Positive public attitudes could be an indicator of acceptance and adoption of internet interventions. Generally, attitudes can be characterized as an aggregate of subjective assessments about an object, ranging, for example, from harmful to helpful [36].

The theory of planned behavior [37] proposes that attitudes, among other factors, shape individuals’ intentions, which then lead to a certain behavior. Individuals’ personal expectancies are assumed to shape such attitudes and to thereby influence behavioral intentions [37]. In accordance with the unified theory of acceptance and use of technology [38], performance expectancy (ie, how useful an individual perceives an intervention to be for reaching a specific goal) might thus play an important role in the adoption and acceptance of internet interventions [39,40] and provide a guideline in overcoming the limitations in the acceptability of eMHSs [34,39,41,42].

Technology acceptance of eMHS can be operationalized by intentions to use these services [39], which can be affected by the individual preference for a specific delivery mode [28]. Treatment preference means to choose a treatment in favor of an alternative option. Research evidence suggests that considering patients’ preferences for a psychological treatment is associated with improved clinical outcomes [43]. However, little is known about preferences for specific delivery modes, such as therapist-guided treatment, unguided internet interventions, and VCP, and their impact on the willingness to use eMHSs. Some studies identified a preference for traditional (face-to-face) over internet-based treatment (eg, [28,29,31,44-47]) and for therapist-guided over unguided eMHSs (eg, [28,29,33]).

Potential determinants of attitudes and preferences for eMHS include sociodemographics such as age, region [45], or professional background [48-50]. Regarding health-related and help-seeking variables, using the internet for mental health information [45,51], previous use of eMHSs [31], a history of mental illness and help-seeking experience such as undergoing psychotherapy [29], knowledge about eMHSs or awareness of electronic therapies (“e-awareness”) [33,41,42], personality traits [28], and perceived stress [42,52] have been reported as predictors. Regarding the role of symptom severity, a recent study [12] illustrated a help-seeking behavior paradox in students, where individuals’ readiness to seek help from face-to-face services declined with increased perceived stress. In contrast, the same study also demonstrated a positive association between distress and seeking help online.

Attachment style may be a further predictor of eMHS uptake [52], since attachment theory [53] has been applied to predict intentions to use face-to-face help services (eg, [54]). Based on early infant-caregiver interactions, relatively stable internal working models of the self and others in terms of mental representations of close relationships are built. These implicit expectations regarding self-efficacy and reliance on significant others in stressful situations are manifested in adulthood [53]. Adult attachment style can play a role in preferences and attitudes toward seeking help in the context of emotionally relevant relationships, such as in mental health care [52,54]. While a secure attachment style (low attachment anxiety and avoidance; ie, positive models of the self and others) is associated with functional coping strategies, insecure attachment styles were identified as a global vulnerability factor for mental health [55,56] and are related to altered stress responses, symptom reporting, and less use of health care resources [57,58]. However, the role of attachment styles in the readiness to use eMHSs remains unclear [52], especially concerning different delivery modes of internet interventions that vary in the degree of human support.

Taken together, the identification of determinants of attitudes toward and preferences for eMHSs is at an early stage [34]. This study addressed this research gap.

We used a modified 17-item version of an e-therapy attitudes measure (ETAM) [42] containing statements about typically cited benefits of internet therapy and its comparability with face-to-face psychotherapy, as well as subjective beliefs (eg, about data security). Participants were asked to rate their agreement with each statement on a 5-point rating scale ranging from 0 (“strongly disagree”) to 4 (“strongly agree”). To ensure comparability, participants were instructed to rate items regarding guided internet interventions (see Textbox 1). Based on previous exploratory factor analysis, we identified two factors, which we termed perceived usefulness and helpfulness, and advantage relative to face-to-face therapy. Multimedia Appendix 1 provides detailed information about the exploratory factor analysis. For classification of attitudes, we used predefined cutoffs in line with previous work using the ETAM [41,42]: mean scores <1.5 (a median score of 0 or 1) were defined as negative, values between 1.5 and 2.49 (median score of 2) as neutral, and scores ≥2.5 (median scores of 3 or 4) as positive attitudes toward guided internet interventions. Cronbach alpha was excellent in this survey (alpha=.92).

We operationalized preference (see Textbox 1) by assessing help-seeking intentions for different delivery modes of internet interventions (e-preference, options 1-3) in contrast to the disinclination to use internet interventions in case of emotional problems (non–e-preference, option 4).

