Prediction Models for Suicide Attempts and Deaths: A Systematic Review and Simulation

Brazilian Psychiatric Association guidelines for the management of suicidal behavior. Part 1. Risk factors, protective factors, and assessment

Suicide is a global public health problem that causes the loss of more than 800,000 lives each year, principally among young people. In Brazil, the average mortality rate attributable to suicide is approximately 5.23 per 100,000 population. Although many guidelines have been published for the management of suicidal behavior, to date, there are no recent guidelines based on the principles of evidence-based medicine that apply to the reality of suicide in Brazil. The objective of this work is to provide key guidelines for managing patients with suicidal behavior in Brazil. This project involved 11 Brazilian psychiatry professionals selected by the Psychiatric Emergencies Committee (Comissão de Emergências Psiquiátricas) of the Brazilian Psychiatric Association for their experience and knowledge in psychiatry and psychiatric emergencies. For the development of these guidelines, 79 articles were reviewed (from 5,362 initially collected and 755 abstracts). In this review, we present definitions, risk and protective factors, assessments, and an introduction to the Safety Plan.

Systematic review registry number: CRD42020206517

Linking Public Safety And Public Health Data For Firearm Suicide Prevention In Utah

In Utah, a state with a high rate of gun ownership, the shared concerns of diverse stakeholders generated bipartisan support for a state-funded study that tracked patterns of firearm suicide. The study linked sensitive public health and public safety data and identified opportunities for firearm suicide prevention. Findings reported to the state legislature included the proportion of suicide decedents who could have passed a background check for legal firearm possession at their time of death, had a permit to carry a concealed firearm, or had been seen in the hospital for a previous suicide attempt or self-harm. Within six months of the report's release, the legislature, health care and religious groups, and state agencies had launched diverse, major initiatives to reduce firearm suicide that were informed by the report's findings. We present the Utah experience as a case study in bringing diverse stakeholders-particularly gun owners-together to find common ground on firearm suicide prevention and in using linked data to support and guide their efforts.

Decomposing implicit associations about life and death improves our understanding of suicidal behavior

The Death/Suicide Implicit Association Test (IAT) is effective at detecting and prospectively predicting suicidal thoughts and behaviors. However, traditional IAT scoring procedures used in all prior studies (i.e., D-scores) provide an aggregate score that is inherently relative, obfuscating the separate associations (i.e., "Me = Death/Suicide," "Me = Life") that might be most relevant for understanding suicide-related implicit cognition. Here, we decompose the D-scores and validate a new analytic technique called the Decomposed D-scores ("DD-scores") that creates separate scores for each category ("Me," "Not Me") in the IAT. Across large online volunteer samples (N > 12,000), results consistently showed that a weakened association between "Me = Life" is more strongly predictive of having a history of suicidal attempts than is a stronger association between "Me = Death/Suicide." These findings replicated across three different versions of the IAT and were observed when calculated using both reaction times and error rates. However, among those who previously attempted suicide, a strengthened association between "Me = Death" is more strongly predictive of the recency of a suicide attempt. These results suggest that decomposing traditional IAT D-scores can offer new insights into the mental associations that may underlie clinical phenomena and may help to improve the prediction, and ultimately the prevention, of these clinical outcomes.

What Differentiates Prisoners Who Attempt Suicide from Those Who Experience Suicidal Ideation? A Nationally Representative Study

OBJECTIVE

Many people who think about suicide do not engage in suicidal behavior. Identifying risk factors implicated in the process of behavioral enaction is crucial for suicide prevention, particularly in high-risk groups such as prisoners.

METHOD

Cross-sectional data were drawn from a nationally representative sample of 17,891 prisoners (79% men) in the United States. We compared prisoners who attempted suicide (attempters; n = 2,496) with those who thought about suicide but never made an attempt (ideators; n = 1,716) on a range of established risk factors.

RESULTS

More than half (59%) of participants who experienced suicidal ideation had also attempted suicide. Violent offending, trauma, brain injury, alcohol abuse, and certain mental disorders distinguished attempters from ideators.

CONCLUSION

Our results fit within recent ideation-to-action theories that emphasize the role of a capability for suicide in the transition from thoughts to acts of suicide.

Genome-Wide Association Study of Suicide Death and Polygenic Prediction of Clinical Antecedents

OBJECTIVE

Death by suicide is a highly preventable yet growing worldwide health crisis. To date, there has been a lack of adequately powered genomic studies of suicide, with no sizable suicide death cohorts available for analysis. To address this limitation, the authors conducted the first comprehensive genomic analysis of suicide death using previously unpublished genotype data from a large population-ascertained cohort.

METHODS

The analysis sample comprised 3,413 population-ascertained case subjects of European ancestry and 14,810 ancestrally matched control subjects. Analytical methods included principal component analysis for ancestral matching and adjusting for population stratification, linear mixed model genome-wide association testing (conditional on genetic-relatedness matrix), gene and gene set-enrichment testing, and polygenic score analyses, as well as single-nucleotide polymorphism (SNP) heritability and genetic correlation estimation using linkage disequilibrium score regression.

RESULTS

Genome-wide association analysis identified two genome-wide significant loci (involving six SNPs: rs34399104, rs35518298, rs34053895, rs66828456, rs35502061, and rs35256367). Gene-based analyses implicated 22 genes on chromosomes 13, 15, 16, 17, and 19 (q<0.05). Suicide death heritability was estimated at an h2SNP value of 0.25 (SE=0.04) and a value of 0.16 (SE=0.02) when converted to a liability scale. Notably, suicide polygenic scores were significantly predictive across training and test sets. Polygenic scores for several other psychiatric disorders and psychological traits were also predictive, particularly scores for behavioral disinhibition and major depressive disorder.

CONCLUSIONS

Multiple genome-wide significant loci and genes were identified and polygenic score prediction of suicide death case-control status was demonstrated, adjusting for ancestry, in independent training and test sets. Additionally, the suicide death sample was found to have increased genetic risk for behavioral disinhibition, major depressive disorder, depressive symptoms, autism spectrum disorder, psychosis, and alcohol use disorder compared with the control sample.

Interpersonal theory of suicide: prospective examination

BACKGROUND

The interpersonal theory of suicide (IPTS) is one of the most intensively researched contemporary theories on the development of suicidal ideation and behaviour. However, there is a lack of carefully conducted prospective studies.

AIMS

To evaluate the main predictions of the IPTS regarding the importance of perceived burdensomeness, thwarted belongingness and capability for suicide in predicting future suicide attempts in a prospective design.

METHOD

Psychiatric in-patients (n = 308; 53.6% (n = 165) female; mean age 36.82 years, s.d. = 14.30, range 18-81) admitted for severe suicidal ideation (n = 145, 47.1%) or a suicide attempt completed self-report measures of thwarted belongingness, perceived burdensomeness, capability for suicide, hopelessness, depression and suicidal ideation as well as interviews on suicide intent and suicide attempts and were followed up for 12 months. Logistic regression and receiver operating characteristics (ROC) analysis were conducted.

