Virtual Reality-Based Alcohol Prevention: The Results of a Cross-Sectional Study with Visitors of an Art Exhibition

Abstract

Virtual reality (VR) can help to prevent alcohol consumption by simulating peer pressure. While existing VR applications for adolescents are primarily tested in schools, expanding them to museums and art galleries offers setting-specific benefits and could enhance their reach. An interactive VR simulation was presented in an art exhibition on substance use reflection. This study aimed to (1) describe visitor experiences with the simulation and (2) analyse differences according to gender and age. Visitors (n = 293) explored Virtual LimitLab for about 15 min before completing a survey on user experience and perceived learning experiences, which also included open questions on its positive and negative aspects. The quantitative user experience and perceived learning experience items were rated positively, except for knowledge increase. Young visitors (≤18 years of age) perceived Virtual LimitLab as more interesting (p = 0.043) and exciting (p = 0.047) than older ones. Female visitors rated the reflective response (p = 0.016) and sensitisation to social pressure (p = 0.024) higher than males. The qualitative evaluation showed that visitors appreciated the simulation’s realism, authenticity and adaptability. VR-based public health interventions in the arts setting need further research to study their preventive potential.

1. Introduction

In Germany, alcohol consumption is high, with around 7.9 million people aged 18 to 64 drinking in risky quantities [1]. It is common among German adolescents, with 65% reporting having drunk alcohol in the past 30 days (average of 35 European countries: 47%) and a higher 30-day prevalence of binge drinking than the European average (54% vs. 34%) [2].

In alcohol prevention, virtual reality (VR) technology can provide immersive, interactive and controlled environments in which resistance skills can be practised in a variety of simulated social (pressure) situations that mimic real-life scenarios. For instance, the Australian project Blurred Minds VR House Party, a VR simulation designed for adolescents to experience peer pressure and the consequences of drinking alcohol in a party environment, found high user satisfaction and positive changes in knowledge about alcohol and attitudes towards binge drinking [3]. VR FestLab from Denmark [4] and Virtual LimitLab from Germany [5] are other applications for which positive user experiences and game satisfaction have been reported, suggesting the potential to engage adolescents in alcohol prevention through the use of an interactive and playful VR simulation.

However, the available VR applications for alcohol prevention were primarily tested in schools (e.g., [3,4,6]). The specifics of this setting, such as its formal atmosphere with school rules or the presence of classmates and teachers, could influence one’s interaction with the VR application (e.g., concerns about stigma, socially desirable rather than realistic actions), suggesting the need for extension to other health promotion settings. From a health promotion perspective, it is thus necessary to test the VR simulation in different settings to extend its reach, to increase its adaptability and scalability by testing its acceptability and effectiveness in different contexts, and to benefit from setting-specific advantages.

One area that is underdeveloped in terms of public health interventions compared to common settings (e.g., schools or workplaces) is the cultural heritage sector. Museums and art exhibitions offer non-stigmatising environments, because the visit is not accompanied by shame or criticism [7] but rather by leisure, education and cultural enrichment, which could make individuals more receptive to stigmatised and sensitive issues such as substance-related addiction or problematic consumption patterns. In addition, museums are often accessible at little to no cost, with discounts for certain groups of people (e.g., students), making them a promising venue for public health interventions [7]. In fact, there is supporting evidence for the use of museums and art galleries for public health purposes [7,8]. For instance, museum object handling was found to increase well-being in people with early- to mid-stage dementia [9]. Moreover, a museum-based nutrition education intervention was found to have a positive effect on knowledge and behaviour change intentions in students aged 8 to 18 years [10].

VR opens up a new dimension of viewing and experiencing art. Applications include VR experiences for museum visitors such as exploring ancient oceans [11] and VR tours of museums and cultural sites from anywhere (e.g., with the Google Arts & Culture app and a cardboard VR viewer) [12]. VR, with its advantages of interactivity and emotional engagement, can also be used by individual artists to draw attention to certain issues. For instance, in Marina Abramović’s “Rising”, viewers confront climate change effects by interacting with the artist in a VR scenario of rising sea levels. They face a virtual Abramović in a flooding glass tank and can choose to save her by committing to environmental action, which symbolically lowers the water level. Whilst museums and art exhibitions have been suggested as sites for public health interventions [7,8] and the arts sector can benefit from VR, to our knowledge, none of the VR applications for alcohol prevention have yet been tested in such a setting.