Attitudes and age were significantly and positively correlated (ρ₍₆₄₃₎=.079, P=.045), with older participants displaying more favorable attitudes than younger participants toward internet-based guided self-help. Unemployed participants (mean 2.30, SD 0.69) showed more positive attitudes than employees in training (mean 1.37, SD 0.37, M_diff 0.93, SE 0.288, 95% CI 0.003-1.85). We found no significant differences in attitudes for sex, marital status, region, native language, education level, or work in the health care or social sector (all P>.05).

Frequency of seeking health information online was associated with differences in internet intervention attitudes (F_4,641=6.675; P<.001, η_p²=.040), with more positive attitudes reported by individuals who sought information weekly (mean 2.01, SD 0.61, M_diff –0.36, SE 0.112, 95% CI –0.67 to –0.04), several times a month (mean 2.11, SD 0.78, M_diff –0.45, SE 0.097, 95% CI –0.72 to –0.18), or rarely (mean 1.91, SD=0.59, M_diff –0.26, SE 0.080, 95% CI –0.48 to –0.04; all P<.05) than by those who never did (mean 1.66, SD 0.59). This was not the case for participants who reported seeking information daily (mean 1.46, SD 0.99); both groups (never, daily) expressed rather negative attitudes. There was a significant positive correlation between attitudes toward guided internet interventions and perceived stress on the PSQ-20 (ρ₍₆₄₃₎=.092, P=.020). No significant differences in attitudes were identified for any of the other help-seeking–related variables (eg, e-awareness, attachment style, all P>.05).

We found no significant differences in e-preferences (preference for guided or unguided internet interventions and VCP) based on age, sex, marital status, region, native language, education level, employment status, or work in the health care or social sector (all P>.05).

E-awareness significantly predicted a preference for different forms of internet-based therapy (χ²₆=12.8; P=.046). Individuals who were aware of internet therapies (97/214, 45.5%) or not sure (40/87, 46.0%) were more likely to prefer guided internet interventions than were those who were not aware (115/343, 33.5%).

We found differences in e-preference based on experience with online counseling (χ²₃=13.8; P=.003); persons with experience were less likely to prefer unguided interventions (7/68, 10.6%) than were those without (117/578, 20.2%).

Experience with psychotherapy also predicted e-preference (χ²₉=21.6; P=.01). A preference for guided internet interventions was most common among persons without experience who were currently seeking a therapist (20/35, 57.1%) and persons with experience with psychotherapy (99/228, 43.4%). All subgroups were nonetheless most likely to prefer guided internet interventions.

This study identified an overall moderate public acceptance, or moderately positive attitudes, toward guided internet interventions in a German sample. This tendency is in line with another study on a psychoeducational intervention using an adapted ETAM version [41]. Participants supported health care insurance coverage of costs for guided internet-based therapies and endorsed the helpfulness of such approaches, their perceived anonymity, and the chance to receive help earlier compared with traditional health care. At the same time, participants disagreed with the supposed comparability of guided internet interventions with face-to-face psychotherapy, for example, with regard to their effectiveness and possibility to develop a good therapeutic relationship. Our findings are, furthermore, consistent with earlier research with respect to a general preference for face-to-face therapies over internet interventions [28,29,44], data security concerns [64,65], and perceived higher compatibility of internet interventions with everyday life [11,29].

We found no relevant differences for sociodemographics as predictors of attitudes toward internet-based guided self-help. Interestingly, neither education level nor sex was associated with attitudes. These results are in line with a study on the acceptance of internet-based interventions in chronic pain [66]. Replication of this finding might indicate that the often-reported overrepresentation of woman and highly educated participants in randomized controlled trials evaluating internet interventions [13,67-72] might not be due to lower acceptance of digital health interventions in general, but due to other relevant barriers such as lower willingness to seek help. Future research should try to shed light on low utilization rates among persons with low education and men.

Moreover, participants with higher levels of perceived stress in the past month tended to express a more positive attitude toward internet interventions, which is consistent with a prior study using the same instrument to assess attitudes [42]. This might point to improved acceptance of such guided internet interventions among participants in a stressful situation with an actual need for support.

This study identified a clear preference for guided over unguided internet interventions, which only few studies have investigated before [34]. Interestingly, guided internet interventions were also preferred over VCP. Approximately four-fifths of the participants were willing to use internet-based approaches for emotional problems, indicating a broad applicability of internet interventions for mental health care.

