This is an exploratory secondary analysis based on data derived from a cross-sectional survey-study that was carried out by a research team of the University of Zurich in summer 2020. For the primary analysis the acceptance and perceived barriers of EMH were calculated as an average of five different EMH services (psychotherapy via telephone, psychotherapy via videoconference, VR, unguided and guided programs) among PiT. The current acceptance scores of EMH services were compared to pre-COVID-19 acceptance scores, which were assessed retrospectively. Results will be reported elsewhere in full length¹. Participants were recruited between June and July of 2020 via recognized educational institutions for psychotherapy within Germany and the German-speaking part of Switzerland. Recruitment was administered solely via e-mail, asking the post-gradual educational institutions to forward the link to the survey to PiT. Thereby, PiT were directed to the survey, which was conducted online and completely anonymous. The survey contained 50 questions and mean processing time was 19.1 min (SD = 5.9). As an incentive, participants could take part in a raffle of book vouchers worth 50 euros. Institutions were contacted again if they did not answer the request after 2 weeks. In total, 29 institutions in Switzerland and 232 institutions in Germany were contacted. Since only a few institutions gave feedback on forwarding the questionnaire, no statement can be made about the response rate on an institutional level. In total, the questionnaire was opened 692 times, with 228 PiT completing the survey which results in a dropout rate of 68.7 %. We could not control for multiple clicking, thus the dropout rate might appear higher than it actually is 0.12 participants were excluded from analyses as they had not started the practical part of their postgraduate clinical training yet. After written consultation with the President of the Ethics Committee of the University of Zurich on 3 March 2020 and the checklist to self assess ethical safety, no further approval of the ethics committee was necessary to garantuee the ethical safety of the study.

Data were analyzed using IBM SPSS Statistics 26 and R (Version 4.0.0). To answer the question whether the acceptance of EMH services varies between application purposes, we used as a statistical model a 2-factor within-subject (repeated measure) ANOVA with the factors EMH services (five steps: psychotherapy via telephone, psychotherapy via videoconference, VR treatment, unguided programs, guided programs) and application purposes (four steps: prevention, treatment substitute, treatment addition, aftercare) and EMH acceptance as dependent variable. The model included both main effects (EMH services and application purposes), as well as their interaction (EMH services x application purposes).

To identify how different barriers to the acceptance of EMH services might differentially affect EMH service types, we adopted a two-step approach. First, we identified an appropriate model of the relation of the barriers to the different EMH services in terms of general acceptance. Specifically, we considered three candidate linear mixed-effects models in our model set. All models included a main effect of EMH service type and a random subject intercept. The random subject intercept was included as acceptance was assessed multiple times, that is once per EMH service for each participant (i.e., as a repeated measure). This is a standard procedure to account for within-subject correlation of measures (e.g., see (74), p. 29). The first model (A1) additionally included a main effect of all five barriers each (data insecurity, impersonality, irresponsibility, legal concerns, concerns about therapeutic alliance), as well as pair-wise interaction terms of each barrier and EMH service type. Hence, this model represented a differential relationship of barriers to EMH service acceptance depending on the type of service. The second model (A2) dropped the interaction terms, hence representing a uniform influence of the barriers on EMH acceptance. The third model (A3) dropped the main effect terms of the five barriers, representing no influence of the barriers on EMH acceptance. Our criterion of model comparison was based on Akaike Information criterion (AIC) weights (75), which express the probability that a model is the best in the model set conditional on the data. Second, we inspected the regression coefficients of the best fitting model specifically for the five barriers to gain insights on which barriers had a significant influence on EMH acceptance.

We followed an equivalent procedure to better understand the influence of advantages of EMH services. Again, we firstly identified an appropriate descriptive model, considering three candidate linear mixed-effects models in our model set. All models included a main effect of EMH service type and a random subject intercept. The first model (B1) additionally included a main effect of all four advantages each (simplified information provision, time flexibility, geographic flexibility, simplified contact maintenance), as well as pair-wise interaction terms of each advantage and EMH service type. Hence, this model represented that the relationship of advantages to EMH service acceptance depended on the type of service. The second model (B2) dropped the interaction terms, hence representing a uniform influence of the advantages on EMH acceptance. The third model (B3) dropped the main effect terms of the four advantages, representing no influence of the advantages on EMH acceptance. Again, we inspected the regression coefficients of our best fitting model specifically for the four advantages, to gain insights on which of them had a significant influence on EMH acceptance.