RESULTS

The interaction of perceived burdensomeness, thwarted belongingness and capability for suicide was not predictive of future suicide attempts, but perceived burdensomeness showed a significant main effect (z = 3.49, P < 0.01; OR = 2.34, 95% CI 1.59-3.58) and moderate performance in screening for future suicide attempts (area under the curve AUC = 0.729, P < 0.01).

CONCLUSIONS

The results challenge the theoretical validity of the IPTS and its clinical utility - at least within the methodological limitations of the current study. Yet, findings underscore the importance of perceived burdensomeness in understanding suicidal ideation and behaviour.

Identifying risk factors for mortality among patients previously hospitalized for a suicide attempt

Age-adjusted suicide rates in the US have increased over the past two decades across all age groups. The ability to identify risk factors for suicidal behavior is critical to selected and indicated prevention efforts among those at elevated risk of suicide. We used widely available statewide hospitalization data to identify and test the joint predictive power of clinical risk factors associated with death by suicide for patients previously hospitalized for a suicide attempt (N = 19,057). Twenty-eight clinical factors from the prior suicide attempt were found to be significantly associated with the hazard of subsequent suicide mortality. These risk factors and their two-way interactions were used to build a joint predictive model via stepwise regression, in which the predicted individual survival probability was found to be a valid measure of risk for later suicide death. A high-risk group with a four-fold increase in suicide mortality risk was identified based on the out-of-sample predicted survival probabilities. This study demonstrates that the combination of state-level hospital discharge and mortality data can be used to identify suicide attempters who are at high risk of subsequent suicide death.

Prospective Validity of the Suicide Cognitions Scale Among Acutely Suicidal Military Personnel Seeking Unscheduled Psychiatric Intervention

Abstract. Background and Aim: The Suicide Cognitions Scale (SCS) was developed to assess a broad range of suicide-related cognitions. Research to date supports the scale's factor structure, internal consistency, and construct validity. The present study tested the scale's prospective validity for suicide attempts among 97 military personnel presenting to an emergency department or psychiatric outpatient clinic for an unscheduled walk-in evaluation. Method: Cox regression and receiver operator characteristic analyses were conducted to test the prospective validity of the SCS. Results: Results supported the prospective validity of the SCS (area under the curve [AUC] = 0.69) and indicate the scale's performance is comparable to an empirically supported measure of suicide ideation (AUC = 0.75). The SCS performance was not reduced by removing items containing the word suicide. Limitations: Homogeneous sample comprised of US soldiers, predominantly male, with recent suicidal ideation. Conclusion: Results support the SCS as an indicator of subsequent risk for suicidal behavior when used in acute care settings, and suggest the scale's performance is similar to more traditional suicide risk screening methods that depend on honest self-disclosure of suicidal thoughts.

Understanding the Transition from Suicidal Desire to Planning and Preparation: Correlates of Suicide Risk within a Psychiatric Inpatient Sample of Ideators and Attempters

BACKGROUND

There is a clear need to better understand the trajectory from suicidal ideation to enactment of lethal suicidal behavior. Identification of factors that promote desire and the transition to intent and behavior is critical for the advancement of theory, risk formulation, and prevention.

METHOD

In this cross sectional study, correlates of suicide risk were examined at theoretically distinct points along the trajectory from suicidal thinking to behavior (i.e., desire, plans and preparations, suicide attempt) in a manner consistent with the Three-Step Theory and an ideation-to-action framework. The sample included 197 adult inpatients (60% male, 40% white) hospitalized due to ideation or a recent suicide attempt.

RESULTS

Psychological pain and fearlessness about death were associated with desire and plans and preparations for suicide. There were no significant differences in suicide risk correlates between ideators and attempters.

LIMITATIONS

The primary limitations of the current study relate to the cross-sectional design and the nature of the sample, which do not allow for inference of causal relations, or generalizability to outpatient and community samples or to individuals who die by suicide.

CONCLUSIONS

Psychological pain and fearlessness about death may function as transitional factors that are associated with the transition from desire to suicidal intent in psychiatric inpatients. Findings have important implications for clinical practice. Treatment interventions should reduce psychological pain, increase safety, and reduce access to means.

Applying machine learning on health record data from general practitioners to predict suicidality

BACKGROUND

Suicidal behaviour is difficult to detect in the general practice. Machine learning (ML) algorithms using routinely collected data might support General Practitioners (GPs) in the detection of suicidal behaviour. In this paper, we applied machine learning techniques to support GPs recognizing suicidal behaviour in primary care patients using routinely collected general practice data.

METHODS

This case-control study used data from a national representative primary care database including over 1.5 million patients (Nivel Primary Care Database). Patients with a suicide (attempt) in 2017 were selected as cases (N = 574) and an at risk control group (N = 207,308) was selected from patients with psychological vulnerability but without a suicide attempt in 2017. RandomForest was trained on a small subsample of the data (training set), and evaluated on unseen data (test set).

RESULTS

Almost two-third (65%) of the cases visited their GP within the last 30 days before the suicide (attempt). RandomForest showed a positive predictive value (PPV) of 0.05 (0.04-0.06), with a sensitivity of 0.39 (0.32-0.47) and area under the curve (AUC) of 0.85 (0.81-0.88). Almost all controls were accurately labeled as controls (specificity = 0.98 (0.97-0.98)). Among a sample of 650 at-risk primary care patients, the algorithm would label 20 patients as high-risk. Of those, one would be an actual case and additionally, one case would be missed.

CONCLUSION

In this study, we applied machine learning to predict suicidal behaviour using general practice data. Our results showed that these techniques can be used as a complementary step in the identification and stratification of patients at risk of suicidal behaviour. The results are encouraging and provide a first step to use automated screening directly in clinical practice. Additional data from different social domains, such as employment and education, might improve accuracy.

Supervised Machine Learning: A Brief Primer

Machine learning is increasingly used in mental health research and has the potential to advance our understanding of how to characterize, predict, and treat mental disorders and associated adverse health outcomes (e.g., suicidal behavior). Machine learning offers new tools to overcome challenges for which traditional statistical methods are not well-suited. This paper provides an overview of machine learning with a specific focus on supervised learning (i.e., methods that are designed to predict or classify an outcome of interest). Several common supervised learning methods are described, along with applied examples from the published literature. We also provide an overview of supervised learning model building, validation, and performance evaluation. Finally, challenges in creating robust and generalizable machine learning algorithms are discussed.