To take up new opportunities and combine addiction and prevention research with the artistic examination of substance use and nightlife, the exhibition “On a Night Trip—Between a High and a Crash?” took place from the end of October 2023 to the beginning of December 2023 at the Defensionskaserne Petersberg in Erfurt, Germany [13]. The exhibition featured 40 objects by up-and-coming and internationally renowned artists. One of these objects was the virtual application Virtual LimitLab, which is dedicated to alcohol prevention among adolescents aged 15–18 years. The application included a simulated party experience designed to enable adolescents to say ‘no’ to alcohol offers from peers in peer pressure situations that encourage alcohol consumption and was developed for delivery in schools. The exhibition was visited by a total of 3400 persons during the seven-week opening period.

Based on this background, the objective of this study was to evaluate the feasibility of Virtual LimitLab as the only currently available German VR application for alcohol prevention in a new setting, namely an art exhibition. Specifically, the aims were to (1) describe user experiences with the simulation, its perceived learning experiences, and positive feedback and suggestions for qualitative and quantitative improvement, and (2) to analyse differences according to gender and age.

2. Materials and Methods

2.1. Study Design

This cross-sectional study consisted of testing the VR simulation Virtual LimitLab for alcohol prevention in an art exhibition and collecting data on user experiences with the application and the assessment of the perceived learning experiences.

2.2. Virtual LimitLab: A VR Application for Alcohol Prevention

Virtual LimitLab is an app developed for Android devices that was installed specifically for the exhibition on seven Oculus Quest devices (Meta Platforms Inc., Menlo Park, CA, USA). The VR application is an overdubbed version of the Danish simulation VR FestLab originally developed at the University of Southern Denmark [14] and represents a 360° filmed virtual house party simulation in which the user can enter a variety of scenes (e.g., the bar or the dance floor). Screenshots illustrating Virtual LimitLab are presented in Figure 1. Users can navigate through the application using gaze control and are regularly prompted to make different decisions (e.g., offer or refuse help to a drunk person, accept or reject flirting options, accept or reject alcohol offers). The user’s decisions and blood alcohol concentration, which reflects their alcohol consumption during the game play, determine the subsequent individual plot. During the simulation, the user encounters both positive (refuse to consume alcohol) and negative role models (encourage user’s alcohol use) and can have either positive (flirting, positive feedback in text messages the next morning) or negative experiences (blacking out, negative feedback) depending on the alcohol consumption. The aim of Virtual LimitLab is to use the simulation of a realistic party environment for adolescents to recognize peer pressure situations, improve their drinking refusal self-efficacy by developing the corresponding action and explore communication strategies to make decisions against the consumption of alcohol and for an alternative action.

2.3. Data Collection and Recruitment

Virtual LimitLab was offered to be tested by visitors of the exhibition “On a Night Trip—Between a High and a Crash?” in Erfurt, Germany, over a seven-week period between October and December 2023. The art exhibition served as a platform for exploring the intersection between addiction, happiness and social behaviour in the context of nightlife. It involved the public viewing of 40 different exhibits, including works by emerging and well-known artists as well as contributions from scientific research. Central to the exhibition was the exploration of euphoria and the potential for substance use (e.g., alcohol, cocaine, cannabis) in the nightlife scene with the cultural elements of music, dancing and flirting. The exhibition had around 3400 visitors, of whom one-third were school students [13]. In addition to self-guided exploration of the exhibition during regular opening hours, various guided tours were offered. It was possible to book tours for adolescents or school classes and tours led by addiction self-help groups. Guided tours for teachers and prevention workers were also offered outside opening hours. The visitors took part via self-selection without any specific age restriction. The research team was not involved in the recruitment of participants. Although Virtual LimitLab was developed for adolescents between the ages of 15 and 18, this selection criterion was not applied for gaining access to the application in the exhibition. School classes and their teachers were invited to visit the exhibition and were able to register specifically for Virtual LimitLab and experience the VR application in a separate time slot. Exhibition mediators provided technical and content-related support for Virtual LimitLab. The playing time of each round was limited to around 15 min. The duration of the simulation varied depending on the decisions made in the simulation. After the visitors had completed a run of Virtual LimitLab, they were invited to participate in the evaluation of the application. After providing informed consent, data from the visitors were subsequently collected using a questionnaire which was provided on tablets. Participation in data collection was voluntary and could be terminated at any time.

