Global trends in teenage suicide: 2003-2014.

The object of this article is to review the past decade of research on teenage suicide, with a particular emphasis on epidemiologic trends by age, gender and indigenous ethnicity. As such, a review of research literature from 2003 to 2014 was conducted via a comprehensive search of relevant psychological and medical databases. Wide gaps in our knowledge base exist concerning the true extent of teenage suicide due to lack of data, particularly in developing countries, resulting in a Western bias. The gender paradox of elevated suicidality in females with higher completed suicide rates in males is observed in teenage populations worldwide, with the notable exceptions of China and India. Native and indigenous ethnic minority teens are at significantly increased risk of suicide in comparison to general population peers. Often those with the highest need for mental health care (such as the suicidal adolescent) have least access to therapeutic support.Globally, suicide in teenagers remains a major public health concern. Further focused research concerning completed suicides of youth below the age of 18 is required across countries and cultures to understand more about risk as children progress through adolescence. Gender and ethnic variations in suicidality are embedded within cultural, historical, psychological, relational and socio-economic domains. Worldwide, the absence of child/adolescent-specific mental health policies may delay the development of care and suicide prevention. Overall, it is vital that clinicians adopt a holistic approach that incorporates an awareness of age and gender influences, and that cultural competency informs tailored and evaluated intervention programmes.

Psychosocial and risk behavior correlates of youth suicide attempts and suicidal ideation.

OBJECTIVE: To identify the independent psychosocial and risk behavior correlates of suicidal ideation and attempts. METHOD: The relationships between suicidal ideation or attempts and family environment, subject characteristics, and various risk behaviors were examined among 1,285 randomly selected children and adolescents, aged 9 through 17 years, of whom 42 (3.3%) had attempted suicide and 67 (5.2%) had expressed suicidal ideation only. The youths and their parents were enumerated and interviewed between December 1991 and July 1992 as part of the NIMH Methods for the Epidemiology of Child and Adolescent Mental Disorders (MECA) Study. RESULTS: Compared with subjects with suicidal ideation only, attempters were significantly more likely to have experienced stressful life events, to have become sexually active, to have smoked more than one cigarette daily, and to have a history of ever having smoked marijuana. After adjusting for sociodemographic characteristics, a statistically significant association was found between suicidal ideation or attempt and stressful life events, poor family environment, parental psychiatric history, low parental monitoring, low instrumental and social competence, sexual activity, marijuana use, recent drunkenness, current smoking, and physical fighting. Even after further adjusting for the presence of a mood, anxiety, or disruptive disorder, a significant association persisted between suicidal ideation or attempts and poor family environment, low parental monitoring, low youth instrumental competence, sexual activity, recent drunkenness, current smoking, and physical fighting. CONCLUSION: Low parental monitoring and risk behaviors (such as smoking, physical fighting, alcohol intoxication, and sexual activity) are independently associated with increased risk of suicidal ideation and attempts, even after adjusting for the presence of psychiatric disorder and sociodemographic variables.

Prescription of psychotropic medications to youths in office-based practice.

OBJECTIVE: This study sought to determine sociodemographic characteristics of treatment of children and adolescents for whom psychotropic medications are prescribed and to describe the clinical management approaches associated with the prescription of each major class of psychotropic medication in office-based medical practices in the United States. METHODS: Data for a four-year period (1992-1996) were drawn from the National Ambulatory Medical Care Survey, a nationally representative survey of office-based medical practices, to determine prescribing patterns, patients' sociodemographic characteristics, and clinical management approaches associated with visits during which psychotropic medications were prescribed to patients aged 19 years or under. RESULTS: Psychotropic medications were prescribed during 2.2 percent of all visits. A majority of the prescriptions for psychotropic medications (84.8 percent) were provided by general practitioners or pediatricians. For the visits during which a psychotropic medication was prescribed, stimulants were the most commonly prescribed (53.9 percent of such visits), but prescription of other classes of medications was not uncommon: antidepressants (30 percent), anxiolytics (7.2 percent), antipsychotics (7.2 percent), and mood stabilizers (12.7 percent). Significant differences were observed in the prescription of each class of medication by sex, race, and payment source. CONCLUSIONS: General practitioners and pediatricians have a role in the office-based treatment of youths with psychotropic medications.

Suicide and the media.