Evaluation of official procedures for suicide prevention in hospital from a forensic psychiatric and a risk management perspective

Background: Suicide is a severe public health problem, in 2008 the Italian ministerial recommendation n° 4 on the management of suicide defined key areas for the identification of suicidal risk in hospital wards. The guidelines are important in defining professional liability issues, in line with Law 24 of 8/3/2017 'Gelli-Bianco'. Our study aimed to investigate the appropriateness of the official documents on suicide prevention delivered by Italian hospitals and their compliance with the ministerial recommendation.Methods: The Italian hospitals' public procedures on suicide prevention issued between 2008 and 2019 (n = 33) were retrieved thorough web search and further evaluated according to their compliance with the 2008 Italian ministerial recommendations.Results: The guidelines documents were generally in line with the ministerial recommendation. However, we found a lack of implementation in the specific training of health professionals. Most guidelines provided no risk stratification, nor specific procedures for different risk degrees or diagnoses. More than half of the documents did not report standardised tools for the assessment of suicidal risk.Conclusions: The public procedures on suicide prevention in Italian hospitals present general indications, leaving room for interpretation. Public procedures should be implemented with greater attention to the elements of judgement in the assessment of suicidal risk.KEY POINTSProcedures for suicide prevention are of uttermost importance for psychiatrist working in hospital.Standards in suicide risk evaluations are needed.Comparison between procedures can improve risk assessment and evaluation.

Violence victimization and suicide attempts among adolescents aged 12-15 years from thirty-eight low- and middle-income countries

OBJECTIVE

The association between violence victimization and suicide attempts in a large representative sample of adolescents from low- and middle-income-countries (LMICs) of multiple continents has never been investigated. Therefore, the aim of the present study was to examine the relationship between being a victim of physical attacks (independent variable) and suicide attempts (dependent variable) in a sample of 117,472 students aged 12-15 years [mean (SD) age 13.8 (0.9) years; girls 49.4%] from thirty-eight LMICs in Africa, the Americas, and Asia.

METHODS

Cross-sectional data from the Global School-based Student Health Survey (GSHS) were analyzed. Self-reported data on past 12-month suicide attempts and exposure to physical attacks were collected. Logistic regression and meta-analysis were conducted.

RESULTS

The overall prevalence of suicide attempts and physical attacks were 10.1% and 39.4%, respectively. Overall, the results of the meta-analysis based on country-wise estimates adjusted for potential confounders (i.e., age, sex, food insecurity, alcohol consumption, bullying victimization, anxiety-induced sleep problems, low parental support/involvement, loneliness) showed that physical attacks were associated with a 1.71 (95%CI = 1.62-1.81) times higher odds for suicide attempt.

CONCLUSIONS

In this large sample of adolescents from multiple LMICs, violence victimization was associated with significantly increased odds of suicide attempts. Future longitudinal studies are required to assess causality, and whether addressing exposure to violence can positively impact on adolescent suicide rates.

The architecture of co-morbidity networks of physical and mental health conditions in military veterans

Co-morbidity between medical and psychiatric conditions is commonly considered between individual pairs of conditions. However, an important alternative is to consider all conditions as part of a co-morbidity network, which encompasses all interactions between patients and a healthcare system. Analysis of co-morbidity networks could detect and quantify general tendencies not observed by smaller-scale studies. Here, we investigate the co-morbidity network derived from longitudinal healthcare records from approximately 1 million United States military veterans, a population disproportionately impacted by psychiatric morbidity and psychological trauma. Network analyses revealed marked and heterogenous patterns of co-morbidity, including a multi-scale community structure composed of groups of commonly co-morbid conditions. Psychiatric conditions including posttraumatic stress disorder were strong predictors of future medical morbidity. Neurological conditions and conditions associated with chronic pain were particularly highly co-morbid with psychiatric conditions. Across conditions, the degree of co-morbidity was positively associated with mortality. Co-morbidity was modified by biological sex and could be used to predict future diagnostic status, with out-of-sample prediction accuracy of 90-92%. Understanding complex patterns of disease co-morbidity has the potential to lead to improved designs of systems of care and the development of targeted interventions that consider the broader context of mental and physical health.

Suicides in Mood Disorders in Psychiatric Settings in Nordic National Register-Based Studies

OBJECTIVE

Although risk factors for nonfatal suicidal behavior in mood disorders have been vastly investigated, rate and risk factors of suicide deaths are less well known. Extensive health care and other population registers in the Nordic countries (Denmark, Finland, Iceland, Norway, and Sweden) allow national-level studies of suicide rates and risk factors. This systematic review examined Nordic studies of suicide in mood disorders.

METHODS

National Nordic studies published after 1.1.2000 reporting on suicide mortality or relative risk in diagnosed unipolar depression or bipolar disorder treated in psychiatric settings; temporal variations in suicide risk after discharge, or risk factors for suicide were systematically reviewed.

RESULTS

Altogether 16 longitudinal studies reported on rate and risk of suicide in depression. They found 2%-8% of psychiatric inpatients with depression to have died by suicide. However, in Finland suicide risk among depressive inpatients halved since the early 1990s. Nine studies investigated suicide risk in bipolar disorder, finding 4-8% of patients having died by suicide in long term. The relative risk of suicide was consistently found extremely high (SMR > 100) during the first weeks postdischarge, declining steeply over time to approximately SMR of five after five years. Male gender, preceding suicide attempts, high severity of depression and substance abuse were found risk factors for suicide in depression, with only minor gender differences in risk factors, but major differences in lethal methods. Three studies investigated risk factors for suicide in bipolar disorder, finding male gender, preceding suicide attempts, and depressive episodes and psychiatric comorbidity to be associated with risk.

CONCLUSIONS

Overall, of psychiatric inpatients with depressive of bipolar disorders in the Nordic countries, 2%-8% have died by suicide in the last few decades, but current rates may be lower. Suicide risk is approximately similar or somewhat higher among patients with bipolar disorder, risk factor studies of whom are fewer. Risk of suicide is remarkably high immediately after discharge, and higher among males than females, those with preceding suicide attempts, high severity of depression, or concurrent substance abuse. Generalizability of findings from these Nordic studies to other countries need to be investigated, and their methodological limitations understood.

The Process of Developing a Unit-Based Army Resilience Program

PURPOSE OF REVIEW

We review military doctrine, military public health data, medical literature, and educational literature with the intent of condensing key precepts into a succinct, pragmatic description of the essential steps for leaders looking to build a resilience program to provide secondary prevention services.

RECENT FINDINGS

Although there continues to be a shortage of high-level evidence in support of specific preventive programs, there are numerous large-scale reviews of prevention and health promotion efforts. When combined with population-specific analyses, several essential concepts emerge as most relevant for smaller-scale prevention programs. The key tenets that program leaders should embrace to optimize program effectiveness include utilization of an instructional design approach, focus on evidence-based practices, and teaching resilience skills in order to decrease risk factors and increase protective factors for improved mental health outcomes.