2.4. Measures

A three-page short questionnaire was developed for the evaluation of Virtual LimitLab. The intention was to keep the questionnaire as short as possible to reach visitors at a low threshold. The survey platform Lime Survey was used to design the questionnaire.

In the first part of the survey, the short version of the user experience questionnaire (UEQ-S) by Schrepp et al. [15] was used. Although the UEQ-S questionnaire asks users about both pragmatic and hedonic quality, only the four questions on hedonic quality were used. Participants were asked on a seven-point bipolar scale whether they found Virtual LimitLab exciting/boring, interesting/not interesting, inventive/conventional and leading edge/usual. The respective evaluation categories were marked on the bipolar scale by the adjectives used to evaluate the simulation.

The second part of the survey dealt with the evaluation of the content and perceived learning experiences of Virtual LimitLab. Informed by the behaviour change wheel taxonomy by Michie et al. [16], participants were asked to rate eight self-formulated statements on Virtual LimitLab about the relevance of peer pressure, decision-making skills, knowledge, reflection skills and the perceived consequences of excessive alcohol consumption, the realism of the characters and the design of the scenes (see statement formulations in the Section 3). The statements could be rated on a four-point Likert-type scale from “strongly agree” to “strongly disagree”.

For the qualitative evaluation of the Virtual LimitLab experience, two open-ended questions were provided to the respondents to obtain information about positive aspects and suggestions for improvement. The first question was as follows: (“I liked this about the game …”). The second question was as follows: (“I would change this about the game …”). Finally, two questions were asked to assess gender in two categories (male and female) and age in eight age groups (<13; 13–5; 16–18; 19–29; 30–39; 40–49; 50–59; ≥60).

2.5. Data Analysis

SPSS statistics 27 (IBM Corp., Armonk, NY, USA) was used for data analysis. For the UEQ-S items, the mean values were calculated. Following the classification by Schrepp [17], values above 0.8 on the seven-point bipolar scale from −3 (most negative) to +3 (most positive) represent a positive evaluation, values between −0.8 and 0.8 represent a neutral evaluation, and values below −0.8 a negative evaluation. Absolute and relative frequencies were calculated both for the UEQ-S items and for the items of assessment of the content. To analyse the agreement with statements about the application, the categories “strongly agree” and “agree” as well as “strongly disagree” and “disagree” were collapsed (agree/disagree).

For bivariate analysis, the variables gender (female and male) and age (18 and younger, 19 and older) were dichotomised. The Mann–Whitney U-test was calculated for the combinations between gender and age with UEQ-S mean values and the chi-square test for the two socio-demographic variables and items of assessment of the content.

Content analysis was used to systematically group responses regarding the positive aspects of Virtual LimitLab and suggested improvements into themes. Categories were inductively derived from the responses. The responses were then assigned to these categories.

3. Results

3.1. Participants

Out of a total of 3400 visitors to the exhibition, 293 people tested and evaluated Virtual LimitLab. The number of visitors who tested Virtual LimitLab but did not participate in the evaluation was not recorded. More women than men took part in the evaluation. A total of 57% (n = 165) of respondents stated that they identified with the female gender and 42% (n = 122) identified with the male gender (

Table 1. Characteristics of study participants.

Gender	n = 291 (100%)
female	165 (56.7%)
male	122 (41.9%)
not specified	4 (1.4%)
Age	n = 289 (100%)
≤18 years	57 (19.7%)
19–39 years	162 (56.1%)
≥40 years	70 (24.2%)

). Different age groups were represented, from “18 years and younger” to “40 years and older”. More than half of the respondents were in the age group “19–39 years” (56%, n = 162).

