The suicidal behaviors are phenomena eminently multifactorial. It is thus always difficult to define univocal strategies of prevention of suicide repetition, during the emergency stay, i.e. almost in general population. One find 23 clinical trials in this topic in the past 25 years, and 18 are negative. The majority of the positive trials have the concern of being dissociated from an assumption of responsibility of care strictly speaking, to adopt a position "méta", nearer to the concept of "case management": how to remain in contact with the suicide attempter, without forcing it in this every day life, replacing a possible proposing, but assumption of responsibility resources reliable and quickly accessible in the event of at risk situation? Each one of these studies tests devices which seem more appropriate to such or such characteristic of this population, by retaining only simple criteria like the sex, the number of former suicide attempts, the proposal or not for an assumption of responsibility of care, the observance or not with the plan of care. Thus, it would seem interesting to combine these approaches in an algorithm entitled "ALGOS".

Main aim: To test the effectiveness of this algorithm of case management, named "ALGOS", in reducing the number of death by suicide, in terms of reduction of suicide re-attempts and the number of loss of contact patients in the ALGOS group during 6 months period, compared to a control group of suicide attempters treated as usual (i.e. primarily transmitted to the attending physician).

Secondary objectives: To evaluate, according to the method validated by Beecham in 1992, direct medico-economic impact in the year which follows the introduction of algorithm ALGOS. Reduction of the other suicidal behaviors in 6 months (reduction in the full number of suicidal repetitions in each group, evolution of the score of suicidal ideation, etc…). To evaluate the effect of the algorithm, at the 13th month. To study the possible differences within the time in terms of suicidal repetitions in the 2 groups. To propose different profiles of answers according to psychopathology, the number of suicide attempts, suicidal character, the sex,…

Methodology: Comparative simple blind prospective multicentric controlled study

Evaluations: The subjects will be included in the participating centers following a suicide attempt (SA). After checking of the criteria, information and signature of the study consent, the subjects will be randomized in one of the 2 groups (ALGOS versus Control). A fax of inclusion will be sent to the centralized research team of the Clinical Investigation Center (CIC) of Lille. All the subjects included will be then evaluated at the 6th month and 13th month after the SA, during a phone call. The final evaluation will be carried out in the current of the 13th month, so as to avoid a recontact at the time of the period birthday, too charged affectively at the majority with the patients. This phone call, personal, will be carried out by an trained psychologist. A mail preventing the subject of the call will be addressed approximately 10 days before. During the evaluations at the 6th month and 13th month, the number of suicidal repetitions will be reported, consumption of care since inclusion will be noted, the scale of suicidal ideation of Beck assessed, and a psychopathological evaluation carried out (Mini International Neuropsychiatric Interview DSM IV).

Inclusion Criteria:

• Surviving to a suicide attempt (SA) directly leaving the Emergency unit or have been hospitalized less than 7 days

Exclusion Criteria:

• recidivists who made 4 SA and more in the 3 past years