribed verbatim. The Interviews were carried out by a female senior qualitative researcher and 2 female doctoral students under the supervision of a qualitative research expert. The senior researcher had more than 10 years of qualitative research experience and taught qualitative research methodology to doctoral students at a university. For client interviews, interviewers had no therapeutic relationship with participants. Counselor interviews were conducted by a researcher who was independent of the program and had no role in intervention delivery or supervision. Interviewers maintained reflexive notes after interviews and held regular debriefings to discuss emerging impressions and potential biases. During analysis, multiple researchers participated in coding and template refinement, and discrepancies were resolved through team discussion under senior qualitative oversight.
Data Analyses
Template analysis was adopted to guide data analyses. Template analysis is a form of thematic analysis, 30 and widely used across disciplines. 31 Template analysis offers structured flexibility, allowing for hierarchical coding of qualitative data and in-depth exploration of complex themes by integrating diverse perspectives of different groups.32,33 This method supports theoretical integration through iterative template adjustments, providing depth and clarity in understanding research phenomena. 34 In this study, we used template analysis because it allowed us to begin with an established alliance framework while still capturing themes that were specific to guided, asynchronous ICBT. We used Bordin’s goal–task–bond domains as the initial organizing structure, then iteratively refined the template as new DTA-relevant themes emerged from the interviews. This approach was a better fit than grounded theory because our goal was to extend an existing model to a digital, asynchronous, triadic configuration rather than build a new theory from scratch.
Nvivo 14.0 was used to support data coding and analysis. 35 We imported transcripts and operationalized the evolving template as a hierarchical network of nodes. The node structure was refined iteratively (eg, adding, splitting, merging, or renaming nodes) with earlier transcripts revisited to ensure consistency. We used code extracts and memos to support theme development. The 6-step template analysis was adopted: (1) Becoming familiar with the text to be analyzed; (2) Carrying out preliminary coding of the data. “A priori” themes were established based on our interview aims, including the roles of clients, the digital platform and counselors; the establishing process of DTA; the characteristics of DTA; the factors influencing with DTA, including sub-themes (eg, the dimensions and attributes); (3) Organize the emerging codes into meaningful themes or subthemes, and begin to define how they relate to each other within and between these themes or subthemes; (4) Developing an initial coding template to align with our 4 research questions, and for the second question how DTA is established, we used Bordin’s model as the initial organizing lens while remaining open to additional DTA-specific themes (shown in Table 2); (5) Applying the initial template to additional data and modifying it as necessary; (6) Finalizing the template and apply it to the full data set. 30 The quotes presented were translated from Chinese to English. Transcripts were coded independently by 2 researchers. Coding and template refinement discrepancies were discussed in regular team meetings and resolved by consensus with remaining disagreements adjudicated by the supervising qualitative expert.
The Analytic Template for Coding DTA in Guided ICBT.
Ethics
The study was approved by the Institutional Review Boards of Yale University (2000029433) and Central South University (E2021128). Informed consent was explained and obtained before the interview started, including the study’s aim, the benefits and potential risks, data confidentiality, and voluntary participation. The transcriptions maintained anonymity, with each individual being given a unique numerical identifier. This study was reported according to the Consolidated Criteria for Reporting Qualitative Research (COREQ) guideline. 36 See checklist in Supplemental File 2.
Results
A conceptual framework of DTA in guided ICBT was developed and is summarized in Figure 1. This framework identified the roles of clients, counselors, and the digital platform, elaborated how DTA was established through the triad interaction process of client, counselors and the digital platform, clarified the characteristics of DTA in guided ICBT, and explored the factors influencing variability in DTA.
Conceptual framework of digital therapeutic alliance (DTA) in guided internet-based cognitive behavioral therapy (ICBT).
Role of Clients, Digital Platform, and Counselors
Role of Clients
Clients consistently emphasized taking the initiative in engaging with the program and driving much of the therapeutic work independently during guided ICBT. They described progressing through sessions via self-learning and self-reflection, and articulating and analyzing thoughts and feelings largely through written assignments. Clients also highlighted autonomy in choosing when and where to engage, setting their own pace, and reaching out to counselors proactively when questions or difficulties arose. As 1 counselor observed, “this form. . .is mainly for clients themselves to learn, to take the initiative to go through the treatment,” likening it to a flipped-class model in which clients “learn and grow by themselves through exercises and homework,” and noting that “the protagonist. . .is the client” (C002). Similarly, a client stressed that “the key is to take the initiative and get it done by ourselves,” investing time and energy to work through the materials and logic, “not led by the counselor, but by ourselves” (P042).
