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

Moving Toward Meaningful Evaluations of Monitoring in e-Mental Health Based on the Case of a Web-Based Grief Service for Older Mourners: Mixed Methods Study

Background

Artificial intelligence (AI) tools hold much promise for mental health care by increasing the scalability and accessibility of care. However, current development and evaluation practices of AI tools limit their meaningfulness for health care contexts and therefore also the practical usefulness of such tools for professionals and clients alike.

Objective

The aim of this study is to demonstrate the evaluation of an AI monitoring tool that detects the need for more intensive care in a web-based grief intervention for older mourners who have lost their spouse, with the goal of moving toward meaningful evaluation of AI tools in e-mental health.

Methods

We leveraged the insights from three evaluation approaches: (1) the F1-score evaluated the tool's capacity to classify user monitoring parameters as either in need of more intensive support or recommendable to continue using the web-based grief intervention as is; (2) we used linear regression to assess the predictive value of users' monitoring parameters for clinical changes in grief, depression, and loneliness over the course of a 10-week intervention; and (3) we collected qualitative experience data from e-coaches (N=4) who incorporated the monitoring in their weekly email guidance during the 10-week intervention.

Results

Based on n=174 binary recommendation decisions, the F1-score of the monitoring tool was 0.91. Due to minimal change in depression and loneliness scores after the 10-week intervention, only 1 linear regression was conducted. The difference score in grief before and after the intervention was included as a dependent variable. Participants' (N=21) mean score on the self-report monitoring and the estimated slope of individually fitted growth curves and its standard error (ie, participants' response pattern to the monitoring questions) were used as predictors. Only the mean monitoring score exhibited predictive value for the observed change in grief (R2=1.19, SE 0.33; t16=3.58, P=.002). The e-coaches appreciated the monitoring tool as an opportunity to confirm their initial impression about intervention participants, personalize their email guidance, and detect when participants' mental health deteriorated during the intervention.

Conclusions

The monitoring tool evaluated in this paper identified a need for more intensive support reasonably well in a nonclinical sample of older mourners, had some predictive value for the change in grief symptoms during a 10-week intervention, and was appreciated as an additional source of mental health information by e-coaches who supported mourners during the intervention. Each evaluation approach in this paper came with its own set of limitations, including (1) skewed class distributions in prediction tasks based on real-life health data and (2) choosing meaningful statistical analyses based on clinical trial designs that are not targeted at evaluating AI tools. However, combining multiple evaluation methods facilitates drawing meaningful conclusions about the clinical value of AI monitoring tools for their intended mental health context.

Automatically extracting social determinants of health for suicide: a narrative literature review

Suicide is a complex phenomenon that is often not preceded by a diagnosed mental health condition, therefore making it difficult to study and mitigate. Artificial Intelligence has increasingly been used to better understand Social Determinants of Health factors that influence suicide outcomes. In this review we find that many studies use limited SDoH information and minority groups are often underrepresented, thereby omitting important factors that could influence risk of suicide.

Sexual and Gender Minority Status and Suicide Mortality: An Explainable Artificial Intelligence Analysis

Objectives: Suicide risk is elevated in lesbian, gay, bisexual, and transgender (LGBT) individuals. Limited data on LGBT status in healthcare systems hinder our understanding of this risk. This study used natural language processing to extract LGBT status and a deep neural network (DNN) to examine suicidal death risk factors among US Veterans. Methods: Data on 8.8 million veterans with visits between 2010 and 2017 was used. A case-control study was performed, and suicide death risk was analyzed by a DNN. Feature impacts and interactions on the outcome were evaluated. Results: The crude suicide mortality rate was higher in LGBT patients. However, after adjusting for over 200 risk and protective factors, known LGBT status was associated with reduced risk compared to LGBT-Unknown status. Among LGBT patients, black, female, married, and older Veterans have a higher risk, while Veterans of various religions have a lower risk. Conclusion: Our results suggest that disclosed LGBT status is not directly associated with an increase suicide death risk, however, other factors (e.g., depression and anxiety caused by stigma) are associated with suicide death risks.

Proximal characteristics of suicide attempts: a study in a public hospital in Spain

INTRODUCTION

Different parameters of suicide attempts treated since the implementation of the Attention to Suicide Risk Program (ARSUIC) in 2012 at the Hospital Ramón y Cajal in Madrid Region are described in this paper.

METHOD

The sample was composed of 107 patients and the information was collected through a questionnaire created ad hoc with the following variables: type of suicidal ideation; drug use immediately prior to the attempt; method (in case of drug overdosing: drug/s used); location; accessibility to rescue; planning; intentionality; criticism; and brakes.

RESULTS

Descriptive statistics were obtained and a comparison by gender was made through the χ2 and contingency coefficients tests. The data from the retrospective longitudinal study showed that the most common profile was of patients with unstructured ideas of death and no previous drug use who took an unplanned drug overdose in the family home, with the intention of self-harm or avoidance of discomfort, especially with benzodiazepines. Patients tend to ask for help afterwards and criticise the attempt, but potential restraints are often not recorded in the clinical report. Regarding the dissimilarities based on gender, statistically significant differences were found in prior alcohol consumption, in favour of men and in the overdose method, specifically with benzodiazepines, in favour of women.