Evidence continues to amass on the significant impact of media coverage on suicide. The research literature on the impact of news reports of nonfictional suicides as well as fictional suicide stories is reviewed in order to determine the nature and scope of the influence of the mass media on suicide. The current review, building upon earlier reviews, is limited to English language publications or English translations of articles and/or abstracts. The interactive factors that may moderate the impact of media stories are also reviewed. Such interactive factors include characteristics of the stories (agent), individuals' attributes (host), and social context of the stories (environment). Recommendations are presented for the reporting of suicide stories, which may minimize the risk of imitative suicides. The media's positive role in educating the public about risks for suicide and shaping attitudes about suicide is emphasized. In summary, the existence of suicide contagion no longer needs to be questioned. We should refocus our research efforts on identifying which particular story components promote contagion under which circumstances and which components are useful for preventive programming.

Risk behavior in a community sample of children and adolescents.

OBJECTIVES: First, to investigate whether there is covariation between risk behaviors, including suicidality, in a community probability sample of children and adolescents; and second, to investigate whether risk behavior is associated with selected potential correlates. METHOD: A sample of 9- to 17-year-old youths (N = 1,285) and their caretakers were interviewed in the Methods for the Epidemiology of Child and Adolescent Mental Disorders (MECA) Study. The risk behaviors were marijuana smoking, alcohol use, intercourse, fighting, cigarette smoking, and suicidal ideation/attempts. Relationships between the risk behaviors were described using odds ratios. Linear regression analyses of an index of risk behavior on the selected potential correlates of risk behavior were conducted. RESULTS: There were significant relationships between all pairs of risk behaviors. The score on the index of risk behavior was associated with stressors, lack of resources, family psychiatric disorder, psychopathology, and functional impairment. CONCLUSIONS: Clinicians should be alerted to the possibility of risk behaviors, especially in children and adolescents engaging in other risk behaviors and those with inadequate resources, stressors, functional impairment, or psychopathology.

Parent-victim agreement in adolescent suicide research.

OBJECTIVE: The necessary absence of the victim's report in psychological autopsy studies is likely to introduce a systematic reporting bias. The authors investigated the nature and extent of this bias by examining agreement between accounts of symptomatology given by adolescents who made a serious suicide attempt and their parents. METHOD: Fifty-two attempters younger than age 20 were matched individually on age, sex, and ethnicity to a sample of suicide completers in two previous investigations. A semistructured interview was used to assess symptoms for DSM-III disorders. Diagnoses were assigned using two distinct computer algorithms, based respectively on youth and parent informant reports. Kappa values were estimated for each diagnostic category and related symptoms. RESULTS: Highest rates of parent-youth agreement were found for substance abuse and disruptive disorders. Rates of conduct disorder (nonaggressive) and major depression were reported less frequently by parents relative to their children. At the symptom level, parents less frequently reported Gets drunk and Frequency of drinking. CONCLUSIONS: Psychological autopsy studies of adolescent suicide are likely to underestimate the prevalence of psychiatric disorders in general and major depression and alcohol abuse in particular. These risk factors may be more predictive than previously thought.

Psychopathology associated with suicidal ideation and attempts among children and adolescents.

OBJECTIVE: To identify the independent and differential diagnostic and symptom correlates of suicidal ideation and suicide attempts and determine whether there are gender- and age-specific diagnostic profiles. METHOD: The relationships between suicidal ideation, suicide attempts, and psychiatric disorders were examined among 1,285 randomly selected children and adolescents, aged 9 to 17 years, of whom 42 had attempted suicide and 67 had expressed suicidal ideation only. Youths and their parents were interviewed as part of the Methods for the Epidemiology of Child and Adolescent Mental Disorders (MECA) Study, using the Diagnostic Interview Schedule for Children Version 2.3 (DISC-2.3). RESULTS: Logistic regression analyses indicated that mood, anxiety, and substance abuse/dependence disorders independently increased the risk of suicide attempts, after controlling for sociodemographic characteristics. There was no significant independent contribution of disruptive disorders to suicide attempts, although its association with suicidal ideation was significant. Substance abuse/dependence independently differentiated suicide attempters from ideators. Noncriterion symptoms that remained significant predictors of suicide risk, after adjusting for psychiatric disorder, included panic attacks and aggressiveness. Perfectionism did not significantly increase suicide risk after adjusting for psychiatric disorder. The association of specific disorders and noncriterion symptoms with suicidality varied as a function of gender and age. CONCLUSION: A monolithic diagnostic risk profile for suicidality, ignoring gender- and age-specific risks, is inadequate. The contribution of substance abuse/dependence in the escalation from suicidal thoughts to suicide attempts is underscored.