Lastly, we aimed to put the different pieces of our data modelling together within the UTAUT framework. Specifically, we wanted to test if adding possible influences of barriers and advantages (depending on the analyses above) presented a meaningful extension to the classic UTAUT predictors and simple comparison model featuring only demographic predictors (age, gender). All models included a main effect of EMH service type, age, gender, and a random subject intercept. In addition, model C1 included the UTAUT predictors (performance expectancy, effort expectancy, social influence, and facilitating conditions), the barriers and advantages, as well as knowledge about, experience with and subjective assessment of the scientific evidence base of different EMH services as they have been shown to have an influence on EMH acceptance. Model C2 only additionally included the UTAUT predictors, while model C3 did not include additional predictors. Again, our criterion of model comparison was based on Akaike Information criterion (AIC) weights.

Figure 1 provides a summary of key sociodemographic characteristics. The sample size was n = 216 participants, with n = 60 participants who trained in Switzerland (27.8%) and n = 156 in Germany (72.2%). Most of them were female (88.4%) and between 25 and 39 years old (85.2%). N = 197 respondents studied psychology (91.2%) and n = 6 medicine (2.8%) before starting with their clinical training to become a psychotherapist and n = 13 indicated completing other degrees (6%). Regarding the theoretical orientation, 67.1% stated that they are trained in behavioral therapy (cognitive/cognitive-behavioral), 16.2% in depth psychology or psychoanalysis, 12.5% in systemic therapy, and 4.2% in humanistic therapy. N = 33 participants named various or different integrative approaches (15.3%).

Based on prior research (33, 70) the mean score of EMH acceptance was categorized as low (1–2.34), moderate (2.35–3.67), or high (3.68–5). In general, results revealed that acceptance of EMH was moderate (M = 3.4, SD = 1.12), while acceptance of psychotherapy via videoconference was highest (M = 3.7, SD = 1.15) and acceptance of unguided programs was lowest (M = 2.55, SD = 1.14). Table 1 gives an overview.

Among respondents, general perceived personal knowledge about EMH was moderate (M = 3.64, SD = 0.86), while psychotherapy via videoconference was most well-known (M = 4.34, SD = 0.72). Practical experience with EMH was generally low, as participants stated using EMH services in only one out of ten therapeutic cases (M = 10.37, SD = 10, range 0–100%) between the onset of the COVID-19 pandemic and the time of data collection (June-July 2021). However, there were considerable differences between EMH services and high variance scores within psychotherapy via telephone and videoconference. Psychotherapy via videoconference (M = 26.55, SD = 28.80) and via telephone (M = 23.05, SD = 25.07) was used in about one out of four therapeutic cases. Participants indicated serving only M = 1.34% (SD = 2.2) of their patients with VR. Lastly, PiT recommended unguided EMH programs to only M = 3.38% (SD = 10.50) of their patients, while they stated that they have accompanied M = 4.19% (SD = 12.95) of their patients with guided programs.

Figure 2 provides an overview of the key results. Mauchly tests for sphericity revealed relevant violations (all p < 0.001) wherefore we report Greenhouse-Geisser corrected statistics. Our results confirmed the expected heterogeneity in the acceptance of different types of EMH services depending on their intended application purpose. Specifically, we found an interaction effect of EMH service and application purpose (F(6.229, 1283.088) = 111.497, p < 0.001, η² =0.21). Post-hoc tests showed that, on average, over all application purposes, psychotherapy via videoconference was the most accepted EMH service (all p_bonferroni < 0.001). Further, EMH services were comparatively less accepted as a treatment substitute in acute care than for other application purposes (all p_bonferroni < 0.001). Interestingly, unguided and guided programs were specifically well accepted in preventive care (more so than all other services, all p_bonferroni < 0.059). VR was comparatively less accepted across all application purposes (all p_bonferroni < 0.001).