CONCLUSIONS

Knowing the types of attempts at self-harm is essential for improving prevention, understanding and patient management.

Development of mental health first-aid guidelines for suicide risk: a Delphi expert consensus study in Argentina and Chile

BACKGROUND

Suicide continues to pose a significant global public health challenge and ranks as one of the leading causes of death worldwide. Given the prevalence of suicide risk in the community, there is a significant likelihood of encountering individuals who may be experiencing suicidal thoughts or plans, creating an opening for non-health professionals to offer support. This study aims to culturally adapt the original Australian Mental Health First Aid Guidelines for suicide risk to the Chilean and Argentine context.

METHODS

A two-round Delphi expert consensus study was conducted involving two panels, one comprising individuals with personal experience in suicide thoughts/attempts or caregiving for those with such experiences (n = 18), and the other consisting of professionals specialized in suicide assessment and support for individuals at risk (n = 25). They rated a total of 179 items mainly derived from guidelines developed by Australian experts and translated into Spanish (168), and new items included by the research team (11). The panel members were requested to assess each item utilizing a five-point Likert scale. During the second round, items that received moderate approval in the initial round were re-evaluated, and new items suggested by the local experts in the first round were also subjected to evaluation in the next round. Inclusion in the final guidelines required an 80% endorsement as "essential" or "important" from both panels.

RESULTS

Consensus of approval was reached for 189 statements. Among these, 139 statements were derived from the English-language guidelines, while 50 locally generated statements were accepted during the second round. A significant difference from the original guideline was identified concerning the local experts' reluctance to discuss actions collaboratively with adolescents. Furthermore, the local experts proposed the inclusion of an entirely new section addressing suicide risk in older individuals, particularly focusing on suicide methods and warning signs.

CONCLUSIONS

A Delphi expert consensus study was conducted to culturally adapt mental health first aid guidelines for assessing suicide risk in Chile and Argentina. This study involved professionals and individuals with lived experience. While many items were endorsed, some related to inquiring about suicide risk and autonomy, particularly for adolescents, were not. An additional section for older individuals was introduced. Future research should explore the implementation and impact of these adapted guidelines in training courses. This is vital for enhancing mental health support and implementing effective suicide prevention strategies in Chile and Argentina.

Predicting risk of suicidal behaviour after initiation of selective serotonin reuptake inhibitors in children, adolescents and young adults: protocol for development and validation of clinical prediction models

INTRODUCTION

There is concern regarding suicidal behaviour risk during selective serotonin reuptake inhibitor (SSRI) treatment among the young. A clinically useful model for predicting suicidal behaviour risk should have high predictive performance in terms of discrimination and calibration; transparency and ease of implementation are desirable.

METHODS AND ANALYSIS

Using Swedish national registers, we will identify individuals initiating an SSRI aged 8-24 years 2007-2020. We will develop: (A) a model based on a broad set of predictors, and (B) a model based on a restricted set of predictors. For the broad predictor model, we will consider an ensemble of four base models: XGBoost (XG), neural net (NN), elastic net logistic regression (EN) and support vector machine (SVM). The predictors with the greatest contribution to predictive performance in the base models will be determined. For the restricted predictor model, clinical input will be used to select predictors based on the top predictors in the broad model, and inputted in each of the XG, NN, EN and SVM models. If any show superiority in predictive performance as defined by the area under the receiver-operator curve, this model will be selected as the final model; otherwise, the EN model will be selected. The training and testing samples will consist of data from 2007 to 2017 and from 2018 to 2020, respectively. We will additionally assess the final model performance in individuals receiving a depression diagnosis within 90 days before SSRI initiation.The aims are to (A) develop a model predicting suicidal behaviour risk after SSRI initiation among children and youths, using machine learning methods, and (B) develop a model with a restricted set of predictors, favouring transparency and scalability.

ETHICS AND DISSEMINATION

The research is approved by the Swedish Ethical Review Authority (2020-06540). We will disseminate findings by publishing in peer-reviewed open-access journals, and presenting at international conferences.