Separation/divorce and child and adolescent completed suicide.

OBJECTIVE: To investigate factors that may modify the effect of separation/divorce on youth suicide. METHOD: A case-control, psychological autopsy study of 120 of 170 consecutive suicides younger than age 20 and 147 community age-, sex-, and ethnic group-matched controls living in the greater New York area was conducted. Fifty-eight suicide victims and 49 community controls came from nonintact families of origin, indicating the permanent separation/divorce of the biological parents. Potential modifiers of separation/divorce include youth's age at separation, custodial parent's remarriage, nonresidential parent's frequency of contact, parent-child relationships, and parental psychopathology. RESULTS: The relatively small impact of separation/divorce was further diminished after accounting for parental psychopathology. An interaction of separation/divorce and the father-child relationship emerged. CONCLUSION: The dramatic increase in youth suicide during the past three decades seems unlikely to be attributable to the increase in divorce rates.

Psychosocial risk factors of child and adolescent completed suicide.

BACKGROUND: Few psychosocial risk factors for completed suicide in children and adolescents have been studied systematically. The present study was designed to examine the environmental, social, and familial characteristics of a large representative sample of child and adolescent suicides. METHODS: A case-control, psychological autopsy of 120 of 170 consecutive suicides younger than 20 years and 147 community age-, sex-, and ethnically matched control participants in the greater New York, NY, area. RESULTS: There was a significant independent impact of the psychosocial factors on increasing suicide risk among children and adolescents, beyond that risk attributable to psychiatric illness. The most notable risks were derived from school problems, a family history of suicidal behavior, poor parent-child communication, and stressful life events. Sex, ethnicity, and age modified the relationships of a few of the psychosocial factors. CONCLUSIONS: Socioenvironmental circumstances add significantly to a teenager's risk of suicide. The overall effect size on increasing suicide risk of the psychosocial factors is comparable with that for diagnostic factors, highlighting the importance of considering socioenvironmental factors when assessing suicide risk.

Psychiatric diagnosis in child and adolescent suicide.

BACKGROUND: The age, sex, and ethnic distribution of adolescents who commit suicide is significantly different from that of the general population. The present study was designed to examine psychiatric risk factors and the relationship between them and demographic variables. METHODS: A case-control, psychologic autopsy study of 120 of 170 consecutive subjects (age, <20 years) who committed suicide and 147 community age-, sex-, and ethnic-matched control subjects who had lived in the Greater New York (NY) area. RESULTS: By using parent informants only, 59% of subjects who committed suicide and 23% of control subjects who met DSM-III criteria for a psychiatric diagnosis, 49% and 26%, respectively, had had symptoms for more than 3 years, and 46% and 29%, respectively, had had previous contact with a mental health professional. Best-estimate rates, based on multiple informants for these parameters, for suicides only, were 91%, 52%, and 46%, respectively. Previous attempts and mood disorder were major risks factors for both sexes; substance and/or alcohol abuse was a risk factor for males only. Mood disorder was more common in females, substance and/or alcohol abuse occurred exclusively in males (62% of 18- to 19-year-old suicides). The prevalence of a psychiatric diagnosis and, in particular, substance and/or alcohol abuse increased with age. CONCLUSION: A limited range of diagnoses--most commonly a mood disorder alone or in combination with conduct disorder and/or substance abuse--characterizes most suicides among teenagers.

Clustering of attempted suicide: New Zealand national data.

BACKGROUND: The extent to which clusters of attempted suicides occur is a significant problem that is complementary to the current available research on the clustering of completed suicide. However, little systematic research on clusters of attempted suicides exists. The present study examines the extent and nature of clustering of suicide attempts. METHOD: The occurrence of clustering of attempted suicide was examined in nationwide data for all New Zealand hospitals, obtained from the New Zealand Health Statistics Services for the years 1988-1990. The Scan statistic and Knox procedure were employed for testing the significance of clusters in time and time-space, respectively. RESULTS: The analyses indicated that significant time clustering occurred in younger age groups, specifically among 15-19 and 20-24 year olds. The results could not be accounted for by seasonal variations in admissions. Age specificity of time-space clusters emerged, exhibiting a similar pattern to that reported for completed suicides in the US. CONCLUSIONS: The results suggest a similar underlying mechanism for the clustering of parasuicide and completed suicides and provide support for the existence of contagion of suicidal behaviour. The implications for prevention are discussed.

Correspondence between statistically derived behavior problem syndromes and child psychiatric diagnoses in a community sample.