3.2. Hedonic Quality of Virtual LimitLab

All items of the UEQ-S were evaluated positively, with 1.4 being the lowest mean value (Figure 2). The best rated item was the evaluation of the simulation as “leading edge”, with a value of 1.9. The overall mean of all four UEQ-S items was 1.67.

3.3. Assessment of Virtual LimitLab

Most participants (89.9%, n = 258) agreed that they were able to influence the course of the simulation through contact with characters in Virtual LimitLab and rated the characters as realistic (87.6%, n = 248) (Figure 3). Moreover, most of the respondents (86.9%, n = 253) agreed that the influence of peer pressure on drinking behaviour became clear to them while playing (Figure 3). Furthermore, most participants (79.7%, n = 228) found the scenes in the simulation interesting and agreed with the statement that the simulation illustrated the negative consequences of excessive alcohol consumption (64.9%, n= 187). The majority (60.5%, n = 173) also confirmed that the content of the application made them reflect about their own behaviour or that of others at parties. More than half of the respondents (58.3%, n = 168) also agreed that playing the simulation made them better able to decide how to react when another person offers them alcohol.

However, knowledge acquisition was rated low, and most participants (71.0%, n = 203) did not agree that they were able to expand their previous knowledge about the effects of alcohol while playing.

3.4. Hedonic Quality by Gender and Age

The mean values did not differ significantly by gender (female/male) for any UEQ-S hedonic quality item. The Mann–Whitney U-tests for differences in hedonic quality by age group showed that participants aged 18 years or younger rated Virtual LimitLab as significantly more exciting (U = 5126.50, Z = −1.98, p = 0.047) and interesting (U = 5061.50, Z = −2.02, p = 0.043) than users aged 19 and older. The mean values were positive for both the younger and older age group for each item, with 1.29 being the lowest mean value (

Table 2. Bivariate analyses of hedonic quality by gender and age group.

UEQ-S Item	Female M ± SD	Male M ± SD	M-U-Test	Z-Value	p-Value	≤18 Years M ± SD	≥19 Years M ± SD	M-U-Test	Z-Value	p-Value
exciting/boring	1.41 (SD = 1.41)	1.39 (SD = 1.36)	9169.00	−0.30	0.765	1.76 (SD = 1.12)	1.29 (SD = 1.44)	5126.50	−1.98	0.047
interesting/not interesting	1.71 (SD = 1.37)	1.73 (SD = 1.29)	9293.00	−0.11	0.916	2.00 (SD = 1.26)	1.63 (SD = 1.37)	5061.50	−2.02	0.043
inventive/conventional	1.76 (SD = 1.36)	1.57 (SD = 1.35)	8177.50	−1.41	0.158	1.42 (SD = 1.24)	1.72 (SD = 1.39)	4830.00	−1.95	0.051
leading edge/usual	1.90 (SD = 1.30)	1.92 (SD = 1.20)	8957.00	−0.05	0.957	1.98 (SD = 1.08)	1.85 (SD = 1.34)	5884.00	−0.28	0.781

M = mean; SD = standard deviation; M-U-test = Mann–Whitney U-test; p-values in bold are significant (p < 0.05).

).

3.5. Agreement with Statements About Virtual LimitLab by Gender and Age

Chi-square analyses between gender and agreement with the statements about Virtual LimitLab found female participants to be more likely to agree with the statements that the game demonstrates the influence of peer pressure on drinking behaviour (χ² (1) = 5.11, p = 0.024), encourages self-reflection on their own and others’ behaviour at parties (χ² (1) = 5.80, p = 0.016) and improves their perceived ability to make better decisions when offered alcohol (χ² (1) = 5.60, p = 0.018).

Chi-square analyses between age and agreement with the statements about Virtual LimitLab found a significant difference between the age groups for one item (

Table 3. Bivariate analyses of the assessment of Virtual LimitLab regarding content and perceived learning experiences by gender and age group.