Beyond comparative statements, we used one-sample, one-sided Wilcoxon signed-rank tests against test value of 3 (neutral) to test which EMH services for which application purposes were seen as a valuable addition to the therapy catalogue on absolute scale. This was the case in 11 of 20 combinations. Specifically, results show that EMH services, except VR, are seen as useful for prevention and aftercare whereas they are not accepted as treatment substitution. Table 2 summarizes the results.

While this study contributes to the understanding of the acceptance of different EMH services for various application purposes and its determinants, it also has some limitations that should be considered. First, we must consider the time point of assessment. Data were gathered during the first months of the outbreak of the COVID-19 pandemic which could explain higher acceptance rates compared to older studies (32, 33, 45). The given circumstances have accelerated the use of remote services and forced psychotherapists to rethink about digital alternatives to treat patients. Additionally, the online survey included a description of structural benefits of psychotherapy via telephone or videoconference, especially in extraordinary conditions such as the COVID-19 crisis, which could have positively influenced acceptance scores for these two EMH services. At the same time, general acceptance rates could also be lower compared to newer studies as experience with EMH was still relatively low among respondents and EMH experience has been shown to be positively related to technology acceptance (28, 35, 88). Even if we consider the early stage of implementation of EMH services in Germany and Switzerland (38, 89, 90), healthcare experts have gained experience with digital medicine during the COVID-19 pandemic, the intention to use EMH services might increase concurrently.

In addition, the gender ratio was not balanced as more female than male psychotherapists in clinical training participated in our study which might have influenced our results. Moreover, the response rate was rather low, as on average less than one respondent per institution completed the survey. Age and gender were no predictors of acceptance in the advanced UTAUT model, which is likely due to the selection bias with few male participants and little variation in age. Female psychotherapists in some European countries like Germany have been shown to be more likely to endorse and provide digital psychotherapy during the first weeks of the COVID-19 outbreak in Europe, especially by those who were more concerned about an infection with COVID-19 (60). However, in our study we did not control for nontherapeutic reasons for providing digital psychotherapy, such as concerns regarding an infection.

Furthermore, the present study only focused on acceptance and fell short in the question of how behavioral intention and actual use behavior might be linked. Even though UTAUT describes behavioral intention as a direct predictor of the actual uptake (28), potential users do not always follow their intentions (“intention-behavior gap”, (91)). Thus, we agree with Philippi et al. (9) that future research should focus on the relationship between the intention to use different EMH services and use behavior (92) and investigate whether identified predictors of EMH acceptance could potentially influence actual uptake rates.

Lastly, the operationalization of technology acceptance was slightly different to other studies focusing on acceptance toward digital interventions, thus comparability is limited. Even though we based our assessment of behavioral intention on the frequently used UTAUT, individual adaptations of the UTAUT questionnaire and the number of items can differ between studies. For instance, acceptance is sometimes operationalized with four items (32, 93) or two items (33) that are rated on a 5-point scale ranging from (1) does not apply at all to (5) applies completely. Apolinário-Hagen et al. (45) only used one item by assessing intentions to use EMH services with an abbreviated version of the procedure applied by Klein and Cook (94), asking participants how likely they would use 10 different conventional and EMH services in case of emotional problems on a 5-point rating scale ranging from (0) very unlikely to (4) very likely. In our study, we used three items to assess behavioral intentions, including two items that were also used by Hennemann et al. (33) and one item asking psychotherapists in clinical training for their intention to use different EMH services in their work ever (range 0–100) that was adapted from Elfeddali et al. (69) to measure intention strength.

