Facteurs associés au risque suicidaire chez les jeunes consultants d'un centre de prévention sanitaire et sociale

The association between suicidal risk and various psychological or biographical factors in teenagers or young adults is already well documented. Yet, the role of stressful life events or contexts during childhood or of the recent past, as well as the respective weight of such determinants, has to be specified. Methods - One thousand one hundred and thirty-nine individuals, aged 16 to 25, who consecutively consulted in a preventive health center supported by the National Health Insurance System, located in Seine-Saint-Denis (a French department characterized by an unfavourable socio-economic context) on the occasion of a free work up were invited to fill out several self-administered questionnaires, aimed at assessing especially the level of psychosocial distress (Golberg's GHQ-28) and the level of hopelessness (Beck's hopelessness scale). They were also invited to meet a psychologist for a semi-structured interview, when the day of their consultation coincided with one of the three days a week the psychologist was present in the center ; the interview was aimed at collecting information upon the biographical context and ancient or recent life events and to determine the level of suicidal risk, on the basis of a scale of suicidal ideation [Ducher's Suicidal Risk Scale (ERSD)]. The concurrent validity of the later has already been previously tested and positive correlation coefficients were found with Beck Depression Inventory, Hamilton's Depression Rating Scale and Beck's Hopelessness Scale. Results - One thousand and four records could be analysed, as regards self-administered questionnaires, and among those, 576 as regards the interview with the psychologist and data related to suicidal risk. The studied population included 61.3 % of females and 59.3 % of individuals aged 20 to 25 : mean age was comparable in males and females. GHQ-28 global score and sub-scores (somatisation, anxiety, social dysfunction and depressive mood) were all higher in women (all the p < 0.001). A high suicidal risk (ERSD score4) was found in 24.1 % of the studied population. Subjects presenting with a high suicidal risk were characterized by higher levels of GHQ-28 psychosocial distress and GHQ-28 sub-scores as well as hopelessness (all the p < 0.001). Several biographical antecedents during childhood were significantly associated with suicidal risk : unknown father (p < 0.001), death of parents (p < 0.001), separation from parents (p < 0.001), severe quarrel between parents (p < 0.001), money problems within the family (p < 0.007), disorders related with alcohol consumption in parents (p < 0.016), drug addiction within the family (p < 0.001). Other predictors were several recent stressful events or contexts : violence within the family (p < 0.001), social isolation (p < 0.001), lack of self-esteem of (p < 0.002), school difficulties (p < 0.001), educational failure (p < 0.001) ; as well as the notion of a consumption of drugs (p = 0,001) or medications : neuroleptics (p < 0.015), antidepressants (p = 0.001) and tranquilizers (p < 0.001). A series of univariate regression analyses allowed to compute the Odds Ratios (OR) and the 95 % Confidence Intervals (95 % CI) of the sub-group characterized by a high suicidal risk for each socio-demographic, psychological and biographical independent variable, linked to suicidal risk at a threshold of p < 0.10. A multiple regression analysis was then performed in 2 steps : in a first step, independent variables were pooled by blocks, according to their nature (psychological characteristics, relational deficiencies among biographical antecedents, other stressful conditions among antecedents, stressful conditions among recent biographical context, recent consumption of drugs or medications) ; in a second step, all the independent variables which still remained associated with suicidal risk within each block were included in a final multiple regression analysis. Five variables continued to independently predict a high suicidal risk : hopelessness at Beck's scale (OR = 4.09), depressive mood at GHQ-28 (OR = 3.75), the notion of an unknown father (OR = 2.95), the notion of a recent destabilising event other than a school problem or an aggression (OR = 1.90) and the notion of an educational failure (OR = 1.78). Conclusion - These results confirm previous scientific data on this topic and underline that childhood context, educational course, psychological vulnerability and the occurrence of recent stressful life events combine their effects, enhancing the risk of a suicidal attempt. They can be useful for better sensitising educational as well as social and health care circles, for settling more efficient screening and preventive programs.