Item	Gender			Age
	Agreement in Percent	Chi-Square (df)	p-Value	Agreement in Percent	Chi-Square (df)	p-Value
I was able to influence the course of the game through contact with the characters in the game.	91.5% (female) 89.1% (male)	0.46 (1)	0.500	87.7% (≤18 years) 90.8% (≥19 years)	0.48 (1)	0.486
The characters in the prevention game were realistic.	89.5% (female) 85.6% (male)	0.98 (1)	0.322	83.9% (≤18 years) 88.5% (≥19 years)	0.86 (1)	0.353
It became clear to me during the game that drinking behaviour can be influenced by peer pressure.	91.5% (female) 82.6% (male)	5.11 (1)	0.024	78.9% (≤18 years) 89.2% (≥19 years)	4.27 (1)	0.039
The game scenes were interesting.	76.7% (female) 85.7% (male)	3.57 (1)	0.059	84.2% (≤18 years) 78.4% (≥19 years)	0.94 (1)	0.332
The game showed me what negative consequences excessive alcohol consumption can have.	64.6% (female) 66.7% (male)	0.13 (1)	0.722	63.2% (≤18 years) 65.1% (≥19 years)	0.07 (1)	0.787
The content of the game made me think about my own behaviour or that of others at parties.	66.7% (female) 52.5% (male)	5.80 (1)	0.016	54.4% (≤18 years) 61.7% (≥19 years)	1.01 (1)	0.315
By playing the game, I can make better decisions about how I react when other people offer me alcohol.	64.0% (female) 50.0% (male)	5.60 (1)	0.018	54.4% (≤18 years) 59.0% (≥19 years)	0.39 (1)	0.532
I was able to expand my previous knowledge about the effects of alcohol while playing.	29.9% (female) 27.1% (male)	0.25 (1)	0.613	34.5% (≤18 years) 27.5% (≥19 years)	1.07 (1)	0.301

df = degrees of freedom; the p-values in bold are significant (p < 0.05).

). Participants aged 19 years or older were significantly more likely to agree that it became clear to them during the simulation that drinking behaviour can be influenced by peer pressure than the group of participants aged 18 or younger (χ² (1) = 4.27, p = 0.039).

3.6. Positive Qualitative Aspects of Virtual LimitLab

Participants were able to specify in free-text fields what they liked about Virtual LimitLab and what they would change about it. A total of 100 of the 195 free-text responses were categorised as positive feedback. Of these, 89 responses were grouped into categories with at least three mentions (Figure 4), while the remaining suggestions were only mentioned once or twice. Thirty-three responses fell into the category of “realism and authenticity of the party experience”. The respondents rated the application as realistic, especially the party setting, the social pressure and the characters (e.g., “Many situations were very realistic and reflected the course of a house party very well.”). The options for action (what a user can do in the simulation, e.g., help a drunk person) and the ability to influence the course of the simulation were the second most frequently mentioned positive aspects (n = 22) (e.g., “You were free to decide what you could and couldn’t do and had many options to choose from.”). The novelty and the 3D experience were also positively emphasized (n = 10). In addition, the idea of the application was mentioned among the positive feedback (n = 8) and the experience in general was described as “interesting” by five respondents. Five responses could be assigned to the “humour and entertainment” category, which included the music in the party scene and the bartender’s jokes, in addition to the general assessment of the simulation as “funny”. Other positive feedback included the presentation of the consequences of alcohol consumption, that self-reflection was well possible (n = 3 for both) and the suitability of Virtual LimitLab for schools (n = 2).