Predicting HIV infection in the decade (2005-2015) pre-COVID-19 in Zimbabwe: A supervised classification-based machine learning approach

The burden of HIV and related diseases have been areas of great concern pre and post the emergence of COVID-19 in Zimbabwe. Machine learning models have been used to predict the risk of diseases, including HIV accurately. Therefore, this paper aimed to determine common risk factors of HIV positivity in Zimbabwe between the decade 2005 to 2015. The data were from three two staged population five-yearly surveys conducted between 2005 and 2015. The outcome variable was HIV status. The prediction model was fit by adopting 80% of the data for learning/training and 20% for testing/prediction. Resampling was done using the stratified 5-fold cross-validation procedure repeatedly. Feature selection was done using Lasso regression, and the best combination of selected features was determined using Sequential Forward Floating Selection. We compared six algorithms in both sexes based on the F1 score, which is the harmonic mean of precision and recall. The overall HIV prevalence for the combined dataset was 22.5% and 15.3% for females and males, respectively. The best-performing algorithm to identify individuals with a higher likelihood of HIV infection was XGBoost, with a high F1 score of 91.4% for males and 90.1% for females based on the combined surveys. The results from the prediction model identified six common features associated with HIV, with total number of lifetime sexual partners and cohabitation duration being the most influential variables for females and males, respectively. In addition to other risk reduction techniques, machine learning may aid in identifying those who might require Pre-exposure prophylaxis, particularly women who experience intimate partner violence. Furthermore, compared to traditional statistical approaches, machine learning uncovered patterns in predicting HIV infection with comparatively reduced uncertainty and, therefore, crucial for effective decision-making.

Risk of death by suicide following self-harm presentations to healthcare: development and validation of a multivariable clinical prediction rule (OxSATS)

BACKGROUND

Assessment of suicide risk in individuals who have self-harmed is common in emergency departments, but is often based on tools developed for other purposes.

OBJECTIVE

We developed and validated a predictive model for suicide following self-harm.

METHODS

We used data from Swedish population-based registers. A cohort of 53 172 individuals aged 10+ years, with healthcare episodes of self-harm, was split into development (37 523 individuals, of whom 391 died from suicide within 12 months) and validation (15 649 individuals, 178 suicides within 12 months) samples. We fitted a multivariable accelerated failure time model for the association between risk factors and time to suicide. The final model contains 11 factors: age, sex, and variables related to substance misuse, mental health and treatment, and history of self-harm. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis guidelines were followed for the design and reporting of this work.

FINDINGS

An 11-item risk model to predict suicide was developed using sociodemographic and clinical risk factors, and showed good discrimination (c-index 0.77, 95% CI 0.75 to 0.78) and calibration in external validation. For risk of suicide within 12 months, using a 1% cut-off, sensitivity was 82% (75% to 87%) and specificity was 54% (53% to 55%). A web-based risk calculator is available (Oxford Suicide Assessment Tool for Self-harm or OxSATS).

CONCLUSIONS

OxSATS accurately predicts 12-month risk of suicide. Further validations and linkage to effective interventions are required to examine clinical utility.

CLINICAL IMPLICATIONS

Using a clinical prediction score may assist clinical decision-making and resource allocation.

A Stakeholder-Informed Ethical Framework to Guide Implementation of Suicide Risk Prediction Models Derived from Electronic Health Records

OBJECTIVE

Develop a stakeholder-informed ethical framework to provide practical guidance to health systems considering implementation of suicide risk prediction models.

METHODS

In this multi-method study, patients and family members participating in formative focus groups (n = 4 focus groups, 23 participants), patient advisors, and a bioethics consultant collectively informed the development of a web-based survey; survey results (n = 1,357 respondents) and themes from interviews with stakeholders (patients, health system administrators, clinicians, suicide risk model developers, and a bioethicist) were used to draft the ethical framework.

RESULTS

Clinical, ethical, operational, and technical issues reiterated by multiple stakeholder groups and corresponding questions for risk prediction model adopters to consider prior to and during suicide risk model implementation are organized within six ethical principles in the resulting stakeholder-informed framework. Key themes include: patients' rights to informed consent and choice to conceal or reveal risk (autonomy); appropriate application of risk models, data and model limitations and consequences including ambiguous risk predictors in opaque models (explainability); selecting actionable risk thresholds (beneficence, distributive justice); access to risk information and stigma (privacy); unanticipated harms (non-maleficence); and planning for expertise and resources to continuously audit models, monitor harms, and redress grievances (stewardship).

CONCLUSIONS

Enthusiasm for risk prediction in the context of suicide is understandable given the escalating suicide rate in the U.S. Attention to ethical and practical concerns in advance of automated suicide risk prediction model implementation may help avoid unnecessary harms that could thwart the promise of this innovation in suicide prevention. HIGHLIGHTSPatients' desire to consent/opt out of suicide risk prediction models.Recursive ethical questioning should occur throughout risk model implementation.Risk modeling resources are needed to continuously audit models and monitor harms.

