To address our research objectives, we employed a qualitative interview study design to focus on the lived experiences and perspectives of each respondent [44]. This study method is an effective strategy for developing theories and exploring experiences of social situations and reasons for behavior [45].

Initially, ethical approval was obtained from the Ethics Committee of the Medical Faculty of the University of Duisburg-Essen (23‐11642-BO). The study was conducted after obtaining informed consent from all participants. Privacy of the participants was safeguarded by pseudonymizing each interview. The audio data were stored on a password-protected device and will be kept for the next 10 years, ensuring adequate confidentiality. Participation was voluntary, and the study participants did not receive any compensation.

A total of 20 interviews were conducted between March and May 2024, involving 10 students from each of the fields of interest: medicine and psychology. The inclusion criteria were fluency in the German language, adult age (>18 y), and therapeutic-related study subject (human medicine or psychology). Participants were recruited in different ways, including flyers distributed at the university campus, in the clinic setting, on social media, and through direct communication, both verbally and through email, within the study team’s environment (professional and personal networks). Every interviewee completed the whole interview.

The majority of qualitative interview formats are semistructured interviews conducted with individuals or groups [46]. A semistructured interview comprises a guide of questions but allows for an individual order of questions, wording, and in-depth inquiry [45]. After reviewing the current relevant literature, we developed an interview guide adapted to our target group. The semistructured questionnaire is provided in Table S3 in Multimedia Appendix 1. Initial pilot testing ensured familiarity with the equipment and the interview process.

The interview was divided into 4 thematic sections. It started with questions about individual sociodemographic data and general and health-related media use. After that, we explored participants’ perspectives regarding our DIPT concept. The fourth section addressed experiences and expectations regarding digitalization in medical and psychotherapeutic contexts, including the possibility of AI-guided psychotherapy. Finally, we concluded by offering participants the opportunity to ask questions about the study or provide feedback. The interview guide, which provides a detailed definition of the DIPT concept, is provided in Table S3 in Multimedia Appendix 1.

After a detailed description of the study was provided verbally and in written form, written informed consent was obtained and documented with signature. Interviews were conducted in person at the facilities of the LVR-University Hospital Essen, Clinic for Psychosomatic Medicine and Psychotherapy, Essen, Germany. Audio recordings were obtained with a voice recorder (Sony-ICD-PX470). We selected a sample of 20 interviews, with each subgroup consisting of 10 interviews, because recent studies indicate a saturation of themes with no more aspects emerging at this point [47,48]. Saturation of comprehension requires approximately 20 interviews [48], which was also achieved in this study when subgroups were analyzed separately. The interview guide was successfully applied. Therefore, all interviews yielded 20 comparable and complete datasets. Thereafter, audio data were transformed into written transcripts using f4x 2024 software [49]. Finally, the data underwent a proofreading process and were carefully collated with the audio recordings to ensure high data quality.

The average duration of an interview was approximately 29 minutes. The conducted interviews were transformed into pseudonymized transcripts. Data were analyzed semantically using MAXQDA 24 software, a tool that can be used for analyzing qualitative data [50,51]. Three independent researchers analyzed the data: 2 of them coding all transcripts, and the third coder coding 25% of the transcripts. The third researcher also acted as a supervisor and mediated in cases of disagreements, as described by Kuckartz [52].

We conducted a thematic analysis following the instructions established and revised by Braun and Clarke [53,54]. This approach recognizes the active role of the researchers in identifying and interpreting patterns of meaning across the data. The 6-phase framework consists of (1) familiarizing oneself with the data, (2) initial coding, (3) searching for themes, (4) reviewing themes, (5) determining themes, and (6) reporting the results. Themes were generated inductively through an iterative and reflective process, which was data-driven without a preestablished, potentially restrictive codebook [53]. The subjectivity and positionality of the research team (consisting of PhD candidates, research assistants, postdoctoral researchers, and a full-time professor) are considered integral to the analytic process. Thematic analysis is described as a 3-step approach of descriptive coding, interpretive coding, and the final definition of relevant themes [55]. After every step of this analysis, we held a meeting of the 3 coders to ensure the validity of the results [56]. The underlying methodological perspective was a coding reliability approach within reflexive thematic analysis, one of the several subtypes that have evolved from thematic analysis [57]. In this approach, multiple coders work independently and subsequently compare their analysis. The degree of consensus reached enables the reliability of the results [57,58]. Initial coding involved creating short summaries from the raw data (eg, “presence therapy is preferred”). Subsequently, the researchers clustered these codes. For example, summaries related to hardware and software were grouped under the theme “technical infrastructure,” while all mentions of AI were sorted to the theme section “AI-Psychotherapy.” In the following team meeting, these themes were merged and organized into a hierarchical category: theme-subtheme system (axial coding). At this stage, 50% (10/20) of the interviews had been coded. With the developed codebook, the 2 independent researchers then coded the remaining half of the interviews (final coding). In the second meeting, the team discussed and finalized the codebook. Final adjustments were made with the assistance of the supervisor or third coder. To underline sufficient data-reporting quality, we applied the checklist for COREQ (Consolidated Criteria for Reporting Qualitative Research) by Tong et al [59]. The checklist is provided in Checklist 1.