High e-awareness was associated with a preference for guided internet interventions. Overall, e-awareness in our sample was low (33.1%), which could be, for instance, seen in context of the early stage of implementation of eMHSs in Germany [50,73] and might rise further in the future. This is supported by previous German surveys reporting even lower rates of e-awareness (14.0%-27.3%; [42,52,64]), including a representative socioeconomic panel (SOEP-Innovation Modules 2016, N=4802) showing 24.4% e-awareness (D Richter, written communication, May 2017). Experience with seeking psychological help formats was also a determinant of preferences, which is consistent with other studies [29,31]. Results also suggest that attachment avoidance was associated with a higher preference for guided and unguided self-help via internet interventions, and very low preference for VCP. This finding contributes to research on links between attachment styles and face-to-face health care use readiness [52,57,58] and might indicate that internet-based (guided) self-help approaches could help to reach individuals for whom attitudinal and other psychological barriers such as attachment avoidance might be a drawback for use of an intervention [74].

Furthermore, participants with higher levels of perceived stress showed a higher preference for internet-based guided self-help than for VCP. This might indicate that individuals with stressful lives have problems adhering to fixed synchronous therapy sessions, and that providing asynchronous treatments might help them to get access to psychological treatments, which they would otherwise not use. Such an assumption is supported by studies that found high proportions of first-time help seekers in internet-based stress management programs [75-79]. However, future research is needed to confirm such an assumption.

First, results indicate that, although internet-based approaches are not an option for some individuals, a large proportion of participants in this study were potentially inclined to use eMHSs for treatment. However, e-preference rates were lowest for VCP, which, as a synchronous delivery format, is the most similar to conventional face-to-face psychotherapy [27]. Results also indicated that individuals with high attachment avoidance were least inclined to use this synchronous format to seek help, but were more willing to use asynchronous internet-based interventions. Perceived stigma and a preference for managing mental health problems on one’s own are known barriers to seeking synchronous treatment [10], and personal characteristics such as attachment style may contribute [57,58]. This suggests that provision of asynchronous treatment options, such as guided or unguided internet interventions, could be a feasible way to reach larger proportions of the general population, especially individuals who would not use synchronous options such as face-to-face psychotherapy or VCP. Matched-care models have been proposed before (cf, [80]), allocating internet-based or face-to-face treatment based on symptom severity; findings in this study, however, pointed out that various asynchronous as well as synchronous treatment formats should be provided simultaneously to reach as many individuals affected by mental health problems as possible.

Second, these results also suggested that offering guidance alongside internet-based self-help internet interventions in routine care could, from a public health perspective, have a major influence on their effects on a population level. Whether to offer guided or unguided interventions in routine care has been debated in the literature since internet-based self-help has emerged. This discussion has since predominately focused on potential differences in adherence, effects, and costs [22,81-83]. Meta-analytic findings clearly indicate that stand-alone guided self-help interventions can be effective in the prevention and treatment of a range of mental health problems, including depression [68], anxiety [72], and stress [77]. However, although more patients could potentially be treated for the same costs using unguided self-help, a basic prerequisite to exploiting the potential of any effective treatment is that affected individuals are willing to use it [84]. This study showed that approximately twice as many participants preferred guided interventions over unguided interventions. Thus, with evidence showing guided interventions to be comparable with face-to-face psychotherapy, for example for depression and anxiety [17,85], and with large effect sizes of guided formats when delivered under routine care conditions [25,86-88], preference should be given at the moment, whenever possible, to guided self-help in routine care. However, it should also be acknowledged that almost 20% of participants preferred unguided self-help; hence, future studies should clarify whether offering both guided and unguided interventions could lead to greater effects on a population level due to higher overall utilization rates, compared with offering only one of the two options.

Awareness about internet-based treatment was rather low in this sample, but was positively associated with higher preference for guided internet interventions. Furthermore, participants did not find internet interventions to be equal in effectiveness and therapeutic relationship to face-to-face therapies. Previous research, however, has shown that the effects of guided internet interventions are comparable with face-to-face therapies [17,85] and that therapeutic relationships are of the same quality as in conventional treatment [1-4]. This finding points to the importance of developing measures to increase awareness of and knowledge about the efficacy of internet-based treatment in the public to raise its acceptance. Acceptance-facilitating interventions using brief, highly scalable educational videos have been shown to be a valid strategy to enhance the acceptability of internet interventions in clinical practice [66,71,84]. As acceptance-facilitating interventions may be easily disseminated through official health care information channels, they might be an auspicious approach to increase e-awareness and knowledge concerning internet interventions, and thus raise their public acceptance.