Catalonia Suicide Risk Code Epidemiology (CSRC-Epi) study: protocol for a population-representative nested case-control study of suicide attempts in Catalonia, Spain

INTRODUCTION

Suicide attempts represent an important public health burden. Centralised electronic health record (EHR) systems have high potential to provide suicide attempt surveillance, to inform public health action aimed at reducing risk for suicide attempt in the population, and to provide data-driven clinical decision support for suicide risk assessment across healthcare settings. To exploit this potential, we designed the Catalonia Suicide Risk Code Epidemiology (CSRC-Epi) study. Using centralised EHR data from the entire public healthcare system of Catalonia, Spain, the CSRC-Epi study aims to estimate reliable suicide attempt incidence rates, identify suicide attempt risk factors and develop validated suicide attempt risk prediction tools.

METHODS AND ANALYSIS

The CSRC-Epi study is registry-based study, specifically, a two-stage exposure-enriched nested case-control study of suicide attempts during the period 2014-2019 in Catalonia, Spain. The primary study outcome consists of first and repeat attempts during the observation period. Cases will come from a case register linked to a suicide attempt surveillance programme, which offers in-depth psychiatric evaluations to all Catalan residents who present to clinical care with any suspected risk for suicide. Predictor variables will come from centralised EHR systems representing all relevant healthcare settings. The study's sampling frame will be constructed using population-representative administrative lists of Catalan residents. Inverse probability weights will restore representativeness of the original population. Analysis will include the calculation of age-standardised and sex-standardised suicide attempt incidence rates. Logistic regression will identify suicide attempt risk factors on the individual level (ie, relative risk) and the population level (ie, population attributable risk proportions). Machine learning techniques will be used to develop suicide attempt risk prediction tools.

ETHICS AND DISSEMINATION

This protocol is approved by the Parc de Salut Mar Clinical Research Ethics Committee (2017/7431/I). Dissemination will include peer-reviewed scientific publications, scientific reports for hospital and government authorities, and updated clinical guidelines.

TRIAL REGISTRATION NUMBER

NCT04235127.

Can machine-learning methods really help predict suicide?

PURPOSE OF REVIEW

In recent years there has been interest in the use of machine learning in suicide research in reaction to the failure of traditional statistical methods to produce clinically useful models of future suicide. The current review summarizes recent prediction studies in the suicide literature including those using machine learning approaches to understand what value these novel approaches add.

RECENT FINDINGS

Studies using machine learning to predict suicide deaths report area under the curve that are only modestly greater than, and sensitivities that are equal to, those reported in studies using more conventional predictive methods. Positive predictive value remains around 1% among the cohort studies with a base rate that was not inflated by case-control methodology.

SUMMARY

Machine learning or artificial intelligence may afford opportunities in mental health research and in the clinical care of suicidal patients. However, application of such techniques should be carefully considered to avoid repeating the mistakes of existing methodologies. Prediction studies using machine-learning methods have yet to make a major contribution to our understanding of the field and are unproven as clinically useful tools.

Routinized categorization of suicide risk into actionable strata: Establishing the validity of an existing suicide risk assessment framework in an outpatient sample

OBJECTIVE

The Suicide Risk Assessment and Management Decision Tree (DT) is a clinician-administered assessment that leads to risk categorizations that correspond with actionable strata. This study investigated the construct validity and test-retest reliability of the DT risk categories across two time points.

METHOD

Outpatients (N = 731) completed a battery of self-report measures. Spearman's correlations were used to examine the relationships between DT suicide risk level and suicidal symptoms, theory-based risk factors, psychiatric correlates, and DT suicide risk level at Timepoint 2. Correlations were analyzed for significant differences to examine the divergent validity of the DT.

RESULTS

Results, overall, were in line with hypotheses, with the exception of depression and thwarted belongingness.

CONCLUSIONS

Findings provide evidence for the reliability, convergent validity, and discriminant validity of the DT. This clinician-administered suicide risk assessment may be useful for standardization of the assessment and management of suicide risk in outpatient clinical settings.

Psychological distress of suicide attempters predicts one-year suicidal deaths during 2007-2016: A population-based study

BACKGROUND/PURPOSE

Of the risk factors for suicide, prior attempt is regarded as one of the strongest for subsequent attempts or completed suicide. This large-scale cohort study aims to examine whether the distress level of individual mental symptoms and general psychopathology measured at the index attempt can predict subsequent suicide death within one year.

METHODS

The participants were 104,441 suicide attempters first reported to the Taiwan National Suicide Surveillance System during 2007-2016, who completed the five-item Brief Symptom Rating Scale (BSRS-5) at the index attempt. Kaplan-Meier survival curve analysis with log-rank test and Cox regression model were used to examine whether the level of psychological distress could predict the cumulative incidence of re-attempted suicidal death in the following year.

RESULTS

In total, 1254 (1.2%) participants subsequently killed themselves within one year. Survival curves analysis and Cox regression modelling indicated that levels of distress of individual items (i.e., suicide ideation, depression, inferiority, anxiety, hostility and insomnia) and total BSRS-5 scores were significantly correlated with the incidence of subsequent suicidal death within one year for both genders.

CONCLUSION

The study revealed that self-rated psychological distress was a significant and sustained predictor of re-attempted suicide death within one year after the index attempt. These results imply that suicide is not only an issue of acute crisis, but also a prolonged problem of lasting psychological distress. The BSRS-5 assessment could provide a symptom profile on which to develop a pertinent person-centered approach to prevent subsequent suicide attempts.

Suicide in the pediatric population: screening, risk assessment and treatment

The number of children and adolescents dying by suicide is increasing over time. Patterns for who is at risk are also changing, leading to a need to review clinical suicide prevention progress and identify limitations with existing practices and research that can help us further address this growing problem. This paper aims to synthesise the literature on paediatric suicide screening, risk assessment and treatment to inform clinical practice and suicide prevention efforts. Our review shows that universal screening is strongly recommended, feasible and acceptable, and that there are screening tools that have been validated with youth. However, screening may not accurately identify those at risk of dying due to the relative rarity of suicide death and the associated research and clinical challenges in studying such a rare event and predicting future behaviour. Similarly, while risk assessments have been developed and tested in some populations, there is limited research on their validity and challenges with their implementation. Several promising suicide-specific treatments have been developed for youth, but overall there is an insufficient number of randomised trials. Despite great need, the research evidence to support screening, risk assessment and treatment is still limited. As suicide rates increase for children and adolescents, continued research in all three domains is needed to reverse this trend.