The correspondence between Diagnostic and Statistical Manual (3rd ed.) (DSM-III) diagnoses and statistically derived syndromes was examined within a community sample of children and adolescents in Puerto Rico. Specifically, the extent to which behavior dimensions, derived from the Child Behavior Checklist and the Youth Self-Report, corresponded to psychiatric diagnoses, derived from parent and child versions of the Diagnostic Interview Schedule for Children, was examined. The alternative approaches for assessing psychopathology in children and adolescents were compared against external validators. The results indicated a meaningful convergence between DSM-III diagnoses and statistical syndromes; however, a one-to-one correspondence did not emerge. Little evidence was found for "diagnostic thresholds." There was no evidence of the superiority of either the statistically derived syndromes or the DSM-III diagnoses. The incorporation of a measure of impairment improved the validity of both approaches. Adding parental reports to the self-reports of adolescents yielded little gain in the validity of either the statistical or diagnostic approach. The implications for the definition and assessment of child and adolescent disorders are discussed.

Patterns of diagnostic comorbidity in a community sample of children aged 9 through 16 years.

Secondary analyses of the data from the Puerto Rico Child Psychiatry Epidemiologic Study were done to provide information on the comorbidity of four major diagnostic domains (attention deficit disorders, conduct/oppositional disorders, depression and anxiety disorders). A high level of comorbidity was found among these four domains of child and adolescent psychopathology. In general the patterns of comorbidity were not affected by whether the data was put together by a clinician or by means of a computer algorithm scoring a structured interview. The patterns were not affected in any major way by who the informants were in the diagnostic process. Minor differences were found in certain comorbidity patterns depending on the sex and age of the subjects. Comorbidity was associated with level of impairment and to service utilization.

Aggregating data from multiple informants in child psychiatry epidemiological research.

One of the issues confronting the field of child psychiatric epidemiology is devising ways to aggregate discrepant reports from multiple informants obtained in structured diagnostic assessments of children and adolescents. The present report uses data from an epidemiological community survey to attempt to address this issue. The results obtained through statistical procedures designed to identify the "optimal" informant for specific symptomatic criteria and diagnoses do not seem to offer a clear-cut advantage over a simple combinatorial rule that identifies symptomatic criteria as positive when they are acknowledged as positive by either parent or child informants.

Reliability and validity of a psychosocial axis in patients with child psychiatric disorder.

The interrater reliability of a classification scheme of psychosocial stressors was assessed among 52 child psychiatrists in rating exercises with 28 case histories. Validity was evaluated in a field study of 377 cases rated by 39 of the same child psychiatrists. The impact of glossary descriptions on reliability and validity was also determined. Reliability was low with most Kappa values being in the range of 0.20 to 0.40. The glossary did not affect interrater reliability, but it significantly decreased the use of abnormal codes. The discriminant validity of the codes was supported in that the most commonly used codes were used with differential frequency in a pattern that conforms with the known psychosocial correlates of the different psychiatric conditions of childhood.

Screening for childhood psychopathology in the community using the Child Behavior Checklist.

One of the uses of the Child Behavior Checklist is as a screening instrument for childhood psychopathology in two-phase designs. The present report involves a two-phase epidemiological survey conducted in Puerto Rico in which the CBCL was used as a screening instrument during the first stage, and children were evaluated clinically during the second stage. The data indicate that in using the CBCL for screening for psychopathology in children, parent information is most informative, particularly for children in the adolescent age group. Nevertheless, the data also reinforce the need to obtain teacher information with the Teacher Report Form to enhance screening sensitivity. In this population, the data obtained with the Youth Self-Report were found to be of limited usefulness for screening purposes.

Impairment in the epidemiological measurement of childhood psychopathology in the community.

The desirability of incorporating a measure of impairment to the categorization of childhood psychopathology in the community is examined. The use of the Children's Global Assessment Scale (CGAS) for this purpose is recommended. The choice of 61 (definite case) and 71 (probable case) as cutpoints on the Children's Global Assessment Scale is supported empirically by the data on service utilization, parental perceived need, and behavior problem scores obtained in the Puerto Rico Child Psychiatry Epidemiological Study.

Suicide clusters: an examination of age-specific effects.

The age specificity of time-space clusters of suicide was examined using National Center for Health Statistics data for 1978-84. Significant clustering of suicide occurred primarily among teenagers and young adults, with minimal effect beyond 24 years of age. Clustering was two to four times more common among adolescents and young adults than among other age groups.