ORIAS is a randomized controlled trial designed to study the efficacy of the ELIOS system in reducing suicidality in the AYA, reinforcing their motivation to seek help and making them access to care.

As ELIOS would be the first suicide prevention system worldwide entirely tailored to the AYA contemporaneous modes of interaction on social media, ORIAS was designed to address the challenge of deciding whether this innovation is worth adding to the current national and international prevention arsenal. While randomized control trials are crucially lacking in the field of suicide prevention, especially on the Internet, the high-level evidence that ORIAS is expected to bring could have a decisive influence on how the French prevention strategies will seize the social media.

ELIOS will consist in a digital system that distressed users can contact via different social media or traditional communication channels. Participants randomized to the ELIOS group will be offered to interact online with a web-clinician who will deliver adapted interventions to alleviate their distress and refer them to appropriate care.The ELIOS web-clinicians will be psychologists and nurses supervised by a psychiatrist, specially trained to carry out crisis, counselling and motivational at distance interventions with AYA. In addition, they will receive a special training module dedicated to the interaction constraints and possibilities that the social media imply, but also to the textual and iconographic communication modalities. Human resources of ELIOS will be shared with those of the VigilanS program.

The ORIAS study primarily aims at showing the superiority of the ELIOS system in decreasing suicidality at 3 months in comparison to simple professional contact delivery among the AYA who seek help online for suicidal ideations.

With that purpose, participants will be randomly allocated to 2 parallel arms

• In addition to receiving the same generic mental health resources as in the control group, the participants of the ELIOS arm will benefit from the intervention of the ELIOS team. This includes an initial systematic contact with an ELIOS web-clinician, systematic recontacts by the ELIOS web-clinicians, scheduled at 24h, 72h and/or 7 days, depending on the participants' level of suicide risk and the opportunity to spontaneously contact the ELIOS system and get in contact with a web-clinician. All the ELIOS interventions will be subordinated to the two main goals of reducing the users' distress and suicidality and achieving an effective referral to the mental health care services. To do so, the web-clinicians will variably combine three complementary psychotherapeutic techniques: crisis intervention, counselling and motivational guidance. The interventions will be constrained by the imperative of rapidly bringing the participants from social media channels to more engaging channels (phone call or video-conference) in order to gain degrees of freedom in the interaction.
• Participants of the control arm will receive generic professional help contacts. Three types of resources will be suggested via the ELIOS website: the General Practitioner (GP), the closest Medical and Psychological Center, the emergency services. The order of presentation of the resource contacts will depend on the participants' intensity of suicidal ideation.

In both arms, data will be collected by means of online self-questionnaire before the intervention or provision of resource contacts (T0) and 3 months later (T1).

As primary efficacy analysis, the change in the Columbia Suicide Severity Rating Scale Intensity of Ideation-subscale score between baseline and M3 will be compared between the 2 arms using the constrained longitudinal data analysis (cLDA) model. To take into account the stratified factor considered in randomization, gender will be included as a covariate in the cLDA model. The treatment effect (mean between-arms difference in the 3-months change from baseline) will be estimated by the time-by-arm interaction. If normality of the model residuals is not satisfied (despite the log-transformation of the data), non-parametric analysis will be used; absolute changes between baseline and M3 will be calculated and compared between the 2 arms using non-parametric analysis of covariance adjusted for baseline values. Primary analysis will be conducted according to the ITT principle after handling missing values. A complete case-analysis will be performed as sensitivity analysis. In a second sensitivity analysis, the primary endpoint will be compared between the 2 arms by including in the cLDA model pre-specified covariates known for their association with suicidal behaviors, namely age, regular consumption of tobacco, regular consumption of alcohol, regular consumption of drugs, known pre-existing psychiatric disorder and personal history of suicide attempt. The necessary subjects number is 386 (193 per arm).

Inclusion Criteria:

• To spontaneously contact the ELIOS online system
• To have been experiencing suicidal ideations in the week prior the contact
• To live in France and speak French
• To provide informed consent

For ethical reasons, we decided not to include minors, due to impossibility of getting the parental authorizations without compromising the facilitation role of the system. However, most recent expert consensus consider that adolescence extends to 25 years old.

Exclusion Criteria:

• To contact the ELIOS online system for a relative or an acquaintance