What is bipolar disorder? A disease model of dysregulated energy expenditure

Advances in the understanding and management of bipolar disorder (BD) have been slow to emerge. Despite notable recent developments in neurosciences, our conceptualization of the nature of this mental disorder has not meaningfully progressed. One of the key reasons for this scenario is the continuing lack of a comprehensive disease model. Within the increasing complexity of modern research methods, there is a clear need for an overarching theoretical framework, in which findings are assimilated and predictions are generated. In this review and hypothesis article, we propose such a framework, one in which dysregulated energy expenditure is a primary, sufficient cause for BD. Our proposed model is centered on the disruption of the molecular and cellular network regulating energy production and expenditure, as well its potential secondary adaptations and compensatory mechanisms. We also focus on the putative longitudinal progression of this pathological process, considering its most likely periods for onset, such as critical periods that challenges energy homeostasis (e.g. neurodevelopment, social isolation), and the resulting short and long-term phenotypical manifestations.

Reaching Those at Highest Risk for Suicide: Development of a Model Using Machine Learning Methods for use With Native American Communities

OBJECTIVE

Suicide prevention is a major priority in Native American communities. We used machine learning with community-based suicide surveillance data to better identify those most at risk.

METHOD

This study leverages data from the Celebrating Life program operated by the White Mountain Apache Tribe in Arizona and in partnership with Johns Hopkins University. We examined N = 2,390 individuals with a validated suicide-related event between 2006 and 2017. Predictors included 73 variables (e.g., demographics, educational history, past mental health, and substance use). The outcome was suicide attempt 6, 12, and 24 months after an initial event. We tested four algorithmic approaches using cross-validation.

RESULTS

Area under the curves ranged from AUC = 0.81 (95% CI ± 0.08) for the decision tree classifiers to AUC = 0.87 (95% CI ± 0.04) for the ridge regression, results that were considerably higher than a past suicide attempt (AUC = 0.57; 95% CI ± 0.08). Selecting a cutoff value based on risk concentration plots yielded 0.88 sensitivity, 0.72 specificity, and a positive predictive value of 0.12 for detecting an attempt 24 months postindex event.

CONCLUSION

These models substantially improved our ability to determine who was most at risk in this community. Further work is needed including developing clinical guidance and external validation.

Does it get easier over time? Psychologists' experiences of working with suicidal patients

Does working with suicidal patients become easier with time? A representative national survey of 375 psychologists (72% women, Mean age 44 years) showed that years of experience (r = -.13, p = .01) and frequency of contact with suicidal patients (r = -.15, p = .004) correlated only weakly with perceived difficulty. Thematic analysis of respondents' descriptions of difficult suicide-related situations on an open survey-question yielded four themes: Unreachable patients, choosing between therapy and security, therapist's boundaries and empathy with death-wishes. We conclude that improved confidence in coping with suicidality may require specific training rather than simply years of work.

Prospective prediction of suicide attempts in community adolescents and young adults, using regression methods and machine learning

BACKGROUND

The use of machine learning (ML) algorithms to study suicidality has recently been recommended. Our aim was to explore whether ML approaches have the potential to improve the prediction of suicide attempt (SA) risk. Using the epidemiological multiwave prospective-longitudinal Early Developmental Stages of Psychopathology (EDSP) data set, we compared four algorithms-logistic regression, lasso, ridge, and random forest-in predicting a future SA in a community sample of adolescents and young adults.

METHODS

The EDSP Study prospectively assessed, over the course of 10 years, adolescents and young adults aged 14-24 years at baseline. Of 3021 subjects, 2797 were eligible for prospective analyses because they participated in at least one of the three follow-up assessments. Sixteen baseline predictors, all selected a priori from the literature, were used to predict follow-up SAs. Model performance was assessed using repeated nested 10-fold cross-validation. As the main measure of predictive performance we used the area under the curve (AUC).

RESULTS

The mean AUCs of the four predictive models, logistic regression, lasso, ridge, and random forest, were 0.828, 0.826, 0.829, and 0.824, respectively.

CONCLUSIONS

Based on our comparison, each algorithm performed equally well in distinguishing between a future SA case and a non-SA case in community adolescents and young adults. When choosing an algorithm, different considerations, however, such as ease of implementation, might in some instances lead to one algorithm being prioritized over another. Further research and replication studies are required in this regard.

Validation of an Electronic Health Record-Based Suicide Risk Prediction Modeling Approach Across Multiple Health Care Systems

Importance

Suicide is a leading cause of mortality, with suicide-related deaths increasing in recent years. Automated methods for individualized risk prediction have great potential to address this growing public health threat. To facilitate their adoption, they must first be validated across diverse health care settings.

Objective

To evaluate the generalizability and cross-site performance of a risk prediction method using readily available structured data from electronic health records in predicting incident suicide attempts across multiple, independent, US health care systems.

Design, Setting, and Participants

For this prognostic study, data were extracted from longitudinal electronic health record data comprising International Classification of Diseases, Ninth Revision diagnoses, laboratory test results, procedures codes, and medications for more than 3.7 million patients from 5 independent health care systems participating in the Accessible Research Commons for Health network. Across sites, 6 to 17 years' worth of data were available, up to 2018. Outcomes were defined by International Classification of Diseases, Ninth Revision codes reflecting incident suicide attempts (with positive predictive value >0.70 according to expert clinician medical record review). Models were trained using naive Bayes classifiers in each of the 5 systems. Models were cross-validated in independent data sets at each site, and performance metrics were calculated. Data analysis was performed from November 2017 to August 2019.

Main Outcomes and Measures

The primary outcome was suicide attempt as defined by a previously validated case definition using International Classification of Diseases, Ninth Revision codes. The accuracy and timeliness of the prediction were measured at each site.

Results

Across the 5 health care systems, of the 3 714 105 patients (2 130 454 female [57.2%]) included in the analysis, 39 162 cases (1.1%) were identified. Predictive features varied by site but, as expected, the most common predictors reflected mental health conditions (eg, borderline personality disorder, with odds ratios of 8.1-12.9, and bipolar disorder, with odds ratios of 0.9-9.1) and substance use disorders (eg, drug withdrawal syndrome, with odds ratios of 7.0-12.9). Despite variation in geographical location, demographic characteristics, and population health characteristics, model performance was similar across sites, with areas under the curve ranging from 0.71 (95% CI, 0.70-0.72) to 0.76 (95% CI, 0.75-0.77). Across sites, at a specificity of 90%, the models detected a mean of 38% of cases a mean of 2.1 years in advance.

Conclusions and Relevance

Across 5 diverse health care systems, a computationally efficient approach leveraging the full spectrum of structured electronic health record data was able to detect the risk of suicidal behavior in unselected patients. This approach could facilitate the development of clinical decision support tools that inform risk reduction interventions.

Pre-deployment predictors of suicide attempt during and after combat deployment: Results from the Army Study to Assess Risk and Resilience in Servicemembers

BACKGROUND

Deployment-related experiences might be risk factors for soldier suicides, in which case identification of vulnerable soldiers before deployment could inform preventive efforts. We investigated this possibility by using pre-deployment survey and administrative data in a sample of US Army soldiers to develop a risk model for suicide attempt (SA) during and shortly after deployment.

