Study for the Use Smartphone Application to Prevent Suicidal Relapse Among 15-35 Years-old With Previous Suicide Attempted

Suicide is the 2nd cause of death during adolescence in Europe (1.200/year, 7.9 % of deaths in this age group) and the prevalence of Lifetime Suicide Attempts (SA) in this population is 4.2%. Within a year of the SA, the repetition rate in adolescent populations ranges from 15% to 28%.

Post-SA management is crucial and treatment compliance is a major concern in adolescent populations. A French consensus psychiatric workgroup recommends that when patient is discharged from emergency room (ER) or hospitalization, it is important to organize health care. Unfortunately, these recommendations are imprecise and, moreover, from 40 to 77% of adolescents with previous SAs do not follow the recommendations of the post crisis program. In real clinical practice, patients barely follow the recommended treatment.

Reducing the SA repetition rate and increasing compliance with the recommended post-crisis program are two synergic issues.

The incredible and recent adoption of smartphone health applications (apps), that collect data (weight, exercises…) and allow the consumer to see graphs and diagrams, illustrates the sociological and psychological power of digital self-care and self-management. This phenomenon is particularly important for adolescents and young adults, who are considered to be digital natives. Medical experience is mainly based on the idea that is could be very useful to ask the patient to fill in scales or questionnaires on his/her smartphone instead of on a computer or a paper sheet. This is called Ecological Momentary Assessment (EMA) as it is a naturalistic method to access clinical data.

There are a few positive experiments in self-management (or self-care) using smartphone applications in psychiatry. In the field of suicide and mood disorders they have very interesting and promising results (high filling rates and better compliance to heath care programs). However, despite a few self-care aspects most of these apps are based on psycho-education programs and suffer in reality from lack of feedback.

Second field of apps use is Ecological Momentary intervention (EMI). Main idea is to use EMA data in algorithm in order to produce personalized comments and advises for patients. With the development of machine learning, it is obvious that EMI will have great implication in the future. There are few experiences with EMI but recent reviews all suggest that it is results are promising in mood disorders and anxiety.

The investigators are proposing an EMA + EMI new approach to prevent patient SA and Suicide based on a mobile healthcare application. This application is not connected. This is noticeably different then connected approach, using connected technology (phone, visual-phone and text messages) such as the Suicide Intervention Assisted by Message (SIAM) program, with text messages sent (sometimes automatically) by healthcare professionals to high-risk suicide subjects has shown positive preliminary results, with a reduced repetition rate. The App will be customised for each user when redeemed with a personal code provided by the practitioner. The App will collect data from the patient regarding anxiety, mood and sleep disorders twice a day (with a decreased frequence along the period of use), essentially based on analogic visual scales and drop-down lists. An algorithm-based feedback will pro-actively inform the patient with comments and advice based on WHO recommendations and/or associated with self-coping or Mindfulness practices. The aim is to position the smartphone application like a health care partner, it's name will be Medical Companion.

In order to assess the efficiency of the complete app the investigators will first lead a multicentre open pilot study with 6 months of follow-up in an adolescents and young adults population with a history of suicide attempts.

The application is an association of a "conventional" EMA program and an innovative algorithm-based response to the patient. Built on the notion of self-care, this app is not connected and the patient's responses are processed in the algorithm: (i) to provide feed-back regarding their mental state (contextualized, i.e. "you seem to be less anxious then yesterday") and, (ii) to advice them about what to do (for example "call a friend" "breath deeply 5 times in a row"…). Each mental status will be analysed by the Bayesian application algorithm. Regular evaluations are also provided on screen with easy to see graphs that might be viewed by the patient. These graphs could (and should) be seen by the psychiatrist, or psychologist involved in the Care as Usual program that all participants will still have. In sum, medical practitioners involved in the patient's usual care will have access to synthetic presentation of patient's responses and the investigators plan to ask participants about their experience.