Role of the Digital Platform
Participants described several vivid roles that the digital platform played. Most clients treated the digital platform as a “tree hole” (a widely recognized metaphor for a safe, anonymous space to express personal emotions and secrets without judgment in China), in which they could share “all kinds of emotions” and “have another perspective on [those emotions]” through noticing and analyzing how those emotions happened (P055); most importantly, “nobody knows what I'm doing” (P016). Some clients viewed the platform as an invisible friend, who “could be trusted compared to [my] parents in some aspect,” and was “much more helpful” to relieve their stress, because of its “more professional, more knowledgeable, and [the platform] speaking more reasonably and gently” (P067, P050). Many clients appreciated that this digital platform served as a “warm harbor,” bringing them a sense of belonging. It fostered comfort and assurance when they were navigating through the platform. When reflecting on the platform, several clients described a sense of “coming home” to a familiar, warm place (P077, P063).
In addition to the emotional roles the platform played, most clients also reported that it served as “a repository of knowledge and tools,” which provided them with valuable resources and skills to cope with negative emotions. They believed it was “more like a manual, with various methods and solutions to possible issues” they might encounter (P040), and it equipped them with the capacity to handle sexual minority stress and identify their core beliefs, and that this was as powerful as “finding a needle in a haystack” (P115).
Role of Counselors
Clients and counselors consistently characterized the counselor role as “guide” embedded within a structured, self-paced program. Counselors primarily reviewed clients’ written assignments and messages and provided feedback that helped clients stay oriented, apply skills, and move through sessions. Clients often described counselors using “guide or teacher” metaphors. One client said, “He’s like a guide, or a teacher. . .showing me the way. . .guiding me through the door” into the program (P016). A counselor similarly noted that over time, they came to see themselves as “just a guide,” emphasizing that clients “need to learn and master the logic by themselves” (C006). Clients described counselor input as particularly valuable when they encountered difficulties, using the platform to reach out and receiving tailored guidance that clarified next steps: “she gave me great guidance. . .after following her advice, it really made me feel enlightened” (P012).
How DTA is Established in Guided ICBT
Establishing Goal Alliance
Goal alliance was typically formed early through a structured onboarding sequence rather than developing gradually through later sessions. Goal setting began with the initial “welcome” call, after which clients entered the platform and were guided to formulate goals through the program’s introductory materials and the first module. As 1 client described, “after the first call, I went onto the platform to start learning the introductory images (a summary of the 10 sessions), and the first session helped me understand how to set goals, I felt like everything would be really organized to help me” (P052). In this process, goals were articulated through structured self-reflection on the platform, with counselor input delivered asynchronously when needed. Another client noted that this early structure helped them “clarify [their] struggles and intentions in the very beginning” (P083), even without real-time interaction. This pattern appeared to be driven more by the onboarding design (ie, welcome call and structured prompts) than by the timing of asynchronous exchange itself.
Establishing Task and Bond Alliance Through Interactions of the Triad
Clients established DTA by proactively learning the counseling materials on the platform. They sent their feedback, thoughts, and analysis to the digital platform, and then received professional therapeutic support from the counselors through the platform. The strength of DTA depended on how proactively and frequently the client provided information to the platform.
In this triadic process, participants described the strength of DTA as varying with how proactively and consistently clients contributed information on the platform. When clients maintained steady progress (eg, logged in regularly, completed homework on time and in full, or raised questions), counselors were able to provide timely guidance through written feedback. By contrast, when clients did not actively provide information on the platform, counselors had limited material to respond to in this asynchronous format, and clients reported that they “could not get any feedback” (P015). In this workflow, counselors’ role in initiating contact was often constrained, with support delivered mainly in response to clients’ inputs.