METHODS

Data came from the Army Study to Assess Risk and Resilience in Servicemembers Pre-Post Deployment Survey (PPDS). Soldiers completed a baseline survey shortly before deploying to Afghanistan in 2011-2012. Survey measures were used to predict SAs, defined using administrative and subsequent survey data, through 30 months after deployment. Models were built using penalized regression and ensemble machine learning methods.

RESULTS

Significant pre-deployment risk factors were history of traumatic brain injury, 9 + mental health treatment visits in the 12 months before deployment, young age, female, previously married, and low relationship quality. Cross-validated AUC of the best penalized and ensemble models were .75-.77. 21.3-40.4% of SAs occurred among the 5-10% of soldiers with highest predicted risk and positive predictive value (PPV) among these high-risk soldiers was 4.4-5.7%.

CONCLUSIONS

SA can be predicted significantly from pre-deployment data, but intervention planning needs to take PPV into consideration.

Clinical risk model to predict 28-day unplanned readmission via the accident and emergency department after discharge from acute psychiatric units for patients with psychotic spectrum disorders

BACKGROUND

Unplanned readmissions rates are an important indicator of the quality of care provided in a psychiatric unit. However, there is no validated risk model to predict this outcome in patients with psychotic spectrum disorders.

AIMS

This paper aims to establish a clinical risk prediction model to predict 28-day unplanned readmission via the accident and emergency department after discharge from acute psychiatric units for patients with psychotic spectrum disorders.

METHOD

Adult patients with psychotic spectrum disorders discharged within a 5-year period from all psychiatric units in Hong Kong were included in this study. Information on the socioeconomic background, past medical and psychiatric history, current discharge episode and Health of the Nation Outcome Scales (HoNOS) scores were used in a logistic regression to derive the risk model and the predictive variables. The sample was randomly split into two to derive (n = 10 219) and validate (n = 10 643) the model.

RESULTS

The rate of unplanned readmission was 7.09%. The risk factors for unplanned readmission include higher number of previous admissions, comorbid substance misuse, history of violence and a score of one or more in the discharge HoNOS overactivity or aggression item. Protective factors include older age, prescribing clozapine, living with family and relatives after discharge and imposition of conditional discharge. The model had moderate discriminative power with a c-statistic of 0.705 and 0.684 on the derivation and validation data-set.

CONCLUSIONS

The risk of readmission for each patient can be identified and adjustments in the treatment for those with a high risk may be implemented to prevent this undesirable outcome.

Stigma, biomarkers, and algorithmic bias: recommendations for precision behavioral health with artificial intelligence

Effective implementation of artificial intelligence in behavioral healthcare delivery depends on overcoming challenges that are pronounced in this domain. Self and social stigma contribute to under-reported symptoms, and under-coding worsens ascertainment. Health disparities contribute to algorithmic bias. Lack of reliable biological and clinical markers hinders model development, and model explainability challenges impede trust among users. In this perspective, we describe these challenges and discuss design and implementation recommendations to overcome them in intelligent systems for behavioral and mental health.

Why Are Suicide Rates Increasing in the United States? Towards a Multilevel Reimagination of Suicide Prevention

Suicide, a major public health concern, takes around 800,000 lives globally every year and is the second leading cause of death among adolescents and young adults. Despite substantial prevention efforts, between 1999 and 2017, suicide and nonfatal self-injury rates have experienced unprecedented increases across the United States - as well as in many other countries in the world. This chapter reviews the existing evidence on the causes behind increased suicide rates and critically evaluates the impact of a range of innovative approaches to suicide prevention. First, we briefly describe current trends in suicide and suicidal behaviors and relate them to recent time trends in relevant suicide risk markers. Then, we review the existing evidence in suicide prevention at the individual and the population levels, including new approaches that are currently under development. Finally, we advocate for a new generation of suicide research that examines causal factors beyond the proximal and clinical and fosters a socially conscious reimagining of suicidal prevention. To this end, we emphasize the need for the conceptualization of suicide and suicidal behaviors as complex phenomena with causes at several levels of organization. Future interdisciplinary research and interventions should be developed within a multilevel causal framework that can better capture the social, economic, and political settings where suicide, as a process, unfolds across the life course.

Suicide prediction models: a critical review of recent research with recommendations for the way forward

Suicide is a leading cause of death. A substantial proportion of the people who die by suicide come into contact with the health care system in the year before their death. This observation has resulted in the development of numerous suicide prediction tools to help target patients for preventive interventions. However, low sensitivity and low positive predictive value have led critics to argue that these tools have no clinical value. We review these tools and critiques here. We conclude that existing tools are suboptimal and that improvements, if they can be made, will require developers to work with more comprehensive predictor sets, staged screening designs, and advanced statistical analysis methods. We also conclude that although existing suicide prediction tools currently have little clinical value, and in some cases might do more harm than good, an even-handed assessment of the potential value of refined tools of this sort cannot currently be made because such an assessment would depend on evidence that currently does not exist about the effectiveness of preventive interventions. We argue that the only way to resolve this uncertainty is to link future efforts to develop or evaluate suicide prediction tools with concrete questions about specific clinical decisions aimed at reducing suicides and to evaluate the clinical value of these tools in terms of net benefit rather than sensitivity or positive predictive value. We also argue for a focus on the development of individualized treatment rules to help select the right suicide-focused treatments for the right patients at the right times. Challenges will exist in doing this because of the rarity of suicide even among patients considered high-risk, but we offer practical suggestions for how these challenges can be addressed.

Cytomegalovirus Seropositivity and Suicidal Behavior: A Mini-Review

Background and objectives: In recent years, a growing body of research has focused on identifying possible biological markers for suicidal behavior, including infective and immunological markers. In this paper, our aim was to review available evidence concerning the association between cytomegalovirus (CMV) infection and suicide. Materials and Methods: A systematic search according to the PRISMA statement was performed on Pubmed. After the screening procedure, we identified five relevant papers. Results: We found inconsistent evidence linking CMV infection and suicide, with some papers reporting an association between CMV seropositivity and suicidal behavior, and others not finding the association. Conclusions: With the evidence available presently, it is not possible to infer whether there is a correlation between suicide and CMV infection.

Prediction of Suicide and Nonfatal Self-harm After Bariatric Surgery: A Risk Score Based on Sociodemographic Factors, Lifestyle Behavior, and Mental Health: A Nonrandomized Controlled Trial

OBJECTIVE

To identify preoperative sociodemographic and health-related factors that predict higher risk of nonfatal self-harm and suicide after bariatric surgery.

BACKGROUND

Evidence is emerging that bariatric surgery is related to an increased risk of suicide and self-harm, but knowledge on whether certain preoperative characteristics further enhance the excess risk is scarce.

