Search

This paper presents a novel method for predicting suicidal ideation from electronic health records (EHR) and ecological momentary assessment (EMA) data using deep sequential models. Both EHR longitudinal data and EMA question forms are defined by asynchronous, variable length, randomly sampled data sequences. In our method, we model each of them with a recurrent neural network, and both sequences are aligned by concatenating the hidden state of each of them using temporal marks. Furthermore, we incorporate attention schemes to improve performance in long sequences and time-independent pre-trained schemes to cope with very short sequences. Using a database of 1023 patients, our experimental results show that the addition of EMA records boosts the system recall to predict the suicidal ideation diagnosis from 48.13% obtained exclusively from EHR-based state-of-the-art methods to 67.78%. Additionally, our method provides interpretability through the t-distributed stochastic neighbor embedding (t-SNE) representation of the latent space. Furthermore, the most relevant input features are identified and interpreted medically.

Suicidal behavior and its prevention constitute a major public health issue. Etiology of suicidal behavior is multifactorial. Whereas current research is mostly focused on clinical and biological risk factors, the sociodemographic risk factors for suicidal behavior, first highlighted by Durkheim, have received less attention. Besides the well-known impact of age and gender, sociodemographic variables such as marital and parental status, education, occupation, income, employment status, religion, migration or minority status, and sexual orientation are repeatedly reported to play an important role in suicidal behavior. This narrative review aimed to summarize recent research on sociodemographic risk factors for suicidal behavior and to elicit possible implications for suicide prevention.

Objective: Suicide, a major cause of death worldwide, has distinct biological underpinnings. The authors review and synthesize the research literature on biomarkers of suicide, with the aim of using the findings of these studies to develop a coherent model for the biological diathesis for suicide., Method: The authors examined studies covering a large range of neurobiological systems implicated in suicide. They provide succinct descriptions of each system to provide a context for interpreting the meaning of findings in suicide., Results: Several lines of evidence implicate dysregulation in stress response systems, especially the hypothalamic-pituitary-adrenal axis, as a diathesis for suicide. Additional findings related to neuroinflammatory indices, glutamatergic function, and neuronal plasticity at the cellular and circuitry level may reflect downstream effects of such dysregulation. Whether serotonergic abnormalities observed in individuals who have died by suicide are independent of stress response abnormalities is an unresolved question., Conclusions: The most compelling biomarkers for suicide are linked to altered stress responses and their downstream effects, and to abnormalities in the serotonergic system. Studying these systems in parallel and in the same populations may elucidate the role of each and their interplay, possibly leading to identification of new treatment targets and biological predictors.

Background: Whether suicide attempters and completers represent the same population evaluated at different points along a progression towards suicide death, overlapping populations, or completely different populations is a problem still unresolved., Methods: 446 Adult suicide attempters and knowledgeable collateral informants for 190 adult suicide probands were interviewed. Sociodemographic and clinical data was collected for both groups using semi-structured interviews and structured assessments. Univariate analyses and logistic regression models were conducted to explore the similarities and differences between suicide attempters and completers., Results: Univariate analyses yielded significant differences in sociodemographics, recent life events, impulsivity, suicide intent, and distribution of Axis I and II disorders. A logistic regression model aimed at distinguishing suicide completers from attempters properly classified 90% of subjects. The most significant variables that distinguished suicide from attempted suicide were the presence of narcissistic personality disorder (OR=21.4; 95% CI=6.8-67.7), health problems (OR=20.6; 95% CI=5.6-75.9), male sex (OR=9.6; 95% CI=4.42-20.9), and alcohol abuse (OR=5.5; 95% CI=2.3-14.2)., Limitations: Our study shares the limitations of studies comparing suicide attempters and completers, namely that information from attempters can be obtained from the subject himself, whereas the assessment of completers depends on information from close family or friends. Furthermore, different semi-structured instruments assessed Axis I and Axis II disorders in suicide attempters and completers. Finally, we have no data on inter-rater reliability data., Conclusions: Suicide completers are more likely to be male and suffer from alcohol abuse, health problems (e.g. somatic illness), and narcissistic personality disorder. The findings emphasize the importance of implementing suicide prevention programs tailored to suicide attempters and completers., (© 2013 Elsevier B.V. All rights reserved.)