Participants’ ages ranged from 20 years to 54 years (mean 25.95, SD 6.91 years). Of the 20 participants, 16 (80%) were identified as female, while 4 (20%) were identified as male. Half of the interviewees (10/20, 50%) reported approximately 3 to 6 hours of digital media use per day. Within the context of their study program, 70% (14/20) had some psychotherapeutic experience and 45% (9/20) reported personal experiences in psychotherapy as patients. Most frequently used personal digital devices were smartphones (20/20, 100%) and laptops (14/20, 70%), followed by tablets (12/20, 60%). Digital media was mentioned for various purposes, primarily for participation in social media (17/20, 85%) and entertainment (16/20, 80%). The use of digital health applications ranged from daily (7/20, 35%) to none (6/20, 30%). In this context, digital support was predominantly used for health tracking purposes (12/20, 60%) or in connection with health insurance applications (10/20, 50%). Further information is presented in Table 1.

Furthermore, we obtained data regarding the individual use of digital devices in general and for health-related purposes. The data are provided in Table S2 in Multimedia Appendix 2. The results indicate a marked usage behavior and high digital literacy.

The coding process resulted in the emergence of 4 overarching categories: (1) evolution of digitalization in medical practice, (2) future directions for digital psychotherapy, (3) technical framework, and (4) AI-based psychotherapy. A comprehensive overview of the finalized codebook is provided in Figure 2. For transparency and to support clear thematic interpretation, all participant quotations referenced in the results are included in Table S4 in Multimedia Appendix 3.

This study explored medical and psychology students’ perspectives on digitalization in medicine and psychotherapy, with a particular focus on our DIPT concept. Participants provided concrete suggestions for supporting digitalization, especially through the expansion of existing structures like teleconsultations, digital appointment scheduling, and digital files. Compared to the existing literature, our findings reflect a more nuanced perspective, containing both enthusiasm and reservation. In recent studies, students’ attitudes appear less focused, claiming a general positive [60,61], while others express an overall skeptical attitude [62]. Positive experiences with telemedicine consultations, as reported by our participants, align with previous literature about patients and providers [63,64]. In a similar setting of integrative medicine during the COVID-19 pandemic, Barth et al [65] reported a particularly good working alliance between the therapists and the patients, which raises hope for the psychotherapeutic sector. Nevertheless, the experience of technical problems was reported repeatedly [63,65], consistent with the concerns expressed in our interviews.

The most extensive thematic focus of this study pertained to students’ opinions toward digital psychotherapy in general and in the context of a possible inpatient setting. We noticed a clear support for the general idea of digital psychotherapy but at the same time a marked hesitation toward a fully digital psychotherapy setting. This rather conservative result was surprising, given the participants’ digital literacy and their frequent use of digital media in general and in health-related contexts. In certain situations—for example, everyday constraints, work-related restrictions, emergency contacts, physical limitations—psychotherapy appears to be accepted; however, fully digital settings seem to be less well-received. Given the assumption that our sample represents a particularly open group with regard to digital treatments, attitudes in the general population may be even more reserved. A closer examination of our study population reveals that our sample consisted of young, predominantly female participants currently enrolled in academic education. This homogeneous study population is consistent with the significantly higher proportion of women in the field of medicine and psychology. The proportion of female students is approximately 70% in medicine [66] and even higher in psychology, at 76.5% [66]. Mackenzie et al [67] reported an association between female gender and a more help-seeking attitude toward mental health treatment, which could consequently lead to a higher response rate of female participants in our study.

Nevertheless, this reported ambivalence toward digital psychotherapy is echoed in current research. In several studies, students show an open attitude toward digital mental health interventions [41,68-70]. Regarding digital psychotherapy, Braun et al [71] reported positive experiences of digital psychotherapy in combination with digital mental health applications. Studies examining students’ opinions of digital psychotherapy remain relatively scarce overall. However, a survey from Gbollie et al [72] revealed that only 12.6% of the students preferred digital solutions over a traditional face-to-face therapy. In another survey that accompanied a group therapy setting, students’ opinions on therapeutic effects were divided [73]. The specific requirements for digital psychotherapy voiced by our participants are consistent with the literature. Building a therapeutic relationship in person prior to digital sessions seems to be favorable [29,74]. Caution regarding changes beyond therapy location seems warranted. Leukhardt et al [74] found that change of location causes a feeling of uncertainty in the psychotherapeutic process. The idea of specific patient suitability seems valid considering studies on factors that influence engagement in digital mental health interventions: older age [75], severe mental health issues [76,77], or even the lack of knowledge of the existence of this type of care seem to be barriers [78]. Conversely, social affiliation [77] or digital literacy [79] facilitate engagement. The stated chances of digital therapy were in line with prior studies. Digital psychotherapy options are valued for their low threshold, greater flexibility, and reduced travel [80]. Nevertheless, important concerns regarding impaired interpersonal and interactional aspects remain significant. Feasibility studies confirm these concerns: patients experienced digital psychotherapy as less personal [29], with more superficial interaction [74]. Frittgen and Haltaufderheide [81] highlight the changes in sensory and embodied aspects of communication. Given both important critical and beneficial aspects, recommendations for a hybrid therapeutic setting, recommended by several of the participants, emerge as the next logical step. The corresponding data seems promising [82], and some even appear better than traditional therapy [83].