Machine learning based identification of structural brain alterations underlying suicide risk in adolescents

Suicide is the third leading cause of death for individuals between 15 and 19 years of age. The high suicide mortality rate and limited prior success in identifying neuroimaging biomarkers indicate that it is crucial to improve the accuracy of clinical neural signatures underlying suicide risk. The current study implements machine-learning (ML) algorithms to examine structural brain alterations in adolescents that can discriminate individuals with suicide risk from typically developing (TD) adolescents at the individual level. Structural MRI data were collected from 79 adolescents who demonstrated clinical levels of suicide risk and 79 demographically matched TD adolescents. Region-specific cortical/subcortical volume (CV/SCV) was evaluated following whole-brain parcellation into 1000 cortical and 12 subcortical regions. CV/SCV parameters were used as inputs for feature selection and three ML algorithms (i.e., support vector machine [SVM], K-nearest neighbors, and ensemble) to classify adolescents at suicide risk from TD adolescents. The highest classification accuracy of 74.79% (with sensitivity = 75.90%, specificity = 74.07%, and area under the receiver operating characteristic curve = 87.18%) was obtained for CV/SCV data using the SVM classifier. Identified bilateral regions that contributed to the classification mainly included reduced CV within the frontal and temporal cortices but increased volume within the cuneus/precuneus for adolescents at suicide risk relative to TD adolescents. The current data demonstrate an unbiased region-specific ML framework to effectively assess the structural biomarkers of suicide risk. Future studies with larger sample sizes and the inclusion of clinical controls and independent validation data sets are needed to confirm our findings.

Psychotic depression and deaths due to suicide

BACKGROUND

The purpose of this study was to establish the risk of suicide associated with incident psychotic depression (PD) compared to incident non-psychotic severe depression (NPD).

METHODS

This cohort study used routine data from nationwide health registers in Finland. Eligible participants were aged 18-59 years at the index diagnosis. Causes of death were defined by the International Classification of Diseases, 10th revision codes. The follow-up time was up to five years. Adjusted Cox regression models were used to analyse risk of death by method of suicide.

RESULTS

We included 17,331 individuals with incident PD and 85,989 individuals with incident NPD. Most of the deaths due to suicides occurred within the first two years after the index diagnosis. Compared to NPD, PD was associated with an overall two-fold increased risk of suicide (adjusted hazard ratio, (aHR) 2.19, 95 % confidence interval (CI) 1.95, 2.46), after adjusting for psychiatric comorbidities. In PD, the highest relative risks were for impact-related suicides (aHR 3.03, 95%CI 2.23, 4.13) and for suffocation-related suicides (aHR 2.72, 95%CI 2.23, 3.30), whereas the lowest relative risk was for intentional poisonings (aHR 1.66, 95%CI 1.37, 2.02).

LIMITATIONS

Information on all potential confounders is not available in studies using routine data.

CONCLUSIONS

Psychotic symptoms doubled the risk of suicides over and above of the risk that was associated with severe depression, after controlling for comorbid psychiatric disorders. The severity of suicidal ideation may be higher in PD than in NPD, which then leads to more lethal methods of self-harm.

Digital cognitive-behavioural therapy to reduce suicidal ideation and behaviours: a systematic review and meta-analysis of individual participant data

QUESTION

Digital interventions based on cognitive-behavioural therapy (iCBT) is associated with reductions in suicidal ideation. However, fine-grained analyses of effects and potential effect-moderating variables are missing. This study aimed to investigate the effectiveness of iCBT on suicidal ideation, effect moderators, effects on suicide attempts and predictors of adherence.

STUDY SELECTION AND ANALYSIS

We systematically searched CENTRAL, PsycINFO, Embase and PubMed for randomised controlled trials that investigated iCBT for suicidal ideation or behaviours. Participants reporting baseline suicidal ideation were eligible. We conducted a one-stage individual participant data (IPD) meta-analysis. Suicidal ideation was the primary outcome, analysed as three indices: severity of suicidal ideation, reliable changes and treatment response.

FINDINGS

We included IPD from nine out of ten eligible trials (2037 participants). iCBT showed significant reductions of suicidal ideation compared with control conditions across all indices (severity: b=-0.247, 95% CI -0.322 to -0.173; reliable changes: b=0.633, 95% CI 0.408 to 0.859; treatment response: b=0.606, 95% CI 0.410 to 0.801). In iCBT, the rate of reliable improvement was 40.5% (controls: 27.3%); the deterioration rate was 2.8% (controls: 5.1%). No participant-level moderator effects were identified. The effects on treatment response were higher for trials with waitlist-controls compared with active controls. There were insufficient data on suicide attempts. Human support and female gender predicted treatment adherence. The main source of potential bias was missing outcome data.

CONCLUSIONS

The current evidence indicates that iCBT is effective in reducing suicidal ideation irrespective of age, gender and previous suicide attempts. Future studies should rigorously assess suicidal behaviour and drop-out reasons.