The digital platform was the bridge between clients and counselors. It received information from clients and counselors, and then exchanged this information between the 2 parties. The platform could be bundled with the counselor in the eye of the client, while it could be bundled with the clients from the perspective of the counselor. Thus, when clients expressed that they missed the platform, they might really miss the counseling content or the professional support from the counselors. To some extent, in the asynchronous ICBT counseling context, clients developed attachment to the platform rather than to the counselors.
P016: I used to login on the platform every week. When it ended, suddenly I felt kind of empty, mainly because there are no sessions. . ..maybe I lost a way of communication. (he completed 10 sessions)
In this ICBT counseling program, the interaction also included 2 phone calls made by counselors to welcome and check-in on the ongoing counseling process for the clients. Most participants reported that the initial welcome phone call and midterm check-in call made them believe they were “chatting with someone online [rather than a robot] and thus they could establish human relationship and connection” (P095), or they could feel the counselors’ attitude “to determine whether they dislike you or stigmatize you” (P053), so as to establish the trust between clients and counselors. Therefore, the 2 phone calls could bring “real-time” connections, which were very helpful to ensure the establishment of DTA.
Characteristics of DTA
Participants described several distinctive characteristics of DTA in asynchronous guided ICBT. First, DTA was often experienced as a more client-driven approach. Clients typically drove engagement and maintained it through ongoing interactions with the platform. Participants described “curiosity about upcoming content” and perceived “benefits from prior sessions” as reasons for continuing (P095, P083). Moreover, participants noted that perceived benefits largely relied on clients’ motivation and capacity for self-learning. The counselors often relied on the digital platform to deliver and sustain alliance-supportive communication, and described their role as “more responsive than initiating” rather than leading the interaction (C004). In this workflow, the counselor’s outreach was largely limited to communication through the digital platform.
P097: This kind of online counseling, if you do not learn consciously, the learning effect may not be so obvious. C005: I think we [counselors] are constrained by the platform. I could only contact them through this platform, so if they didn't log on to the platform, I couldn't do anything about it, so I just waited and left messages on the platform. . . . If he did not want to be involved anymore, then I couldn't reach out to him.
Second, clients and counselors developed an “equal relationship” in DTA, which was very different from in-person TA. Clients expressed that they did not fear to show their weakness or inner vulnerability on the platform, and they felt that there was no coercion since they controlled the rhythm of the interactions. They felt they were treated equally as a friend.
P016: Because they can't see me, and I can't see them, so we both engage in a dialogue on equal footing, and just feel like there's a thread connecting us. P027: It didn't feel like being in school [as expected], there was no sense of hierarchy, the atmosphere [of dialogue with counselors] was great. It felt like communicating with a friend.
Third, participants perceived initiating DTA as more challenging compared with their understanding of in-person TA. Both clients and counselors said it was challenging to establish trust at the beginning of the internet-based interaction, as it is an anonymous and invisible, lacking physical interaction, voice communication, or sensory connection.
C003: At the beginning of setting up a counseling relationship through a digital platform and I was wondering, "You can't see the person, how are you supposed to build that counseling relationship?" . . . Because this counseling relationship largely depends on communication through the platform by written words, without that face-to-face visual, voice or sensory connection that you get offline. P040: When it comes to online stuff [e.g., establishing a relationship with someone], I feel like it's a bit fake. . . because you can't see or touch it.
Finally, emotional bonding was perceived as less strong, and participants described DTA as easier to terminate. They attributed this mainly to asynchronous, text-based communication with fewer real-time exchanges and nonverbal cues, while the platform-based format also made disengagement low-effort.
C002: Relationships like the face-to-face consultation tend to be more intimate [than DTA], because I only know a story [through the platform], and it doesn't correspond to specific real-life people, so I think this counseling relationship will be a little more distanced. P097: . . .on this website, it [the communication] might be too mechanized. I feel it lacks that kind of, you know, that face-to-face, an environmental feeling. The relationship is not as close as [that in] face-to-face [counseling], I think.