Objectives: A major interest in the assessment of suicide risk is to develop an accurate instrument, which could be easily adopted by clinicians. This article aims at identifying the most discriminative items from a collection of scales usually employed in the assessment of suicidal behavior., Methods: The answers to the Barrat Impulsiveness Scale, International Personality Disorder Evaluation Screening Questionnaire, Brown-Goodwin Lifetime History of Aggression, and Holmes & Rahe Social Readjustment Rating Scale provided by a group of 687 subjects (249 suicide attempters, 81 non-suicidal psychiatric inpatients, and 357 healthy controls) were used by the Lars-en algorithm to select the most discriminative items., Results: We achieved an average accuracy of 86.4%, a specificity of 89.6%, and a sensitivity of 80.8% in classifying suicide attempters using 27 out of the 154 items from the original scales., Conclusions: The 27 items reported here should be considered a preliminary step in the development of a new scale evaluating suicidal risk in settings where time is scarce., (Copyright © 2012 Elsevier Ltd. All rights reserved.)

The objective of this research was to examine whether the Holmes-Rahe Social Readjustment Rating Scale, a life event scale, can be used to identify suicide attempters. The Holmes-Rahe Social Readjustment Rating Scale's ability to identify suicide attempters was tested in 1183 subjects (478 suicide attempters, 197 psychiatric inpatients, and 508 healthy controls) using the Fisher Linear Discriminant Analysis and traditional psychometric methods. The Fisher Linear Discriminant Analysis outperformed traditional psychometric approaches (area under the curve: 0.85 vs. 0.78; p < 0.05) and indicated that this scale may be used to identify suicide attempters. The life events that better characterized suicide attempters were change in frequency of arguments, marital separation, and personal injury. The Holmes-Rahe Social Readjustment Rating Scale may help identify suicide attempters.

Objective: Patients with bipolar disorder are prone to suicidal behavior, yet possible protective mechanisms are rarely studied. We investigated a possible protective role for moral or religious objections to suicide against suicidal ideation and attempts in depressed bipolar patients., Method: A retrospective case control study of 149 depressed bipolar patients (DSM-III-R criteria) in a tertiary care university research clinic was conducted. Patients who reported religious affiliation were compared with 51 patients without religious affiliation in terms of sociodemographic and clinical characteristics and history of suicidal behavior. The primary outcome measure was the moral or religious objections to suicide subscale of the Reasons for Living Inventory (RFLI)., Results: Religiously affiliated patients had more children and more family-oriented social networks than nonaffiliated patients. As for clinical variables, religiously affiliated patients had fewer past suicide attempts, had fewer suicides in first-degree relatives, and were older at the time of first suicide attempt than unaffiliated patients. Furthermore, patients with religious affiliation had comparatively higher scores on the moral or religious objections to suicide subscale of the RFLI, lower lifetime aggression, and less comorbid alcohol and substance abuse and childhood abuse experience. After controlling for confounders, higher aggression scores (P = .001) and lower score on the moral or religious objections to suicide subscale of the RFLI (P < .001) were significantly associated with suicidal behavior in depressed bipolar patients. Moral or religious objections to suicide mediated the effects of religious affiliation on suicidal behavior in this sample., Conclusions: Higher score on the moral or religious objections to suicide subscale of the RFLI is associated with fewer suicidal acts in depressed bipolar patients. The strength of this association was comparable to that of aggression scores and suicidal behavior, and had an independent effect. A possible protective role of moral or religious objections to suicide deserves consideration in the assessment and treatment of suicidality in bipolar disorder., (© Copyright 2011 Physicians Postgraduate Press, Inc.)

Background: Prior studies on the association between sleep disturbances and suicidal behavior did not explore whether or not short sleep is a marker of suicide intent, lethality or risk., Design: Cross-sectional., Participants: Suicide attempters (SAs) (n=434). Controls included 83 psychiatric inpatients who have never been SAs, and 509 healthy controls., Measurements: Short sleep was defined by self-assessment as ≤ 5 h per day. The MINI and the DSM-IV version of the International Personality Disorder Examination Screening Questionnaire were used to diagnose Axis I and Axis II diagnoses, respectively. Suicide intent and lethality were evaluated through the Beck's Suicidal Intent Scale (SIS) and the Risk-Rescue Rating Scale (RRRS), respectively. Beck's Medical Lethality Scale (BMLS) was administered to assess the degree of medical injury, and the SAD PERSONS mnemonic scale was used to evaluate suicide risk., Statistical Analyses: Chi-square tests and logistic regression analyses explored frequencies of short sleep in 3 samples. Chi-square tests explored whether or not suicide intent, lethality and risk were greater in SAs with short-sleep versus those without short-sleep., Results: Short sleep was more prevalent in SAs than in psychiatric controls only in males. In female SAs, short sleep was significantly associated with several SIS items and high scores in the SAD PERSONS., Limitations: Sleep duration was assessed only by self-report., Conclusions: The association between short sleep and suicidal behavior may be partly explained by confounders. Short sleep may be a marker of severity of suicidal behavior among female SAs., (Copyright © 2011 Elsevier B.V. All rights reserved.)