A genetically informed prediction model for suicidal and aggressive behaviour in teens

Suicidal and aggressive behaviours cause significant personal and societal burden. As risk factors associated with these behaviours frequently overlap, combined approaches in predicting the behaviours may be useful in identifying those at risk for either. The current study aimed to create a model that predicted if individuals will exhibit suicidal behaviour, aggressive behaviour, both, or neither in late adolescence. A sample of 5,974 twins from the Child and Adolescent Twin Study in Sweden (CATSS) was broken down into a training (80%), tune (10%) and test (10%) set. The Netherlands Twin Register (NTR; N = 2702) was used for external validation. Our longitudinal data featured genetic, environmental, and psychosocial predictors derived from parental and self-report data. A stacked ensemble model was created which contained a gradient boosted machine, random forest, elastic net, and neural network. Model performance was transferable between CATSS and NTR (macro area under the receiver operating characteristic curve (AUC) [95% CI] AUCCATSS(test set) = 0.709 (0.671-0.747); AUCNTR = 0.685 (0.656-0.715), suggesting model generalisability across Northern Europe. The notable exception is suicidal behaviours in the NTR, which was no better than chance. The 25 highest scoring variable importance scores for the gradient boosted machines and random forest models included self-reported psychiatric symptoms in mid-adolescence, sex, and polygenic scores for psychiatric traits. The model's performance is comparable to current prediction models that use clinical interviews and is not yet suitable for clinical use. Moreover, genetic variables may have a role to play in predictive models of adolescent psychopathology.

Assessment of suicide risk in mental health practice: shifting from prediction to therapeutic assessment, formulation, and risk management

Suicide prevention in psychiatric practice has been dominated by efforts to predict risk of suicide in individual patients. However, traditional risk prediction measures have been shown repeatedly in studies from high income countries to be ineffective. Several factors might contribute to clinicians' preoccupation with risk prediction, which can have negative effects on patient care and also on clinicians where prediction is seen as failing. The model of therapeutic risk assessment, formulation, and management we outline in this article regards all patients with mental health problems as potentially at increased risk of suicide. It is aimed at reducing risk through use of a person-centred approach. We describe how a move towards therapeutic risk assessment, formulation, and risk management, including collaborative safety planning, could help clinicians develop a more tailored approach to managing risk for all patients, incorporating potentially therapeutic effects as well as helping to identify other risk reduction interventions. Such an approach could lead to enhanced patient safety and quality of care, which is more acceptable to patients.

Machine Learning Prediction of Suicide Risk Does Not Identify Patients Without Traditional Risk Factors

Objective: To determine whether predictions of suicide risk from machine learning models identify unexpected patients or patients without medical record documentation of traditional risk factors.

Methods: The study sample included 27,091,382 outpatient mental health (MH) specialty or general medical visits with a MH diagnosis for patients aged 11 years or older from January 1, 2009, to September 30, 2017. We used predicted risk scores of suicide attempt and suicide death, separately, within 90 days of visits to classify visits into risk score percentile strata. For each stratum, we calculated counts and percentages of visits with traditional risk factors, including prior self-harm diagnoses and emergency department visits or hospitalizations with MH diagnoses, in the last 3, 12, and 60 months.

Results: Risk-factor percentages increased with predicted risk scores. Among MH specialty visits, 66%, 88%, and 99% of visits with suicide attempt risk scores in the top 3 strata (respectively, 90th-95th, 95th-98th, and ≥ 98th percentiles) and 60%, 77%, and 93% of visits with suicide risk scores in the top 3 strata represented patients who had at least one traditional risk factor documented in the prior 12 months. Among general medical visits, 52%, 66%, and 90% of visits with suicide attempt risk scores in the top 3 strata and 45%, 66%, and 79% of visits with suicide risk scores in the top 3 strata represented patients who had a history of traditional risk factors in the last 12 months.

Conclusions: Suicide risk alerts based on these machine learning models coincide with patients traditionally thought of as high-risk at their high-risk visits.

Military mental health professionals' suicide risk assessment and management before and after experiencing a patient's suicide

OBJECTIVE

This study examines the association between a patient's suicide and the therapist's suicide risk assessment (SRA) and suicide risk management (SRM) of patients, following the occurrence.

METHOD

SRA values range from "absence of suicidality" to "immediate suicidal intent to die". SRM consists of therapists' written recommendations. Rates of the various SRA and SRM values in therapists' evaluations were assessed 6-months prior to the suicide and at the two three- and six-month time-points thereafter.

RESULTS

Of the 150 soldiers who died by suicides, 30 (20%) visited 50 military therapists in the 6 months preceding their deaths. Using Wilcoxon signed rank test, lower SRA rates of "threatens suicide" were found 2 months after a patient's suicide. Regarding SRM, the mean rates for "recommendations for psychotherapy treatment" were higher at the two (p = 0.022) and the 3 month time-points (p = 0.031) after a suicide.

CONCLUSIONS

The SRA findings may indicate therapists' fear of treating suicidal patients, causing them to overlook patients' non-prominent suicide-risk indicators. In SRM, the higher rate of recommendations for additional therapy sessions rather than military release or referrals to other therapists may relate to over-caution and attempts to control the patient's therapy ensuring it's done properly.

Controversies In The Pharmacotherapy Of Adolescent Depression

BACKGROUND

Although fluoxetine and, in the USA, escitalopram are approved for depression in adolescence, substantial concern surrounds antidepressant use in youth. Major controversies regarding efficacy and safety (increased suicidality).

