What’s hidden beneath the surface? Suicidal behavior in adolescents: trends, risk groups, risk factors, course and prediction

Omschrijving

Suicidal behavior is a major public health problem in adolescents. Worldwide, about one-third of adolescents report lifetime suicidal thoughts, one in ten adolescents have attempted suicide and almost 46,000 adolescents end their own lives every year. In the Netherlands, one adolescent dies by suicide every week. Adolescents with suicidal behavior also have a higher risk of problems later in life. Therefore, early screening and prevention of suicidal behavior is important. Although much epidemiological evidence on suicidal behavior is available, there are still gaps in our understanding of suicidal behavior in adolescents. The general aim of the research presented in this dissertation was therefore to contribute to a better understanding of suicidal behavior among adolescents (aged 13-16 years) in Amsterdam and vicinity. We used data from the school-based screening questionnaires and from the Digital Child Health Care Registry, which is a registration system for the outcomes of health assessments. Both of these sources are part of the routine health assessments of the Public Health Service (GGD) of Amsterdam. We considered the data to be representative of all adolescents 13-16 years old living in Amsterdam and vicinity. Part I of this dissertation describes time trends for emotional and behavioral problems and suicidal behavior. An important finding was that sociodemographic risk groups for emotional and behavioral problems and suicidal behavior changed over time. The prevalence of total difficulties in adolescents increased by 19% between school years 2004-2005 and 2013-2014 and the time trends differed by sex. The prevalence of suicidal thoughts decreased between 2010-2011 and 2014-2015 from 18% to 13% and of suicide attempts from 3% to 2%. Time trends differed between subgroups according to ethnicity and educational level. Part II of this dissertation describes possible risk factors of suicidal behavior during adolescence. Adolescents with overweight or obesity were more likely to have emotional and behavioral problems and suicidal behavior than their non-overweight peers, which is probably because they were also victimized more often. These associations were stronger for adolescents with obesity than those with overweight. Adolescents who reported physical abuse, emotional neglect, emotional abuse or victimization had an increased risk of suicidal thoughts and suicide attempts. This risk was higher when maltreatment or victimization was more frequent or when several forms of maltreatment were involved simultaneously. For physical abuse, emotional neglect and multiple forms of maltreatment, the risk of suicidal behavior was higher for girls than for boys. Part III presents longitudinal research into the course of suicidal behavior and the possibilities of prediction. Relative to their peers without suicidal behavior, adolescents with suicidal thoughts reported suicidal thoughts five times more often (26% versus 5%) and suicidal attempts six times more often (3% versus 0.5%) two years later. Adolescents who reported a suicide attempt also reported suicidal thoughts four times more often (22% versus 5%) and suicide attempts 25 times more often (12.5% versus 0.5%). Persistent suicidal behavior two years later was associated with sex, education level, family situation, reporting emotional and behavioral problems and abuse. We also investigated whether machine learning techniques improved the prediction of future suicidal behavior in adolescents compared to usual practice, i.e. risk based on previous suicidal behavior. Random Forest achieved slightly higher sensitivity, with the same specificity. With Lasso Regression, sensitivity increased significantly but at the expense of specificity. In conclusion, recommendations for future research and policy are the following: (i) account for different risk groups and the fact that these risk groups can change over time; (ii) focus on modifiable environmental risk factors; (iii) ensure early screening and early treatment; and (iv) improve prediction of increased risk of adolescent suicidal behavior.