Background: Recent studies have suggested that genetic predisposition to suicidal behavior may be independent of the risk of suicide associated to mental disorders, such as affective disorders, schizophrenia or alcohol dependence. Given the suicidal behavior heterogeneity and its hereditary complexity, the need to find demonstrable intermediate phenotypes that may make it possible to establish links between genes and suicide behaviors (endophenotypes) seems to be necessary. The main objective is to review which are the candidate endophenotypes of suicidal behaviors., Methods: We carried out a non-systematic review of all published literature in English, French and Spanish in MEDLINE. The search terms were endophenotypes and suicide behaviors., Conclusions: The main candidate endophenotypes of suicidal behaviors are neuropsychological (decision-making, executive functions), personality traits (impulsivity, aggressiveness, and neuroticism), neurochemistry (5-HIAA in CNS) and neuroimaging (fMRI of cerebral amygdala or PET of prefrontal cortex metabolism).

The aim of the study is to compare the prevalence of suicidal ideation and attempts in the United States in 1991-1992 and 2001-2002, and identify sociodemographic groups at increased risk for suicidal ideation and attempts. Data were drawn from the National Institute on Alcohol Abuse and Alcoholism 1991-1992 National Longitudinal Alcohol Epidemiologic Survey (n=42,862) and the 2001-2002 National Epidemiologic Survey on Alcohol and Related Conditions (n=43,093), two nationally representative household surveys of non-institutionalized civilians aged 18 years and older, residing in the United States. The lifetime prevalence of suicide attempts remained unchanged in the United States between 1991-1992 and 2001-2002. Specific groups, namely 18- to 24-year-old white and black women, 25- to 44-year-old white women and 45- to 64-year-old Native American men were identified as being at high risk for suicide attempts. Despite prevention and treatment efforts, the lifetime prevalence of suicide attempts remains unchanged. Given the morbidity and mortality associated with suicide attempts, urgent action is needed to decrease the prevalence of suicide attempts in the United States.

Objective: Suicidal behavior is a clinically significant but underestimated cause of mortality in narcissistic personality disorder. Currently, there are no reliable estimates of suicidal behavior for this population. The main objective of this study was to test whether or not suicide attempters diagnosed with narcissistic personality disorder are different in terms of impulsivity and expected lethality from suicide attempters with other cluster B personality disorders., Method: In a sample of 446 suicide attempters, patients with cluster B personality disorder diagnoses (n = 254) as assessed by the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV), version of the International Personality Disorder Examination-Screening Questionnaire (IPDE-SQ) were compared in terms of expected lethality and impulsivity (measured by the Beck Suicidal Intent Scale and the Barratt Impulsiveness Scale, respectively). The subjects were admitted to the emergency departments of the Ramón y Cajal Hospital and the Fundación Jiménez Diaz University Hospital in Madrid, Spain, between January 1999 and January 2003., Results: Suicide attempts of subjects diagnosed with narcissistic personality disorder had higher expected lethality than those of subjects without narcissistic personality disorder (t = -4.24, df = 439, P < .001). There were no differences in expected lethality of the attempts when comparing subjects with and without histrionic personality disorder (t = 0.28, df = 439, P = .795), antisocial personality disorder (t = 0.66, df = 439, P = .504), and borderline personality disorder (t = 1.13, df = 439, P = .256), respectively. Suicide attempters diagnosed with narcissistic personality disorder did not significantly differ from suicide attempters without narcissistic personality disorder in terms of impulsivity measures (t = -0.33, df = 442, P = .738), while suicide attempters diagnosed with antisocial personality disorder, histrionic personality disorder, and borderline personality disorder were significantly more impulsive than suicide attempters without these diagnoses (t = -3.96, df = 442, P < .001; t = -3.88, df = 442, P < .001; and t = -7.44, df = 442, P < .001, respectively)., Conclusions: Narcissistic personality disorder seems to be a distinct group among cluster B personality disorders with regard to suicidal behavior. In contrast to suicide attempters with other cluster B personality disorders, suicide attempters diagnosed with narcissistic personality disorder are less impulsive and have suicide attempts characterized by higher lethality. These distinctions may offer a basis for targeted therapies aimed at decreasing suicidal risk in patients with narcissistic personality disorder., (Copyright 2009 Physicians Postgraduate Press, Inc.)