INTRODUCTION

The cathegory of depression is very broad and overinclusive, in terms of etiology, role of psychosocial adversities severity, episodicity, presentation, relationship with bipolarity. This heterogeneity, not fully controlled considered in Randomized Controlled Trials (RCTs), may account for the disappointing results on both efficacy and safety.

METHOD

Based on the available literature, we will address the following topics: a) controversies regarding the definition of depression as a unique homogeneous condition with a unique type of pharmacological treatment; b) controversies about the interpretation of data from Randomized Controlled Trials (RCTs) on the efficacy of pharmacological treatments in adolescent depression; c) the interpretation of data regarding the safety of antidepressant treatment in adolescent depression, particularly in terms of increased suicidal risk.

RESULTS

According to RCTs, antidepressants are minimally to moderately more effective than placebo, principally based on very high placebo responses, and only fluoxetine showed more evidence of efficacy. These differences in meta-analyses are sometimes statistically, but not clinically significant. Depression is a heterogeneous condition in terms of etiology, role of psychosocial adversities severity, episodicity, presentation, relationship with bipolarity. This heterogeneity may partly explain the low drug-placebo difference and the high placebo response (possibly related to a high level of natural recovery of the adolescent depression). In the National Institute of Mental Health (NIMH)-funded studies, including a lower number of study sites and more reliable enrollment procedures, lower placebo response rates and greater group differences between medication and placebo were found. Robust evidence supports an increased risk of emergent suicidality after starting antidepressants. A clear age effect on suicidal risk after antidepressants is supported by a comprehensive meta-analysis, showing that suicidal risk increased with decreasing age, being markedly greater in subjects aged between 18 and 25 years. However, the term suicidality is too broad, as it includes suicidal ideation, suicidal attempts, and completed suicide, with a hugely wide range of severity and pervasiveness. If emergent suicidality should be actively and carefully explored, empirical evidence, albeit weak, suggests that combined pharmacotherapy (antidepressant and/or lithium) associated with psychotherapy may be helpful in reducing pretreatment suicidal ideation and suicidal risk.

DISCUSSION

Moderate to severe depression should be treated with psychotherapy and/or fluoxetine, the best-supported medication, and treatment-resistant adolescents should always receive combined treatment with psychotherapy. Suicidal ideation, particularly with a plan, should be actively explored before starting an antidepressant, as a reason for the closest monitoring. Emergent suicidality after starting antidepressants, as well as antidepressant-related activation, should also be closely monitored and may lead to antidepressant discontinuation. Although no response to pharmacotherapy and psychotherapy may occur in up to 40% of depressed adolescents, possible predictors or mediators of poorer response in adolescents are uncertain, and only a few studies support possible treatment strategies. Finally, studies exploring the efficacy of antidepressants in specific depression subtypes, i.e., based on prevalent psychopathological dimensions (apathy, withdrawal, impulsivity), are warranted.

Risk factor analysis and nomogram construction for predicting suicidal ideation in patients with cancer

BACKGROUND

Suicidal ideation in cancer patients is a critical challenge. At present, few studies focus on factors associated with suicidal ideation, and predictive models are still lacking. This study aimed at investigating the risk factors for suicidal ideation among cancer patients, and developed a predictive nomogram to screen high risk cancer patients for early prevention and intervention.

METHODS

A questionnaire survey was conducted among cancer patients between May 2021 and January 2022. The factors associated with suicidal ideation were used to construct a multivariate logistic regression model, which was visualized as a predictive nomogram to evaluate the risk of suicidal ideation. Areas under the curve, calibration plot, decision curve analysis, and internal and external validation were used to validate the discrimination, calibration and clinical usefulness of the model.

RESULTS

A total of 820 patients with cancer were recruited for this study and 213 (25.98%) developed suicidal ideation. Levels of demoralization, depression and cancer staging, marital status, residence, medical financial burden, and living condition were influence factors for suicidal ideation. Comparing nomogram with Self-rating Idea of Suicide Scale (SIOSS), the nomogram had a satisfactory discrimination ability with an AUC of 0.859 (95% CI: 0.827-0.890) and 0.818 (95% CI: 0.764-0.873) in the training and validation sets, respectively. The calibration plot and decision curve analysis revealed that this nomogram was in good fitness and could be beneficial in clinical applications.

CONCLUSIONS

Suicidal ideation is common in cancer patients. Levels of demoralization, depression and cancer staging were independent predictors of suicidal ideation. The nomogram is an effective and simple tool for predictive suicidal ideation in cancer patients.

Classification of suicidal thoughts and behaviour in children: results from penalised logistic regression analyses in the Adolescent Brain Cognitive Development study

BACKGROUND

Despite efforts to predict suicide risk in children, the ability to reliably identify who will engage in suicide thoughts or behaviours has remained unsuccessful.

AIMS

We apply a novel machine-learning approach and examine whether children with suicide thoughts or behaviours could be differentiated from children without suicide thoughts or behaviours based on a combination of traditional (sociodemographic, physical health, social-environmental, clinical psychiatric) risk factors, but also more novel risk factors (cognitive, neuroimaging and genetic characteristics).

