Co-creating and evaluating a digital Escape Room for students on mental health in the Covid-19 era – EscapeCovid

The project was structured in two main parts: the co-creation of the game and its evaluation. The game was co-created using the PRODUCES technique which clearly defines: what problem (concerning a health behavior) we want to address with the intervention; the objective of the co-creation process; what participatory methodology we use; who are the end users; who is involved as co-creators; how the intervention will be evaluated; and what model to adopt to ensure its scalability. Following these indications, we created a first version of the game which was then tested by 45 students who were recruited in a snowball fashion from the school of public health attached to our research center. These were exclusively health students. First, they completed a scale on youth mental health literacy. Then, they played the Escape Game and gave their feedback on the game. This was done through a questionnaire with items such as "How much did you like the graphics?" or "Was the game difficult?" The scale on mental health literacy was re-proposed to observe an improvement in the score after the game, which was the case. Ten participants were randomly selected to answer a semi-directed interview. As health students, they were able to detail the game's flaws in terms of form and also content, explaining how to better talk about depression and suicide risk in a scientifically valid way that was adapted to a young audience. Back and forth with the developers and researchers helped finalize the game. The debriefing part was particularly important. A mediator was to help the player navigate through the rooms of the Escape Game and, above all, carry out a final debriefing by sharing key messages on mental health problems: how to recognize them, how to overcome them.

Once the game had been perfected thanks to the contribution of the target audience, we set up a randomized controlled trial for evaluation under experimental conditions. We sent emails with the presentation of the project to the main student associations of the Universities of Bordeaux and Bordeaux Montaigne, listed on the official websites of the universities. The email explained the stages of the project and its objectives. Tricky contributed to the recruitment of students from the Nursing Training Institute of the University Hospital Center Hôpital Pellegrin. Tricky's CEO, David Labrosse, spoke in lecture halls to boost recruitment. Emails were sent to the 40 partner schools of French Tech. Respondents to the first questionnaire (mental health literacy + scales on emotions, resilience and beliefs + access to care) were then randomized (by gender and field of study) into an intervention arm (game) and a control arm. The first 20 participants, 10 intervention and 10 control, were interviewed using the mixed methods approach: quantitative data from the questionnaires are analyzed together with qualitative data from the interviews using a convergent model.

The cocreation phase allowed us to obtain a game adapted to the needs of students in terms of form (gamifying approach) and content (accessibility of storytelling) within a solid theoretical framework, including mental health literacy. After the test, the terms used in the game were corrected to provide a realistic medical framework. The game characters were also modified to be closer to the image of a typical student. The colors, clues, debriefing, etc. were modified according to the preferences of the group of 45 testers.

The evaluation part has been started and should be completed beyond the end of the ANR funding. Indeed, additional funding from the Nouvelle Aquitaine Region has been obtained to complete the recruitment until the end of 2026. We are currently at 53 complete participations for the non-exposed arm and 20 complete participations for the exposed arm. We encountered difficulties because the game, although online, must be carried out in person at Tricky's premises to ensure that the session runs smoothly. Unstable internet connections or an obsolete computer can damage the session and, consequently, the experiment. Traveling after a long day of classes seemed to us to be the major obstacle to participation. For this, gift cards are offered to participants in both arms to boost their recruitment.

The preliminary results show a trend of improvement in the mental health literacy score and the desire to consult for a psychological problem. Nothing is significant and the sample remains unrepresentative.

As for the qualitative data, the interviews showed that the game was appreciated for the possibility of interacting and learning by doing. Information was more easily retained thanks to the immersive approach. According to the interviewees, understanding the experiences of others helped reduce stigma. The game also made it possible to learn more about support resources. The control group received an informative email on the definition of depression and useful addresses. This email was appreciated and the participants said they wanted to keep it in case of need. However, it was a tool without interaction, just to consult, without reinforcement of knowledge.

While waiting for the quantitative data (we are now aiming for 200 students, 100 per arm, while maintaining representativeness), we have already prepared on R all the programs to be launched to produce the statistics, such as the multivariate models. These programs are documented in a manual to be used once recruitment is complete.

At the same time, we wrote with Tricky a document that summarizes the "business model" based on the experience of this ANR funding.

The quantitative and qualitative data collected in this study will demonstrate the effectiveness of the Escape Game with the aim of making it a "medical device" to be marketed across universities in France. In addition, the general aim is to obtain a "gold standard" for the creation of other games on the mental health of populations.

Regarding the technology, the qualitative data collected during the "human" debriefing of the Escape Game will be used to automate exchanges with the mediator through reinforcement learning (artificial intelligence). The spoken text will be used to train ChatGPT to provide increasingly refined answers in the final part of the game.

The Covid-19 pandemic and the protracted lockdowns have heavily impacted French University students’ mental health. Digital Escape Rooms represent a good means to reach students and propose them solutions for their psychological well-being. This project aims to co-create, optimize and evaluate a digital Escape Room on students’ mental health in the Covid-19 era. The intervention should: increase mental health literacy, help appraise and change beliefs about mental health, support recognition and management of emotions, and teach positive coping strategies with regards to Covid-19. Co-creation and optimization will imply students, researchers, health professionals and IT developers involved in focus groups/workshops (hackathon). For the evaluation, a randomized controlled trial with pre- and post-test data collection (online questionnaires with validated scales) will be conducted among 500 students in Bordeaux. A sample of 30 students will be randomly selected for responding to a semi-structured interview following a mixed-methods design.