This article provides an update on suicidal behaviors in young women. The rates of completed suicide and suicide attempts among young women are reviewed, and the impact of race and ethnicity on these rates is described. The risk and protective factors associated with suicidal behaviors in young women are discussed, including stressful life events, mental disorders, and hormonal factors. Finally, some considerations for treating suicidal young women are included.

Five factor analyses with limitations explored the Suicidal Intent Scale (SIS) subscales reflecting suicidal behavior dimensions. This larger sample study conducts an exploratory factor analysis of the SIS. Two large samples of suicide attempters (N= 435 and N= 252) from a general hospital were studied. The validity of SIS subscales obtained from the factor analysis was investigated by examining the association between the subscales and clinical variables. There were two factors: expected lethality and planning. In both samples, male gender and depression tended to be associated with higher scores in both subscales (small to medium effect sizes). Hospitalization was associated with higher scores in both SIS subscales (medium to large effects) suggesting that these subscales were reasonably good predictors of suicide attempt severity. Clinicians assessing patient reports to establish the severity of suicide attempts need to ask questions regarding both dimensions: expected lethality and planning.

Objective: Severity of personality disorders (PDs) may be more useful in estimating suicide risk than the diagnosis of specific PDs. We hypothesized that suicide attempters with severe PD would present more attempts and attempts of greater severity/lethality. Method: Four hundred and forty-six suicide attempters were assessed. PD diagnosis was made using the International Personality Disorder Questionnaire – Screening Questionnaire. PDs were classified using Tyrer and Johnson’s classification of severity (no PD, simple PD, diffuse PD). Severity/lethality of attempts was measured with the Suicide Intent Scale, Risk-Rescue Rating Scale and Lethality Rating Scale. Results: Attempters with severe (diffuse) PD had more attempts than the other groups. After controlling for age and gender, this difference remained significant only for the younger age group and women. There was no relationship between severity of PDs and severity/lethality of attempts. Conclusion: Younger female attempters with severe PD are prone to repeated attempts. However, the severity of PD was not related to the severity/lethality of suicide attempts. [ABSTRACT FROM AUTHOR]

Introduction: Goodman expanded the conceptualization of addictions to embrace not only drug addiction but other behavioral addictions. In some cases, suicidal behaviour can be viewed as a behavioral addiction. Objectives: The main objective of the present study is to analyze the relationships between major suicide repeaters (> or =5 lifetime suicide attempts) and measures of suicidal behaviour addiction. Aims: To characterize major suicide repeaters. Methods: Sample and procedure: This is a transversal study of 954 suicide attempters (Montpellier, France). All suicide attempters were assessed using a protocol including: MINI (Axis I disorders), TPQ (personality traits) and BIS-10 (impulsivity), among others. Statistical Analyses: Comparisons between groups was made using logistic regression models with crude odds ratios and 95% confidence intervals. Results: Major repeaters were more likely to be female and having low educational level than non-major repeaters (OR[95%]=6.95[3.19–15.10]; p<0.0001; and OR[95%]=2.17[1.38–3.33]; p<0.001, respectively). As compared with non-major repeaters, major repeaters more often met criteria for bipolar disorder (OR[95%]=1.82[1.22–2.74]; p<0.05), anxiety disorders (OR[95%]=1.77[1.03–3.07]; p<0.05) and eating disorders (OR[95%]=2.81[1.79–4.41]; p<0.0001). Furthermore, compared with non-major repeaters, major repeaters were more frequently diagnosed with cigarette smoking (63.5% vs. 53.5%), alcohol use (29.3% vs. 25.4%) and substance use (15.4% vs. 13.2%), but none of them reached statistical significance. Finally, major repeaters, as compared with non-major repeaters, were more likely to score high in harm avoidance (OR[95%]=2.52[1.52–4.18];p<0.001), BIS-10 global score (OR[95%]=2.09[1.25–3.47]; p<0.05) and BIS-10 non-planning impulsiveness (OR[95%]=3.31[1.37–7.99]; p<0.05). Conclusions: Our preliminary results give partial support to the addictive hypothesis of suicidal behaviour. [Copyright &y& Elsevier]