The study protocol was approved by the local Ethics Committee (EK 039/20). The study was designed and conducted in accordance with the latest version of the Declaration of Helsinki [10]. Moreover, the study was registered by clinicaltrials.gov on September 28, 2020 (NCT04579354).

All participants were patients of the Clinic for Anesthesiology, RWTH Aachen University Hospital. The inclusion criteria included age over 18 years, elective surgery with general anesthesia, and a good knowledge of German. Exclusion criteria included emergency surgery, thoracic surgery, neurosurgery, and tumor surgery, as these can be associated with increased anxiety or inability to provide consent. All participants signed an informed consent form before participating in the study.

Sample size planning was based on a meta-analysis by Carl [4] on the effect of virtual exposure on anxiety. For comparisons of VR exposure with participants on a waiting list that receive treatment after the treatment groups, this study reported a mean effect size (Hedges's g) of 0.9, analogous to Cohen's d. Sample size planning using g multiplied by power for this effect size resulted in a sample size of 68 patients (34 per group) for a two-sided null hypothesis test (t test for independent samples), with a power of 0.95 and an α error probability of .05. To compensate for an expected dropout of 10%-15%, we targeted a total sample size of 80 patients (40 per group).

The study employed a randomized parallel-group design to compare VORT to the standard operation preparation procedure (STOPP). We employed the state and trait anxiety scales from the state-trait operation anxiety inventory (STOA) [2] as outcome variables. STOA is a validated inventory with good psychometric properties which assess operation-related trait and state anxiety on separate scales. Trait anxiety (STOA-T, 20 items) is the relatively stable disposition of a person to be anxious about surgeries, whereas state anxiety (STOA-S, 10 items) assesses acute fear reactions in situations shortly before or after surgery on 2 dimensions (cognitive anxiety and affective anxiety, 5 items each). Measuring a relatively stable trait, STOA-T was rated only before the surgery, whereas STOA-S was assessed before (T1) and after the surgery (T2). As an additional outcome, we employed an in-house designed questionnaire to assess the participants’ perception and evaluation of VORT after surgery. The questionnaire assessed different statements about VORT with 8 items on a 5-point Likert scale from 1 (do not agree) to 5 (strongly agree). Moreover, in 1 item, patients were asked to give VORT a school grade (according to the German grade system) between 1 (excellent) and 6 (very poor).

Participants were randomly assigned to one of the study arms (VORT vs STOPP) by a person not involved in data acquisition using the following procedure: for all patients (N=80), we created opaque assignment envelopes containing a folded paper with a study arm assignment (VORT or STOPP), with a total of 40 participants in each arm. The order of the envelopes was randomized before the beginning of the study. The assignment envelope for each patient was opened after these steps to avoid any bias in providing study information to the patient or the standardized explanatory interview. Thus, the assignment to the study arms was double-blind (to the study nurse or anesthetist and the patient) until after the standardized explanatory anesthesia interview. The procedural steps of the study can be described as follows:

Moreover, concerning the application of VORT, we were interested in its overall patient evaluation. Also, we wanted to explore the possibility of perceived usefulness, especially in patients with higher trait anxiety, in other words, if the VORT rating was associated with perioperative trait anxiety. Finally, we wanted to evaluate whether the evaluation of VORT depended on the patient’s age (eg, older patients’ limited experience with VR). Therefore, we additionally investigated the following explorative research questions:

Data were analyzed with SPSS Statistics (version 25; IBM). STOA-S sum scores for the two dimensions (cognitive and affective anxiety) were compared between study arms with two-tailed independent samples t tests separately for T1 (H1) and T2 (H2). To quantify agreement to the Likert scale items in the evaluation questionnaire (ER1), we calculated the median of the responses. Furthermore, we recoded the responses to disagreement (values 1 and 2), neutral (value 3), and agreement (values 4 and 5) and descriptively quantified the percentage of agreement. To evaluate a possible relationship between trait operation anxiety and the evaluation of VORT (ER2), we calculated Pearson correlations between the items from the evaluation questionnaire and STOA-T values. Finally, we calculated correlations between the items from the evaluation questionnaire and the patients’ age. A threshold of an α error probability of .05 (two-sided testing) was applied to all analyses.