3.7. Suggestions for Improvement in Virtual LimitLab

A total of 95 responses were collected to the question of possible improvements, 80 of which could be assigned to categories with three or more mentions (Figure 5). The respondents most frequently wished for more options for action (n = 18) (e.g., “I would be happy if one would have even more options to make the game even more realistic and to have more options for own actions.”). The answers in this category were predominantly formulated in general terms (“more options for action”), although specific points such as more non-alcoholic drinks or making up an appearance were also mentioned. Participants also frequently (n = 16) wished for a clear simulation of the influence of alcohol, with some people noting that they had not recognized any effects of alcohol consumption (e.g., “I would change the end scene so that if you’ve had too much to drink, the consequences are shown more clearly.”). Specifically mentioned were the simulation of a drunken state, a change in visual perception (blurred vision), and dizziness and (going to the toilet) vomiting before blacking out due to alcohol. A clear reaction from the other characters to the user’s alcohol consumption was also mentioned as a point for improvement. Among twelve responses, “improving the quality of dialog and interaction” was also a frequently represented category. Responses focused primarily on the realism of character interactions, which should be increased through more realistic eye contact, more options in conversations and interactions in different locations (e.g., toilet conversations). In addition, the respondents saw a need for technical optimization, particularly in terms of image quality (four out of nine responses in this category). Other requests for improvement included a multiplayer function (n = 6), additional narcotics such as cannabis or psychedelics as options (n = 6) and a longer simulation duration (n = 5).

4. Discussion

The aim of this study was to investigate the feasibility of Virtual LimitLab, a VR application for alcohol prevention in adolescents, in an art exhibition as a new setting. Quantitative and qualitative data were collected to (1) describe visitor experiences with the simulation and (2) analyse differences according to gender and age.

Regarding the first aim of this study, the evaluation of Virtual LimitLab was generally positive among visitors of the art exhibition, indicating the acceptance of VR-based alcohol prevention outside the school setting. We found a positive hedonic quality that was generally better than in previous studies. While the UEQ-S mean values in our sample ranged from 1.4 to 1.9, they were between 0.8 and 1.7 in a sample of 31 boarding school students who tested the Danish version of the application, VR FestLab [18]. In a later study of the hedonic quality of VR FestLab in a larger sample (n = 181) of adolescent users (aged 15–18 years), the mean values were even lower at 0.5 to 0.9 [4]. We have also previously tested the hedonic quality of Virtual LimitLab with the UEQ-S, including 13 adolescents aged 15 to 18 years. Here, in line with the other tests, the evaluation of hedonic quality was lower than in the present study, with mean values of 0.7 to 1.8 [5]. Findings on how the assessment of the four hedonic quality items differs are inconsistent. For instance, while “leading edge” received the highest rating in this study, the mean value of this item was lowest among the four items in two of the previous studies [5,18].

The statements that the participant had the opportunity to influence the course of the simulation through contact with the characters and that the characters were realistic received the highest levels of agreement in this study of nearly 90% of respondents. In agreement with this, the most positive free-text responses were in the categories “realism and authenticity of the party experience” and “options for action and ability to influence the course of the game”, with both statements being part of these broader categories. Although realism was not evaluated in the same way across the studies, the positive assessment of realism is nevertheless mostly consistent with findings for VR FestLab [18], Blurred Minds VR House Party [3] and also for Virtual LimitLab [5]. The idea of the application and that the experience was interesting and novel were other frequently mentioned free-text responses that were consistent with the positive rating of the UEQ-S hedonic quality items (e.g., “interesting”, “inventive”).

Most participants also agreed with the statement that the simulation had made them more able to decide how to react when other people offer them alcohol and that the content of Virtual LimitLab had made them reflect about their own behaviour or that of others at parties. They further agreed that the application had shown them the negative consequences of excessive alcohol consumption and that it had become clear during the simulation that drinking behaviour can be influenced by peer pressure. When interpreting these findings, it needs to be considered that we did not assess the effect of the simulation on the intention to drink and whether users are able to refuse drinking alcohol in situations where they are under peer pressure. For Blurred Minds VR House Party, it was found that positive attitudes towards binge drinking were reduced and behavioural intention remained low among Australian adolescents after experiencing the simulation [3]. The researchers of VR FestLab investigated the influence of the simulation on drinking refusal self-efficacy in Danish adolescents aged 15 to 18 and reported a small, but non-significant positive effect [6]. These studies however were conducted in the school context, allowing for additional educational influence through teachers, which is limited or at least largely unstructured in an art exhibition context. While it is difficult to accurately measure a change in real-life alcohol consumption associated with the VR simulation experience, it is nevertheless necessary to go beyond the user experience and assessments of perceived learning experiences surveyed in this study and examine attitudes and behavioural intentions towards alcohol use among German adolescents using Virtual LimitLab while also using a control group exposed to an application without preventive intention.