METHODS

The nonrandomized, prospective, controlled Swedish Obese Subjects study was linked to 2 Nationwide Swedish registers. The bariatric surgery group (N = 2007, per-protocol) underwent gastric bypass, banding or vertical banded gastroplasty, and matched controls (N = 2040) received usual care. Participants were recruited from 1987 to 2001, and information on the outcome (a death by suicide or nonfatal self-harm event) was retrieved until the end of 2016. Subhazard ratios (sub-HR) were calculated using competing risk regression analysis.

RESULTS

The risk for self-harm/suicide was almost twice as high in surgical patients compared to control patients both before and after adjusting for various baseline factors [adjusted sub-HR = 1.98, 95% confidence interval (CI) = 1.34-2.93]. Male sex, previous healthcare visits for self-harm or mental disorders, psychiatric drug use, and sleep difficulties predicted higher risk of self-harm/suicide in the multivariate models conducted in the surgery group. Interaction tests further indicated that the excess risk for self-harm/suicide related to bariatric surgery was stronger in men (sub-HR = 3.31, 95% CI = 1.73-6.31) than in women (sub-HR = 1.54, 95% CI = 1.02-2.32) (P = 0.007 for adjusted interaction). A simple-to-use score was developed to identify those at highest risk of these events in the surgery group.

CONCLUSIONS

Our findings suggest that male sex, psychiatric disorder history, and sleep difficulties are important predictors for nonfatal self-harm and suicide in postbariatric patients. High-risk patients who undergo surgery might require regular postoperative psychosocial monitoring to reduce the risk for future self-harm behaviors.

Health outcomes associated with emergency department visits by adolescents for self-harm: a propensity-matched cohort study

BACKGROUND

Self-harm is increasing among adolescents, and because of changing behaviours, current data are needed on the consequences of self-harm. We sought to investigate the trends related to hospital presentation, readmission, patient outcome and medical costs in adolescents who presented with self-harm to the emergency department.

METHODS

We used administrative data on 403 805 adolescents aged 13-17 years presenting to Ontario emergency departments in 2011-2013. Adolescents with self-harm visits were 1:2 propensity matched to controls with visits without self-harm, using demographic, mental health and other clinical variables. Five years after the index presentation, hospital or emergency department admission rates for self-harm, overall mortality, suicides and conservative cost estimates were compared between the 2 groups.

RESULTS

Of 5832 adolescents who visited Ontario emergency departments in 2011-2013 after self-harm (1.4% of visits), 5661 were matched to 10 731 adolescents who presented for reasons other than self-harm. Adolescents who presented with self-harm had a shorter time to a repeat emergency department or hospital admission for self-harm (hazard ratio [HR] 4.84, 95% confidence interval [CI] 4.44-5.27), more suicides (HR 7.96, 95% CI 4.00-15.86), and higher overall mortality (HR 3.23, 95% CI 2.12-4.93; p < 0.001). The positive predictive value of self-harm-related emergency department visits for suicide was 0.7%. Adolescents with self-harm visits had mean 5-year estimates of health care costs of $30 388 compared with $19 055 for controls (p < 0.001).

INTERPRETATION

Adolescents with emergency department visits for self-harm have higher rates of mortality, suicide and recurrent self-harm, as well as higher health care costs, than matched controls. Development of algorithms and interventions that can identify and help adolescents at highest risk of recurrent self-harm is warranted.

Hopelessness in persons with epilepsy: Relationship with demographic, clinical, and social variables

Persons with epilepsy show a higher risk of suicidal ideation and behavior than the general population. Hopelessness, as a feature of demoralization, is considered an emerging risk factor for suicidal ideation. Psychopathological comorbidity, mainly depression, has to be taken into account to predict suicidal attempts but the relationship between suicidality and epilepsy has been also reported independently from depressive disorders. The aim of the study was to investigate hopelessness in a sample of persons suffering from epilepsy without comorbidity with psychiatric disorders and its association with demographic, social, and clinical variables, putatively predictive of suicidal behaviors. Beck Hopelessness Scale (BHS) has been used as measure of suicidal ideation in 72 consecutive outpatients afferent to a third level epilepsy center. Exclusion criterion was psychiatric comorbidity evaluated by clinical approach and quantified by Clinical Global Impression (CGI) Scale. Clinical (focus localization, age at onset, number of drugs), demographic, social variables, and intellectual level were considered. Age, age at onset, gender, intellectual level, socioeconomic status, duration of illness and therapy, number of drugs, seizure frequency, and localization of the epileptic focus and side involved were found associated with BHS score increase. In a sample of persons with epilepsy, without psychiatric comorbidity, our data show an association between factors related to the biological pathophysiology and the clinical course of the disease with the sociodemographic status, as a risk factor for suicidal ideation.

Machine learning and big data analytics in bipolar disorder: A position paper from the International Society for Bipolar Disorders Big Data Task Force

OBJECTIVES

The International Society for Bipolar Disorders Big Data Task Force assembled leading researchers in the field of bipolar disorder (BD), machine learning, and big data with extensive experience to evaluate the rationale of machine learning and big data analytics strategies for BD.

METHOD

A task force was convened to examine and integrate findings from the scientific literature related to machine learning and big data based studies to clarify terminology and to describe challenges and potential applications in the field of BD. We also systematically searched PubMed, Embase, and Web of Science for articles published up to January 2019 that used machine learning in BD.

RESULTS

The results suggested that big data analytics has the potential to provide risk calculators to aid in treatment decisions and predict clinical prognosis, including suicidality, for individual patients. This approach can advance diagnosis by enabling discovery of more relevant data-driven phenotypes, as well as by predicting transition to the disorder in high-risk unaffected subjects. We also discuss the most frequent challenges that big data analytics applications can face, such as heterogeneity, lack of external validation and replication of some studies, cost and non-stationary distribution of the data, and lack of appropriate funding.

CONCLUSION

Machine learning-based studies, including atheoretical data-driven big data approaches, provide an opportunity to more accurately detect those who are at risk, parse-relevant phenotypes as well as inform treatment selection and prognosis. However, several methodological challenges need to be addressed in order to translate research findings to clinical settings.

Real-Time Monitoring of Suicide Risk among Adolescents: Potential Barriers, Possible Solutions, and Future Directions

Recent advances in real-time monitoring technology make this an exciting time to study risk for suicidal thoughts and behaviors among youth. Although there is good reason to be excited about these methods, there is also reason for caution in adopting them without first understanding their limitations. In this article, we present several broad future directions for using real-time monitoring among youth at risk for suicide focused around three broad themes: novel research questions, novel analytic methods, and novel methodological approaches. We also highlight potential technical, logistical, and ethical challenges with these methodologies, as well as possible solutions to these challenges.