Influencing Factors of DTA
The communication mode of asynchronous guided ICBT is a structured, text-based, delayed interaction process mediated by a secure digital platform, where clients and counselors engage in time-independent therapeutic dialog. This communication structure, which replaces synchronous verbal exchange with written, asynchronous feedback, fundamentally shapes how therapeutic goals are negotiated, tasks are implemented, and bonds are formed. As such, the communication mode represents a core determinant—if not the most critical factor—in the establishment and quality of DTA in guided ICBT settings. Therefore, any factors from the client, platform and counselor’s perspective could influence the DTA establishment process.
Clients’ Influencing Factors
Two client-related factors appeared to influence the development and maintenance of DTA. First, motivation shaped sustained engagement and help-seeking during guided ICBT. Clients who reported stronger motivation described taking sessions more seriously and benefiting more from the program, which supported continued participation. For example, 1 client stated, “I really hoped to understand my emotions. . .reduce my stress. . .So, I took my sessions seriously, and I’ve gained a lot. . .As long as you’re willing to learn, you will definitely benefit” (P115). Counselors also noted that motivation could fluctuate over time and might decrease when participants encountered a steep learning curve as the program progressed: “[having a learning curve] is quite normal. . .people tend to become less motivated later on, even as the workload keeps increasing” (C004).
Second, personality traits were described as shaping how comfortable clients felt in asynchronous guided care. Some participants suggested that more introverted or socially anxious clients experienced the text-based and anonymous format as less socially demanding and more comfortable. As 1 client explained, “I’m more of an introverted type. . . I would prefer [an online] counselor who gives more space. . . face-to-face [counseling] might always be some discomfort” (P089). In contrast, counselors observed that more extroverted clients often sought more direct interpersonal contact, and could feel less engaged in primarily text-based communication: “Extroverted clients often need more follow-up and live interaction—they don’t always feel engaged through just writing” (C001).
Counselor Influencing Factors
Counselors perceived ICBT as less time-intensive than in-person CBT, as their work primarily involved reviewing clients’ assignments and providing written responses and guidance. However, clients emphasized that this format still required strong professional counseling skills to quickly grasp clients’ core concerns from limited text-based input and provide precise guidance. They also noted the importance of conveying empathy, a non-judgmental stance, and caring through written feedback to establish rapport and trust. One client said that the counselor’s understanding was “really strong,” that she could “immediately get what I mean,” and was “really good at picking up subtle cues from a single sentence,” making the interaction “a very enjoyable journey” (P067). Another client described experiencing the counselor as “not just a cold artificial machine,” but as “real human’s care,” with “some human touch” conveyed through messages (P061). A counselor similarly noted that they “needed to be sensitive to clients’ emotions” through the platform and respond in ways that “help clients feel cared for” and recognize the counselor’s professionalism (C005). Counselors in our program not only needed to identify key points from the limited text-based information clients provided, but also needed to consistently affirm and encourage clients, and create a supportive environment, so as to establish a good DTA.
In addition, both clients and counselors highlighted response timeliness and integrity as important for maintaining DTA over time. Clients described that prompt, detailed feedback and consistent encouragement increased their motivation to continue learning and stay engaged in the program. For example, 1 client stated that the counselor “can give me motivation [to continue the next session],” that her replies were “quite detailed,” and that “if there is feedback. . .every time” along with encouragement, he would “definitely be more motivated to learn” (P097).
Finally, counselors described supervision and peer support as helpful for sustaining their confidence and for managing relational challenges that arise in guided ICBT, especially when clients were not proactive in responding or interacting. In our study, counselors reported benefiting from weekly peer support and monthly senior supervision to discuss how to interact with clients and how to maintain relationships when engagement was low. One counselor noted that weekly consultation and supervision were “very important,” as they shared cases and discussed what skills to use and how to respond to common difficulties, such as clients “not being proactive in responses or interactions,” and how to increase motivation while maintaining a good relationship (C003).
Digital Platform Influencing Factors
In this online counseling, the digital platform served as a mediator between clients and counselors. Any factors that affect the use of the platform would critically affect the establishment and maintenance of DTA.
Clients would base their trust in the source and confidentiality of the digital platform, including who was running this platform. The reputation of the implementing agency (eg, a renowned university) brought clients high trust in the confidentiality of the platform. One client noted, “this is the collaboration with XXX University, right? So I don’t need to question it. . . I won’t have any doubts about the other aspects of the program” (P061). Another similarly stated that having “well-known hospitals or schools as a guarantee” increased trust, “maybe in terms of confidentiality. . . so I really trust you guys” (P012).