Most users (71.0%) did not report an explicit increase in knowledge about the effects of alcohol while playing. As the simulation is experience-based and not knowledge-creating in nature, a knowledge increase was not to be expected in the first place. A study on the Blurred Minds VR House Party simulation found a significant increase in knowledge about alcohol among adolescents [3], but comparison with our study is difficult due to differences in study design.

The qualitative user feedback revealed that it was not only the variety of options for action in Virtual LimitLab that was an important positive aspect of the simulation, but the provision of further options was also the most frequently mentioned suggestion for improvement. In agreement with this, previous qualitative findings showed that the selection options were positively assessed, but that some users would like more variety [5], and further, suggestions for more options in the other two VR applications were made [3,18]. We also identified the clear simulation of the influence of alcohol and its consequences as a frequently mentioned point for improvement. This is consistent with previous studies on Virtual LimitLab [3,5], where specific suggestions such as blurred vision were made [5]. In the context of health behaviour change models such as the Health Belief Model, a more dramatic portrayal of alcohol consequences could influence the perceived threat of alcohol consequences and provide a cue to action. Investigations into how realism translates into perceived threat, and what degree of dramatic portrayal of alcohol consequences is acceptable, would further advance VR-based alcohol prevention. Here, user involvement is essential, as Virtual LimitLab users stated that they value the absence of fear appeals, and a one-sided negative portrayal of alcohol use and cyber sickness should be avoided as far as possible [5].

Regarding the second aim of this study, we found consistent differences in the assessment by gender. Female participants agreed significantly more often than males that the simulation made it clear to them that drinking behaviour can be influenced by peer pressure, made them think about their own or others’ behaviour at parties and improved their perceived ability to make better decisions about how to react when offered alcohol. These findings can be linked to the results of a review by Dir et al. [19] which found that girls react more sensitively to peer pressure than boys. This could explain why it was mainly female visitors who became aware that peer pressure can influence their own drinking behaviour. Research indicates that women are generally more open to preventive and health-promoting programmes than men [20] and also show greater health awareness [21], which may also explain why women tended to report a higher perceived preventive impact than men in our study. Therefore, there is a need for gender-sensitive approaches in developing VR-based alcohol prevention [22].

No differences were found for hedonic quality, which is consistent with a study of the simulation VR FestLab [4] that showed that hedonic quality was not associated with socio-demographic factors such as gender. In addition, a review by Grassini and Laumann [23] showed that user experience of head-mounted displays such as VR did not differ between genders.

Moreover, the findings also differed by age. Adolescents (≤18 years) rated Virtual LimitLab as significantly more exciting and interesting than visitors aged 19 and older, but they were less likely to rate it as effective in demonstrating the influence of peer pressure on drinking behaviour compared to older users. Although the technical penetration of smartphones, laptops and tablets among adolescents is well advanced, it is possible that some adolescents have not yet had any experience with VR. The integration of a novel technology that has yet to become a regular feature in the lives of adolescents may have stimulated their interest and enthusiasm, thereby providing a potential explanation for these results. Conversely, the perceived learning experiences, particularly the influence of peer pressure, were more pronounced among older visitors, which may be attributed to the more frequent experience of peer pressure in personal situations. While adolescents may have demonstrated greater interest and enthusiasm for the novel technology, older visitors exhibited a greater focus on the specific design and preventative focus of the simulation. As the existing VR applications for alcohol prevention are targeted toward adolescents [4,5], there are no data on differences in user experience and perceived learning experiences between adolescents and older users. Investigating differences in user experience and (perceived) learning experiences of an intervention regarding gender and age is fundamental for tailoring it to specific target groups. Such knowledge can enhance the intervention’s efficacy and demonstrate sensitivity to the specific characteristics of a demographic group.

The literature on VR-based public health interventions in the arts is currently sparse. One example of public health-related use of VR in a cultural heritage setting is the Mind Space museum in Hong Kong, where visitors can experience psychiatric symptoms in VR simulations as part of the exhibition for a better understanding of mental health disorders [24], but no scientific evaluation is available. Art and museums for the improvement in health outcomes and knowledge transfer can potentially also be experienced without being physically present in such settings. Indeed, an immersive art exhibition during the COVID-19 pandemic was found to improve mood [25], and a “VR Diet Museum” received initial positive user feedback, suggesting the potential to use museums in VR to educate users about healthy eating and influence food choices [26].