To expand the uptake of EMH, there is a need to focus on increasing psychotherapists' acceptance of EMH services as they play a crucial part in patients' attitude formation and thus on the implementation of EMH services (35). Our results provide evidence of the need to focus on informing prospect psychotherapists about advantages of various EMH services when applied in different contexts such as prevention and aftercare, but also on how potential barriers such as data security or legal concerns could be overcome. Confirmatory, a study by Humer et al (61) revealed that several psychotherapists in Austria wished for more information on data protection and security. Even before the COVID-19 pandemic, lack of personal contact, data protection and security were already seen as most important disadvantages of online interventions to prevent common mental health disorders by stakeholders such as psychotherapists, policymakers and potential users in Germany, Switzerland, Austria and Spain (95). Thus, these aspects of EMH services should be addressed in training and further education of psychotherapists. Additionally, a clear regulatory framework is needed to reduce legal concerns of psychotherapists. Countries in an earlier stage of digital health implementation into healthcare, such as Switzerland or Germany, could learn from countries that are more advanced in the implementation of EMH services such as the Netherlands or the United Kingdom (27). As a starting point, van Daele et al. (96) have recently formulated an association with the European Federation of Psychologists' Associations (EFPA) general guidelines for mental health workers, health services, regulatory agencies as well as developers to promote the implementation of evidence-based EMH services. The strong need for training and further education also becomes visible in a recent study by De Witte et al. (43), in which participants were asked whether they received any form of training on online consultations about EMH. Results revealed that only 11% of the sample received a form of training, however, only half of these training programs were specific to EMH and lasted just <4 h in every second case. In accordance, Gerlinger et al. (38) indicate that mental health workers do not feel well informed about possible benefits and risks of EMH services, while only few have already gained practical experience with EMH services (97, 98).

To address misconceptions and knowledge gaps through information provision, AFIs have been found to be an established tool in educating individuals about novel treatment options such as EMH services and in increasing their acceptance (30, 32, 45, 99). For instance, Baumeister et al. (32) could show that an AFI such as receiving a short video of blended therapy can increase performance expectancy, effort expectancy, facilitating conditions and overall acceptance toward blended therapy. In the future, similar AFIs could be integrated into the curricula of postgraduate training programs and continuous professional education to increase knowledge about our identified drivers (i.e., simplified information provision, simplified contact maintenance, time flexibility, and geographic flexibility) as well as barriers (i.e., impersonality, legal concerns, and therapeutic alliance) to the acceptance of EMH services. By making EMH an integral part of the education, PiT could gain valuable experience in integrating EMH services into their therapeutic work with patients.

Furthermore, PiT with varying theoretical backgrounds might need different education. For instance, unguided EMH programs such as mental health apps are most often based on cognitive behavioral therapy, which could lead to the assumption that psychotherapists with a background in cognitive behavioral therapy might be more open to use such EMH services (40, 100). In line with this assumption, Baumeister et al. (32) pointed out that particularly psychodynamic oriented psychotherapists could profit from AFIs as they initially seem to be rather skeptical about unguided EMH programs. Furthermore, There are already several studies that have identified associations between theoretical orientation (e.g., psychodynamic, cognitive behavioral, and others) and attitudes toward the use of EMH services (40, 100, 101), however findings are comparatively inconsistent, thus to deduce practical implications future research in this area is needed.

This study is one of few to examine the acceptance of different EMH services (i.e., psychotherapy via videoconference, psychotherapy via telephone, VR, unguided and guided programs) across varying application purposes (i.e., prevention, treatment substitute, treatment addition, aftercare) from the perspective of PiT. We could show that acceptance for several EMH services differed for application purposes among PiT. The results showed that acceptance of EMH services was best predicted with an extension of the UTAUT model, including barriers (i.e., data insecurity, impersonality, liability, legal concerns, and concerns about therapeutic alliance), advantages (i.e., time flexibility, simplified information provision, geographic flexibility, and simplified contact maintenance), EMH experience, EMH knowledge, and EMH evidence assessment. As the use of EMH services will most probably increase in the next years because they offer quick and location-independent help for the prevention, treatment and aftercare of CMDs, our results highlight the need to distinctly inform PiT about different EMH services and their possible application areas. At the same time, our results provide support for stakeholders that are planning and designing training for PiT by highlighting factors that should be addressed if the goal is to increase EMH acceptance.