Clients also highlighted user-friendliness as an important factor shaping their attachment to the platform and willingness to return. This included whether the interface was designed with a modern style and included multiple types of illustrations (words, graphics, or videos), attractive background elements, and other design factors. In our study, some young clients complained about the interface of the platform, and regarded it as “dull and boring,” “old-fashioned” or “antiquated,” and thus it greatly affected their motivation and enthusiasm for logging on the platform. Clients preferred more situational, interactive, and engaging presentation formats (eg, Wechat), which could create a more immersive experience, enhance user satisfaction, and thus improve the development of DTA. For example, 1 participant said the interface felt “rather dull and boring. . . [and] brings you that kind of age-old traditional feeling,” adding, “sometimes I just don’t like logging in” (P067). Another client suggested that more interactive and contemporary formats (eg, short videos or episodic animations) might better match current digital habits, noting that platforms like WeChat and TikTok have shaped people’s expectations for digital content and engagement (P027).
Discussion
This study developed a conceptual framework of DTA within the context of guided ICBT, informed by Bordin’s goal-task-bond model as a reference point and grounded in real-world asynchronous therapy interactions.1,10 Our findings suggested that DTA is not merely a replication of traditional TA in digital form, but a distinct, dynamic process shaped by the triad interactions between clients, the platform and counselors, and the asynchronous structure of guided ICBT. By specifying each role of the triad, the DTA establishment through the triad interactions, and the characteristics of DTA and the influencing factors of DTA, our framework offered a practical map for understanding where alliance-building succeeds or breaks down in an asynchronous guided ICBT context.
Using Bordin’s model as an analytic lens,1,10 our findings suggested how DTA was established through the 2 linked processes: early establishing goal alliance during onboarding and subsequent establishing task and bond alliance through the triadic interaction. The goal alliance in DTA was found to be largely consistent with that in traditional settings, as clients were able to articulate meaningful goals through the initial “welcome” call and structured prompts in early sessions. This suggested that the digital format does not hinder the ability to collaboratively define therapeutic goals, even in the absence of real-time dialog. By contrast, task and bond alliance are more context-dependent and are reconfigured by the asynchronous, self-paced design. Task alliance shifts toward a more client-driven form of therapeutic work, where clients must initiate reflection, complete structured assignments, and decide when to seek counselor input, rather than continuously co-negotiating tasks in real time. This format emphasizes individual responsibility and aligns with the client-driven nature of DTA, contrasting with face-to-face therapy, where tasks are often co-negotiated and adjusted in real time. Bond alliance is likewise impacted by the asynchronous communication structure. The absence of synchronous dialog and nonverbal cues limits opportunities for emotional attunement and relational depth. However, for clients with social anxiety or a preference for written communication, the asynchronous environment can function as psychological safety. These findings highlight that while the structural elements of the alliance remain, their expression is transformed through the triadic interaction between client, counselor, and platform,37,38 and it also extends Berger’s observations by illustrating how task and bond constructs are reshaped through platform-mediated, time-independent exchanges. 27
Recent work agrees that alliance can exist in digital care, but it still disagrees on what DTA should include and what role technology or platform element actually plays. Some authors argue that DTA should explicitly include users’ relationships with digital systems and human–computer interaction, rather than treating technology as a neutral channel. 19 By contrast, a scoping review of text-based digital psychotherapy shows that most empirical studies still measure alliance mainly as a client–therapist relationship, with the platform treated as delivery infrastructure. 24 An integrative review further highlights how inconsistent current definitions and measures are, and questions whether DTA is truly distinct from traditional TA, usability, or engagement. 25 Our findings help move this debate forward by specifying what DTA looks like in a concrete configuration: guided ICBT delivered through an asynchronous, platform-mediated triad. Our contribution is to specify the boundary conditions of DTA in guided ICBT, where digital platform is an active part of the alliance system rather than a neutral channel, because alliance work is distributed across the client–counselor relationship and the platform’s structures and affordances. This is consistent with qualitative findings by Taylor et al, 39 who reported that clients experienced alliance as emerging not only from human interaction but also from platform affordances that felt secure, personalized, and emotionally responsive. This also fits with qualitative work in fully online care showing that programme-directed trust and feeling “considered” can become central when relational cues are filtered through text and interface design. 26 Taken together, our framework shifts DTA theory from broad, modality-level descriptions to a context-specific map of how Bordin’s domains are enacted and reshaped in guided, asynchronous ICBT, which can inform more precise measurement and more actionable platform design.