More generally, museums and art galleries have been suggested as sites for public health programmes without the use of VR in interventions, as studies show positive correlations between participation in cultural activities and health [7,8], and museums offer favourable conditions for reaching individuals (often non-stigmatising and inexpensive to visit) [7]. Our finding that visitors to an art exhibition evaluate the VR application for alcohol prevention as even more positive than in other settings (school or laboratory) supports this suggestion. Given the dearth of the existing literature on this topic, we contend that our findings can be regarded as a contribution to the broader field of VR-based alcohol prevention. Our findings provide a rationale for furthering the connection between VR, health promotion and arts, as well as for testing non-virtual public health interventions in cultural heritage settings. Future research should investigate which type of individuals are reached in such settings, as museums and galleries are much more frequently visited by people of upper socio-economic status [27].

Strengths and Limitations

This study’s principal strength lies in its pioneering application of a VR-based alcohol prevention intervention in a cultural heritage setting, specifically an art exhibition. Both quantitative and qualitative data of the participants were gathered. The use of the UEQ-S is also a strength as it enabled comparisons with previous studies that used the same measurement instrument. Apart from proposing a new setting for VR-based alcohol prevention, this study also confirmed earlier qualitative findings for Virtual LimitLab from an adolescent user sample [4] with quantitative data, thus expanding the knowledge base on the user experience and perceived learning experience of the only VR application for alcohol prevention currently available in Germany.

However, this study has limitations. We collected the data via convenience sampling, which may have reduced the representativeness of our sample. The data presented should only be described as exploratory and cannot be used to draw general conclusions. The ages of the participants ranged from adolescents ≤18 years to people aged ≥40 years. Users aged ≤18 years made up around 20% of our participants, while they comprised 30% of the entire visitor group of the exhibition. This may have affected the comparability with other work focussing on adolescents and the applicability of our findings to the practice of VR-based alcohol prevention. The fact that adults who were either public visitors or teachers and prevention experts were part of the sample may have had an impact on the results. Due to this reason, we conducted age-specific analyses. Even though the sample size of 293 participants posed challenges for these analyses, it can be derived from the comparisons by age group that adolescents generally rate Virtual LimitLab positively, which reflects the non-stratified results. In fact, adolescents rated Virtual LimitLab as even more exciting and interesting than the participants aged 19 and above. Another limitation lies in the design of this study which consisted of cross-sectional data collection after game play without a control group not exposed to the game; this design limits conclusions on the effectiveness of Virtual LimitLab to effects of promoting learning experiences as perceived by study participants. As already stated above, controlled intervention studies would be needed to draw robust conclusions on the effects.

People interested in VR and/or alcohol prevention may have been more likely to participate in the art exhibition, which could have biased the results in different ways. For instance, they could have evaluated the application more critically, which could have led to a less positive rating compared to the general population. However, they may also have rated the simulation as more exciting and interesting because of their interest in new technology. The diversity of perspectives in the suggestions for improvement may also have been influenced by an unbalanced sample. Moreover, we only collected gender and age as socio-demographic information, which limited possible subgroup analyses. However, limited individual data were deliberately collected in order to keep the questionnaire short and fully anonymous and to encourage participation. Finally, it should be noted that we used Oculus Quest devices, and the experience of users may not be generalisable to other and future VR devices [4].

5. Conclusions

In summary, the VR application for alcohol prevention Virtual LimitLab was generally positively rated by visitors to an art exhibition in terms of the hedonic quality and perceived learning experiences, indicating the promising potential of (VR-based) public health interventions in the arts setting. The results can stimulate further research to better understand the effects of combining health promotion and art. Future plans involve establishing a permanent exhibition enabling research that compares the effects of virtual reality-based and traditional art exhibits on user attitudes regarding substance use and addiction. In general terms, we recommend combining the simulation with other evidence-based prevention measures to leverage synergistic effects.