In total, 84 patients were included in the study. Dropouts (n=12; 14.3%) were in the expected range; however, since all dropouts occurred during the enrolment of the first 62 patients, we decided at this time point to create and randomize 20 additional assignment envelopes. The primary reason for dropouts was the early discharge of the patient. Other reasons included postponed or additional surgeries and postoperative complications. VORT-associated concerns or worries were not reported; therefore, VORT did not contribute to participation dropouts. One patient withdrew from study participation. Data from the remaining 72 patients (mean age 54.19, SD 15.94, with a range of 20 to 81 years), with 35 female and 37 male subjects, was analyzed (VORT: n=35 vs STOPP: n=37). Due to single missing data points, the numbers of included patients vary slightly for the different analyses. Results of the study are presented according to our hypotheses.

No significant differences in state anxiety emerged between VORT and STOPP before surgery. This was the case for both the affective (n=71; VORT: mean 10.49, SD 4.61 vs STOPP: mean 10.83, SD 4.09; t₆₉=–0.37; P=.74) and the cognitive dimension (n=67; VORT: mean 10.88, SD 4.32 vs STOPP: mean 10.70, SD 4.38; t₆₅=–0.18; P=.86).

Similarly, no significant differences in state anxiety emerged between VORT and STOPP after the surgery. This was the case both for the affective (n=68; VORT: mean 8.53, SD 4.95 vs STOPP: mean 7.91, SD 3.26; t₆₆=0.61; P=.55) and the cognitive dimension (n=71; VORT: mean 9.11, SD 4.19 vs STOPP: mean 9.17, SD 4.00; t₆₉=–0.05; P=.96).

The median values of the responses and proportion agreement to the statements are listed in Table 1. Altogether, the patients’ ratings of the VORT are positive. The results indicate that most patients perceived the VORT as helpful for preparation and can recommend it to other patients. A median school grade of 2 (corresponding to “good” in the German grade system) corresponds well with these ratings.

There were no significant correlations between any of the items from the evaluation of the VORT (Table 1) and STOA-T values of the patients (all P>.19).

There were no significant correlations between any of the items from the evaluation of the VORT (Table 1) and the age of the patients (all P>.10).

Comparing VORT and STOPP with a two-tailed independent samples t test for all 7 items related to anesthesia treatment, the difference was only found in favor of VORT slightly below the predefined significance threshold (P=.047). Comparisons for all other items were not significant, resulting in no evidence that VORT addresses anesthesia-specific fear components.

Moreover, we calculated comparisons between men and women for all single VORT rating items in two ways to assure that men and women felt equally addressed by our video using a male protagonist. First, we calculated independent samples t tests for the Likert scale ratings. Second, we compared male and female participants using Χ² tests to assess possible sex-dependent differences in the agreement to the statements (recoded ratings). For the t tests, no significant sex differences emerged concerning the ratings of VORT (all P>.17); for the Χ² tests, we found one significant difference for item 6 (“the virtual surgery tour was not detailed enough”), with higher agreement among female participants (21.4%) compared to males (0%). Moreover, we compared male and female participants with respect to their overall rating of VORT (school grade) in an independent samples t test. Men and women did not differ in their school grades (P=.25), with women tending towards better grades (mean 1.86, SD 0.66) compared to men (mean 2.10, SD 0.54).

Although VORT did not affect perioperative anxiety, the tour received good patient ratings and was perceived as helpful for surgery preparation. Remarkably, the positive evaluation of the virtual surgery tour was independent of the fear of surgery (ie, the personality trait or trait anxiety). In other words, the subjective benefit of VORT for the patients could not be attributed to anxiety reduction. A possible explanation for this finding may be the beneficial effect of preoperative education and information. The assumption of uncertainty being a central element in the emergence of hospitalization stress was first raised by Mishel [12] and later specifically adapted to operation situations by Kagan and Bar-Tal [13], who reported that preoperative uncertainty negatively affected mental and physical recovery after surgery. This aligns with the findings of Moerman et al [14], showing that more than 80% of surgical patients had a positive attitude towards receiving information, with more than half of them having high information requirements. Receiving information about health care interventions positively affects several aspects of patients’ mental and physical health, such as recovery, pain levels, and psychological distress [15-17]. Thus, in our study, the positive ratings of VORT may be attributed to reducing uncertainty and satisfying the need for information. Similar to Moerman et al [14], we conclude that surgical fear and need for information are related but distinguishable factors and addressing either of them can reduce patients’ distress. For future studies, physiological measures such as heart rate, blood pressure, and skin conductivity may be relevant indicators for assessing the arousal of the patients. Including these physiological measures might help disentangle physiological processes and their cognitive appraisal [18]. For example, it seems conceivable that VORT evokes physiological arousal due to the operation-associated settings and situations but reduces subjective stress via cognitive control.