Building on these establishment processes, our findings also highlight why DTA in guided ICBT can feel “hard to initiate but easy to terminate.” At the beginning, alliance is often described as being initiated largely through client-driven engagement and asynchronous, text-based exchanges, which offer fewer immediate relational cues and may therefore require more effort to build trust and momentum. However, when the platform reliably supports ongoing task work (eg, clear structure, low friction, and a private space for disclosure) and counselors respond in a timely, personalized manner, the triadic connection can remain stable over time. Once established, the bond may still be vulnerable to emotional detachment in the absence of real-time interaction and nonverbal cues, making discontinuation more likely when motivation drops, responses are delayed, or platform friction accumulates. Related qualitative research suggests that therapists can take a more active role in supporting adherence in therapist-assisted iCBT, indicating that the balance of client and therapist contributions may vary across guided ICBT configurations. 40 Our findings also help explain why some clients experienced the relationship as more equal in guided ICBT. First, the platform provided a shared and transparent structure that organizes tasks and expectations, which reduced the extent to which progress depends on the counselor “directing” the session in real time. Second, asynchronous text-based communication gave clients greater control over timing, pace, and depth of disclosure, allowing them to reflect, edit, and respond when ready. This sense of agency could make the interaction feel less evaluative and less hierarchical. Third, because much of the therapeutic work was carried out through clients’ self-learning and written reflection, the counselor’s role is often experienced as supportive and collaborative, focused on clarifying, affirming, and helping clients apply skills rather than positioning the counselor as an authoritative expert. Taken together, an equal relationship in guided ICBT may reflect a redistribution of control from therapist-led interaction to client-led task work supported by the platform and counselor feedback.
Taken together, this discussion contributes a refined conceptualization of DTA in guided ICBT. By disentangling its core constructs, identifying its structural characteristics, and analyzing the mechanisms of interaction, we offer a framework that both builds on and extends existing alliance theory in digital mental health contexts.
Implications
Our findings suggest several actionable implications for practice, digital platforms, and future research in the context of guided, asynchronous ICBT. First, because participants described DTA as difficult to establish but easy to terminate, services may benefit from lowering the “start-up” cost of alliance and building in opportunities for repair. Platform design can support this by making trust and confidentiality cues explicit, reducing friction in key workflows (eg, log-in, navigation, homework submission), and using clear prompts that help clients initiate task work when motivation fluctuates. Second, counselor responsiveness remained important in sustaining alliance through text-based communication. Timely, complete, and empathic feedback, together with consistent encouragement, may compensate for the limits of asynchronous interaction and help clients stay engaged. Brief, well-timed phone check-ins may facilitate DTA development by strengthening emotional bonding. Third, organizational routines such as supervision and peer support can function as scaffolding for maintaining the quality and consistency of counselor communication, especially when clients disengage or misunderstand session tasks. Future research should test and refine this framework across other guided ICBT contexts and populations, and examine whether specific platform features and implementation strategies shift alliance trajectories and early termination over time.
Transferability
Although this framework was developed in guided, asynchronous ICBT for young sexual minority men in China, several elements may transfer to other guided ICBT contexts, including the triadic client–counselor–platform structure and platform-mediated alliance processes. Other elements may be more context-specific, such as stigma-related privacy concerns and program-specific features. Transferability may also vary across digital modalities (eg, asynchronous text-based guidance vs synchronous teletherapy vs fully automated apps), which warrants future studies to test and refine the framework across settings and delivery formats.