Suicide and suicide risk

Although recent years have seen large decreases in the overall global rate of suicide fatalities, this trend is not reflected everywhere. Suicide and suicidal behaviour continue to present key challenges for public policy and health services, with increasing suicide deaths in some countries such as the USA. The development of suicide risk is complex, involving contributions from biological (including genetics), psychological (such as certain personality traits), clinical (such as comorbid psychiatric illness), social and environmental factors. The involvement of multiple risk factors in conveying risk of suicide means that determining an individual's risk of suicide is challenging. Improving risk assessment, for example, by using computer testing and genetic screening, is an area of ongoing research. Prevention is key to reduce the number of suicide deaths and prevention efforts include universal, selective and indicated interventions, although these interventions are often delivered in combination. These interventions, combined with psychological (such as cognitive behavioural therapy, caring contacts and safety planning) and pharmacological treatments (for example, clozapine and ketamine) along with coordinated social and public health initiatives, should continue to improve the management of individuals who are suicidal and decrease suicide-associated morbidity.

Predicting 3-month risk for adolescent suicide attempts among pediatric emergency department patients

BACKGROUND

The incidence of adolescent suicide is rising in the United States, yet we have limited information regarding short-term prediction of suicide attempts. Our aim was to identify predictors of suicide attempts within 3-months of an emergency department (ED) visit.

METHODS

Adolescents, ages 12-17, seeking health care at 13 pediatric EDs (Pediatric Emergency Care Applied Research Network) and one Indian Health Service Hospital in the United States were consecutively recruited. Among 10,664 approached patients, 6,448 (60%) were enrolled and completed a suicide risk survey. A subset of participants (n = 2,897) was assigned to a 3-month telephone follow-up, and 2,104 participants completed this follow-up (73% retention). Our primary outcome was a suicide attempt between the ED visit and 3-month follow-up.

RESULTS

One hundred four adolescents (4.9%) made a suicide attempt between enrollment and 3-month follow-up. A large number of baseline predictors of suicide attempt were identified in bivariate analyses. The final multivariable model for the full sample included the presence of suicidal ideation during the past week, lifetime severity of suicidal ideation, lifetime history of suicidal behavior, and school connectedness. For the subgroup of adolescents who did not report recent suicidal ideation at baseline, the final model included only lifetime severity of suicidal ideation and social connectedness. Among males, the final model included only lifetime severity of suicidal ideation and past week suicidal ideation. For females, the final model included past week suicidal ideation, lifetime severity of suicidal ideation, number of past-year nonsuicidal self-injury (NSSI) incidents, and social connectedness.

CONCLUSIONS

Results indicate that the key risk factors for adolescent suicide attempts differ for subgroups of adolescents defined by sex and whether or not they report recent suicidal thoughts. Results also point to the importance of school and social connectedness as protective factors against suicide attempts.

Establishing an Evidence Synthesis Capability For Psychological Health Topics in the Military Health System

BACKGROUND

To promote evidence-based health care, clinical providers and decision makers rely on scientific evidence to inform best practices. Evidence synthesis (ES) is a key component of this process that serves to inform health care decisions by integrating and contextualizing research findings across studies.

OBJECTIVE

This paper describes the process of establishing an ES capability in the Military Health System dedicated to psychological health topics.

RESEARCH DESIGNS

The goal of establishing the current ES capability was to facilitate evidence-based decision-making among clinicians, clinic managers, research funders, and policymakers, through the production and dissemination of trustworthy ES reports. We describe how we developed this capability, provide an overview of the types of evidence syntheses products we use to respond to different stakeholders, and detail the procedures established for selecting and prioritizing synthesis topics.

RESULTS

We report on the productivity, acceptability, and impact of our efforts. Our reports were used by a variety of stakeholders and working groups, briefed to major committees, included in official reports and policies, and cited in clinical practice guidelines and the peer-reviewed literature.

CONCLUSIONS

Our experiences thus far suggest that the current ES capability offers a needed service within our health system. Our framework may help inform other agencies interested in developing or sponsoring a similar capability.

Fatal and Non-fatal Self-Injury in the USA: Critical Review of Current Trends and Innovations in Prevention

PURPOSE OF REVIEW

To examine current trends in suicide and self-injury in the USA, as well as potential contributors to their change over time, and to reflect on innovations in prevention and intervention that can guide policies and programs to reduce the burden of suicide and self-injury in the USA.

RECENT FINDINGS

Suicide and non-fatal self-injury are on the rise in the USA. Reasons for such trends over time remain speculative, although they seem linked to coincident increases in mood disorders and drug use and overdose. Promising innovative prevention and intervention programs that engage new technologies, such as machine learning-derived prediction tools and computerized ecologic momentary assessments, are currently in development and require additional evidence. Recent increases in fatal and non-fatal self-harm in the USA raise questions about the causes, interventions, and preventive measures that should be taken. Most innovative prevention efforts target individuals seeking to improve risk prediction and access to evidence-based care. However, as Durkheim pointed out over 100 years ago, suicide rates vary enormously between societal groups, suggesting that certain causal factors of suicide act and, hence, should be targeted at an ecological level. In the next generation of suicide research, it is critical to examine factors beyond the proximal and clinical to allow for a reimagining of prevention that is life course and socially focused.

Building the case for actionable ethics in digital health research supported by artificial intelligence

The digital revolution is disrupting the ways in which health research is conducted, and subsequently, changing healthcare. Direct-to-consumer wellness products and mobile apps, pervasive sensor technologies and access to social network data offer exciting opportunities for researchers to passively observe and/or track patients 'in the wild' and 24/7. The volume of granular personal health data gathered using these technologies is unprecedented, and is increasingly leveraged to inform personalized health promotion and disease treatment interventions. The use of artificial intelligence in the health sector is also increasing. Although rich with potential, the digital health ecosystem presents new ethical challenges for those making decisions about the selection, testing, implementation and evaluation of technologies for use in healthcare. As the 'Wild West' of digital health research unfolds, it is important to recognize who is involved, and identify how each party can and should take responsibility to advance the ethical practices of this work. While not a comprehensive review, we describe the landscape, identify gaps to be addressed, and offer recommendations as to how stakeholders can and should take responsibility to advance socially responsible digital health research.

How accurate are suicide risk prediction models? Asking the right questions for clinical practice

Prediction models assist in stratifying and quantifying an individual's risk of developing a particular adverse outcome, and are widely used in cardiovascular and cancer medicine. Whether these approaches are accurate in predicting self-harm and suicide has been questioned. We searched for systematic reviews in the suicide risk assessment field, and identified three recent reviews that have examined current tools and models derived using machine learning approaches. In this clinical review, we present a critical appraisal of these reviews, and highlight three major limitations that are shared between them. First, structured tools are not compared with unstructured assessments routine in clinical practice. Second, they do not sufficiently consider a range of performance measures, including negative predictive value and calibration. Third, the potential role of these models as clinical adjuncts is not taken into consideration. We conclude by presenting the view that the current role of prediction models for self-harm and suicide is currently not known, and discuss some methodological issues and implications of some machine learning and other analytic techniques for clinical utility.