METHOD

The study included 5885 unrelated children (50% female, 67% White, 9-11 years of age) from the Adolescent Brain Cognitive Development (ABCD) study. We performed penalised logistic regression analysis to distinguish between: (a) children with current or past suicide thoughts or behaviours; (b) children with a mental illness but no suicide thoughts or behaviours (clinical controls); and (c) healthy control children (no suicide thoughts or behaviours and no history of mental illness). The model was subsequently validated with data from seven independent sites involved in the ABCD study (n = 1712).

RESULTS

Our results showed that we were able to distinguish the suicide thoughts or behaviours group from healthy controls (area under the receiver operating characteristics curve: 0.80 child-report, 0.81 for parent-report) and clinical controls (0.71 child-report and 0.76-0.77 parent-report). However, we could not distinguish children with suicidal ideation from those who attempted suicide (AUROC: 0.55-0.58 child-report; 0.49-0.53 parent-report). The factors that differentiated the suicide thoughts or behaviours group from the clinical control group included family conflict, prodromal psychosis symptoms, impulsivity, depression severity and history of mental health treatment.

CONCLUSIONS

This work highlights that mostly clinical psychiatric factors were able to distinguish children with suicide thoughts or behaviours from children without suicide thoughts or behaviours. Future research is needed to determine if these variables prospectively predict subsequent suicidal behaviour.

Local implementation evaluation of a suicide prevention predictive model at a large VA health care system

BACKGROUND

The Veterans Health Administration (VHA) implemented REACH VET, which analyzes health records to identify veterans at statistically elevated risk for suicide and other adverse outcomes compared to other veterans in VHA. This project evaluated REACH VET program implementation at a large VA health care system by examining program fidelity and treatment engagement, receipt of suicide prevention interventions, and suicide-related behaviors in the 6 months following identification.

METHODS

Over a 12-month period, 218 unique cases were identified by REACH VET within a local VA system. Data were extracted from the VA's electronic medical records.

RESULTS

Protocol adherence for required clinical and administrative steps was 94% and above. After identification, 88% received outpatient mental health treatment, 21% had a psychiatric hospitalization, and 83% engaged in Safety Planning around the time of identification or in the following six months. Twenty-six percent of cases were identified by another existing method for identifying high-risk veterans. Five percent had a medically documented suicide attempt, and none were known to die by suicide in the following 6 months.

CONCLUSIONS

Local evaluation suggested high protocol fidelity and high engagement in mental health and suicide prevention services following identification among veterans who remained at elevated risk in the 6 months that followed.

A Machine-Learning-Based Risk-Prediction Tool for HIV and Sexually Transmitted Infections Acquisition over the Next 12 Months

Background: More than one million people acquire sexually transmitted infections (STIs) every day globally. It is possible that predicting an individual’s future risk of HIV/STIs could contribute to behaviour change or improve testing. We developed a series of machine learning models and a subsequent risk-prediction tool for predicting the risk of HIV/STIs over the next 12 months. Methods: Our data included individuals who were re-tested at the clinic for HIV (65,043 consultations), syphilis (56,889 consultations), gonorrhoea (60,598 consultations), and chlamydia (63,529 consultations) after initial consultations at the largest public sexual health centre in Melbourne from 2 March 2015 to 31 December 2019. We used the receiver operating characteristic (AUC) curve to evaluate the model’s performance. The HIV/STI risk-prediction tool was delivered via a web application. Results: Our risk-prediction tool had an acceptable performance on the testing datasets for predicting HIV (AUC = 0.72), syphilis (AUC = 0.75), gonorrhoea (AUC = 0.73), and chlamydia (AUC = 0.67) acquisition. Conclusions: Using machine learning techniques, our risk-prediction tool has acceptable reliability in predicting HIV/STI acquisition over the next 12 months. This tool may be used on clinic websites or digital health platforms to form part of an intervention tool to increase testing or reduce future HIV/STI risk.

Translating promise into practice: a review of machine learning in suicide research and prevention

In ever more pressured health-care systems, technological solutions offering scalability of care and better resource targeting are appealing. Research on machine learning as a technique for identifying individuals at risk of suicidal ideation, suicide attempts, and death has grown rapidly. This research often places great emphasis on the promise of machine learning for preventing suicide, but overlooks the practical, clinical implementation issues that might preclude delivering on such a promise. In this Review, we synthesise the broad empirical and review literature on electronic health record-based machine learning in suicide research, and focus on matters of crucial importance for implementation of machine learning in clinical practice. The challenge of preventing statistically rare outcomes is well known; progress requires tackling data quality, transparency, and ethical issues. In the future, machine learning models might be explored as methods to enable targeting of interventions to specific individuals depending upon their level of need-ie, for precision medicine. Primarily, however, the promise of machine learning for suicide prevention is limited by the scarcity of high-quality scalable interventions available to individuals identified by machine learning as being at risk of suicide.