Limitations
There are several limitations of this study. First, the framework of DTA is developed based on our ICBT project, and the participants are young sexual minority men, which may not accurately reflect the perspectives of all demographic groups. Subsequent studies should consider examining this framework among diverse client populations and different age groups to validate the framework. Second, the study is conducted in the context of guided ICBT with asynchronous counselors’ support, which may not generalize to other digital therapeutic settings given the asynchronous nature of digital communication. 41 Further research is needed to explore DTA across different digital platforms to investigate how specific platform features, such as response latency, message personalization, or interface design affect alliance quality. Third, we included all implementers in the program (5 counselors and 1 supervisor). Given the small size of the service team, additional provider perspectives in other settings may reveal further nuances beyond what was captured here. Fourth, the findings are primarily based on self-reported data, which may involve biases such as social desirability or memory recall issues.
Conclusion
In this guided asynchronous ICBT context, DTA was described as being co-constructed within a client–counselor–platform triad. DTA is further characterized by a client-driven structure, equal counselor-client dynamics, and an interactional process that is often hard to initiate but easy to disengage from. These features offer both opportunities for autonomy and flexibility, and challenges for emotional engagement and continuity. By analyzing the dynamic interactions among clients, counselors, and digital platforms, this study proposes a multifaceted conceptual framework for DTA in guided ICBT, contributing to future research and the development of scalable, effective digital mental health interventions.
Supplemental Material
sj-docx-1-inq-10.1177_00469580261436993 – Supplemental material for A Conceptual Framework of Digital Therapeutic Alliance in Internet-Based Cognitive-Behavioral Therapy: A Qualitative Study
Supplemental material, sj-docx-1-inq-10.1177_00469580261436993 for A Conceptual Framework of Digital Therapeutic Alliance in Internet-Based Cognitive-Behavioral Therapy: A Qualitative Study by Mengyao Yi, Ashley Hagaman, Lloyd Goldsamt, Shufang Sun, Mengshu Li and Xianhong Li in INQUIRY: The Journal of Health Care Organization, Provision, and Financing
Supplemental Material
sj-docx-2-inq-10.1177_00469580261436993 – Supplemental material for A Conceptual Framework of Digital Therapeutic Alliance in Internet-Based Cognitive-Behavioral Therapy: A Qualitative Study
Supplemental material, sj-docx-2-inq-10.1177_00469580261436993 for A Conceptual Framework of Digital Therapeutic Alliance in Internet-Based Cognitive-Behavioral Therapy: A Qualitative Study by Mengyao Yi, Ashley Hagaman, Lloyd Goldsamt, Shufang Sun, Mengshu Li and Xianhong Li in INQUIRY: The Journal of Health Care Organization, Provision, and Financing
Footnotes
Acknowledgements
The authors would like to acknowledge the clients and counselors who participated in this study for their contributions to the research.
Ethical Considerations
This study was approved by the Yale University Human Subjects Committee (No. 2000029433) and the Institutional Review Board of the Behavioral and Nursing Research of Central South University (No. E2021128).
Consent to Participate
Electronic informed consent was obtained from all participants prior to the interviews, covering the study aims, procedures, potential benefits and risks, confidentiality, and the voluntary nature of participation.
Consent for Publication
Not applicable, as all data were de-identified prior to analysis and no individual personal information is included in this manuscript.
Author Contributions
Mengyao Yi contributed the conceptualization, formal analysis, investigation, data curation, project administration, and visualization, and drafted the original manuscript. Ashley Hagaman and Xianhong Li contributed to conceptualization and methodology. Xianhong Li additionally contributed to investigation, resources, visualization, supervision, and funding acquisition. Lloyd Goldsamt and Mengshu Li contributed to formal analysis, and Shufang Sun and Ashley Hagaman provided supervision. All authors participated in reviewing and editing the manuscript, approved the final version, and agree to be accountable for the integrity of the work.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by the National Institutes of Health (NIH, R21TW011762). Mengyao Yi, Ashley, and Shufang received support from the NIH. Xianhong Li received support from the NIH, the China Medical Board (CMB; grant 22-465), and the Science and Technology Department of Hainan Province (grant ZDYF2024SHFZ042). Mengshu Li received support from the China Medical Board (CMB; grant 22-465).
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
Due to the sensitive nature of the qualitative data involving vulnerable populations, the data supporting the findings of this study are not publicly available. Access may be granted upon reasonable request and with appropriate ethical approvals.
Supplemental Material
Supplemental material for this article is available online.