Students also identified the technical framework underlying digital care as a central issue. Studies about digital interventions emphasize the importance of easy usability [69,71], likewise a user-centered design [84]. These features are also mentioned by the participants in the present study. Important problems highlighted by the current findings are lack of technical competences or issues with the digital hardware and software infrastructure. These arguments cannot be dismissed and have repeatedly been reported as a relevant implementation hurdle for digital interventions or digital psychotherapy [76,77,83,85]. Another aspect in this theme was awareness regarding the handling of sensitive information, which is reflected in previous research works [69,75,79,84].

Considering the latest developments in digital care, AI was also addressed. In our study, it became clear that AI-guided psychotherapy was largely rejected as a standalone psychotherapeutic agent, whereas AI-guided assistance was widely supported. In recent literature, the use of AI in psychotherapeutic settings appears promising in terms of effectiveness [86-88] and may offer low-threshold access [89]. Critical evidence highlights a lack of emotional depth [90], which aligns with our participants’ skepticism. Kolding et al [87] state that recent studies about AI need to be interpreted with caution because they often lack quality and have limited clinical relevance. At the European level, the AI Act is a regulation that came into force on August 1, 2024 [91], and is intended to protect recipients of AI health services from physical and psychological harm by preventing the approval of applications with potential adverse effects [91]. The Ethics Committee of the Bundesärztekammer has issued a clear statement on this topic: the responsibility for diagnosis, indication, and therapy lies with the physician [92], and only certified treatments should be used [92]. Similarly, the Kassenärztliche Bundesvereinigung emphasizes that AI should be seen as a support tool, not as a replacement for medical services [93].

Despite the insights gained, we still have limitations. The qualitative nature of this interview study can only give results that reflect the outlook and lived experience of the selected study participants. It enables us to understand the investigated themes and develop theories, but the study findings are not generalizable [45]. Nevertheless, we gained valuable knowledge of important factors for mental health treatment and how future mental health professionals want to experience digital psychotherapy, grounded in their life experiences [46]. Although the study group as a whole demonstrated digital literacy and an open-minded attitude toward potential digital psychotherapy treatments, it is noteworthy that only 3 out of 20 (15%) interviewees had received information about digital psychotherapeutic care during their studies, all of whom were psychology students. Given that digital (mental) health treatments are increasingly integrated into routine health care, it is crucial that their application, benefits, and limitations be systematically incorporated into academic curricula at the beginning of professional training. Likewise, Wibberly [94] as well as Kröplin et al [95] argue that the integration of these topics into study curricula is necessary to address the future demands placed on health care providers. In a few cases, study participants were unfamiliar with the concept of a day clinic, and we noticed that this influenced their ability to envision the DIPT concept. Consequently, discussions often shifted toward the aspects of digital outpatient psychotherapy. This context appeared to be more familiar and easier for participants to envision, likely due to their experiences with digital teaching during the COVID-19 pandemic, personal exposure to digital psychotherapy lessons or apps, and contact with digital therapy options addressed during their studies. Furthermore, it is important to keep in mind that researchers, even when following a paradigm of asking open-ended questions without influencing responses, are not free from their own theoretical background and epistemological perspective on the investigated topic [53]. The researchers’ different professional backgrounds allowed for a balanced perspective on both medical and therapeutic topics. First coder and interviewer, a medical doctor (26 years old), was trained in an open-ended interview approach. Students were not directly dependent, neither professionally through work as student assistants, nor in the context of academic performance assessments. Second and third coders were student assistants (study subject was psychology), well trained in qualitative research. Overall, we assume that a pressure-free interview atmosphere was maintained. A logical next step would be a feasibility study of a real-life implementation of DIPT. With the insights gained here, a more successful concept can be developed, expected hurdles can be evaded, and identified opportunities can be realized.

The results of this interview study led to the conclusion that students are open to more digitalization in health care. Concerning digital psychotherapy, support is more conditional. Hybrid settings were favored that incorporate establishing a therapeutic alliance in person and/or continuing it through in-person meetings during the therapeutic process. A fully AI-guided psychotherapy was rejected, whereas therapeutic supportive tools were welcomed. Technical problems appear to be a relevant hurdle for committing to a digital setting. The obtained aspects can help in implementing adequate digital therapeutic care in a digital inpatient-like or outpatient setting.