Surprisingly, VORT is perceived as beneficial irrespective of the patient’s age, indicating 2 interesting aspects. First, the need for information and the wish to reduce uncertainty seem independent of the patient’s age. Second, the acceptance of VORT was equally high among elder and younger patients. Therefore, there is no evidence that experience with VR may play any role in the evaluation or that elderly patients may feel uncomfortable with the VR setup. Given the high number of elderly patients undergoing surgeries with general anesthesia, these findings favor the general use of VORT in preoperative preparations.

Furthermore, there was no evidence that VORT addresses anesthesia-specific fear components. Our experience shows us that there are two patient groups: those who fear the anesthesia and its consequences in isolation and those who are only afraid of the procedure, the surgical part. As far as we know, no questionnaire provides information on fear of anesthesia in isolation.

There is mixed evidence concerning the potential of VR in reducing perioperative anxiety. Some studies show no reduction in anxiety [19-21] by interventions as in our study, and others offer a detectable influence [22-26]. However, anxiety reduction has been applied preoperatively with good results, especially in children [9,26,27]. These inconsistencies may be partly attributed to the use of different anxiety inventories. The Spielberger State and Trait Anxiety Inventory (STAI) is frequently used [28]. However, the suitability of the STAI has been questioned for operation situations since it primarily measures anxiety in assessment situations [2]. Accordingly, more specific inventories have been developed, such as the STOA [2], the Amsterdam Perioperative Anxiety and Information Scale [14], and the Yale Preoperative Anxiety Scale [29], which was designed especially for children. The content of the items, length of the questionnaire, and other aspects vary considerably between these inventories, influencing study outcomes. For instance, Krohne et al [2] demonstrated that trait scores of STAI and STOA were correlated at r=.40, indicating that both inventories reflected a common factor of general anxiety but measured distinguishable facets of anxiety. Therefore, when comparing studies on VR interventions for perioperative anxiety, one must consider which inventories were used.

One possible limitation of our study is that all patients received the VR tour after the “educating interview.” Regarding patient compliance, it would be relevant to know whether there are fewer canceled surgeries when patients receive a VR education beforehand. Moreover, with the changed order, the VR tour could prepare patients for the anesthesia interview, enabling the patient to ask more detailed and specific questions. The second important limitation concerns the limited generalizability of our results to other types of anesthesia (eg, regional procedures). This is relevant in terms of distinguishing separate aspects of preoperative anxiety. Specifically, we assume that a distinction between fear of anesthesia and fear of the operation procedure by itself can be much more precise since several aspects of general anesthesia are not provided here (eg, fear of not waking up, lack of trust, or losing control while being unconscious).

Another limitation concerns whether the patients already had previous surgery with general anesthesia. Our study did not evaluate this aspect, and we cannot exclude the possibility it played a role in the VORT assessment. However, after careful consideration, we conclude that it is rather unlikely that prior surgical experience with anesthesia affected the VORT assessment. The probability of undergoing surgery increases with life age, suggesting that there was most likely a substantial correlation between life age and operations in the past in our study sample as well [30]. Our data found no correlation between life age and VORT ratings. We are aware the latter cannot distinguish between people who have experienced an operation and those who have not; however, we conclude that there is at least no indication from our data that having undergone an operation or not plays a role in the VORT rating.

VORT is beneficial for a patient’s preparation for surgery with general anesthesia. Remarkably, this effect did not result in reduced perioperative anxiety. Instead, VORT seems to serve the need for information and to reduce uncertainty. Moreover, it is both accepted and recommended by patients. Considering the manageable organizational and financial effort and minimal time required in its implementation and the amount of time spent waiting in the perioperative outpatient clinic, the general use of VORT can be recommended.