Longitudinal trajectories of suicidal ideation and attempts in adolescents with psychiatric disorders in Chile: study protocol

INTRODUCTION

Adolescent suicide is a worldwide public health problem, being the second and the third leading cause of death in the 15-29 and the 15-19 age groups, respectively. Among adolescents, it is estimated that for every suicide, there are 100-200 suicide attempts. Although 79% of suicides in the world occur in low/middle-income countries, most of scientific evidence comes from high-income and low-risk countries. In recent years, adolescent suicide rates have steadily increased in Chile. Deaths caused by self-harm increased by 220% in the population aged 10-19 years between 2000 and 2015. The Maule Region is one of the regions of Chile with the highest levels of suicide among those aged 15 and 19 years old. The objective of this study is to evaluate the trajectories of ideation and suicidal attempts in adolescents with psychiatric disorders treated within the public health system of the Maule Region, Chile, based on different clinical, psychological and neuropsychological factors.

METHOD

A prospective naturalistic study of a clinical sample of adolescents under psychiatric treatment in the Maule Region, Chile. Adolescents will be evaluated using a thorough protocol that includes suicide-related clinical variables. The study seeks to establish patterns of change in the trajectories of ideation and suicide attempts among adolescents.

ETHICS AND DISSEMINATION

Ethical approval was granted by the Scientific Ethics Committee of the Universidad Católica del Maule in Chile. This protocol was registered in ClinicalTrials.gov. The results of this study will be disseminated to health centres through executive reports and feedback sessions. In addition, the most relevant findings will be presented in scientific articles, conferences and seminars open to the community.

TRIAL REGISTRATION NUMBER

NCT04635163.

Screening and Assessment for Psychological Distress among Burn Survivors

Given the high rates of psychological distress after burn injury, thorough screening and assessment for psychosocial factors and psychiatric pathology should be routinely completed for individuals with burn injuries. Burn survivors experience unique psychosocial changes and injury sequelae, such as body image concerns, trauma-related pathology, and itching. Screening for these factors is integral to understanding how these may be contributing to psychological distress. Proactively identifying distress and psychiatric pathology is important to optimize physical and emotional outcomes. The aim of this manuscript is to summarize information about the available screening and assessment tools for psychological distress among burn survivors.

Childhood trauma and the severity of past suicide attempts in outpatients with cocaine use disorders

INTRODUCTION

Suicide attempts have been associated with both cocaine use disorder (CocUD) and childhood trauma. We investigated how childhood trauma is an independent risk factor for serious and recurrent suicide attempts in CocUD. Method: 298 outpatients (23% women) with CocUD underwent standardized assessments of substance dependence (Diagnostic and Statistical Manual-mental disorders, fourth edition, text revised), impulsiveness, resilience, and childhood trauma, using validated tools. Suicide attempts history was categorized as single vs. recurrent or non-serious vs. serious depending on the lifetime number of suicide attempts and the potential or actual lethality of the worst attempt reported, respectively. Bivariate and multinomial regression analyses were used to characterize which childhood trauma patterns were associated with the suicide attempts groups. Results: 58% of CocUD patients reported childhood trauma. Recurrent and serious suicide attempts clustered together and were thus combined into "severe SA." Severe suicide attempt risk increased proportionally to the number of childhood traumas (test for trend, p = 9 × 10^-7). Non-severe suicide attempt risk increased with impulsiveness and decreased with resilience. In multinomial regression models, a higher number of traumas and emotional abuse were independently and only associated with severe vs. non-severe suicide attempts (effect size = 0.82, AUC = 0.7). The study was limited by its cross-sectional design. Conclusion: These preferential associations between childhood trauma and severe suicide attempts warrant specific monitoring of suicide attempts risk in CocUD, regardless of the severity of addiction profiles.

Gatekeeper training for friends and family of individuals at risk of suicide: A systematic review

AIMS

Gatekeeper training (GKT) is an important suicide prevention strategy. Studies have evaluated the effectiveness of GKT in different populations, often neglecting family and friends who play a vital role in caring for people with suicide risk. This review evaluated GKT programs targeting family and friends to determine their effectiveness in this specific population.

METHODS

Academic databases were searched for studies on GKT programs. Programs involving family and friends caring for people with suicide risk were assessed for any impact on knowledge, self-efficacy, attitudes, and suicide prevention skills.

RESULTS

Seventeen studies were reviewed. GKT showed significant gains on outcomes of interest. Three studies targeted family and friends, with one involving them in program creation and conduction and another adjusting the program after their input.

CONCLUSIONS

GKT programs have potentially positive effects on family and friends caring for people with suicide risk. Few programs address the specific needs of this group, and programs adapted specifically for them are scarce. Future program development recommendations are discussed.

Suicide Risk Screening and Suicide Prevention in Patients With Cancer

Background

Suicide rates are up to 4 times greater in cancer compared with the general population, yet best practices for institutional suicide prevention are unknown. The objective of this study was to examine the association between suicide risk screening (SRS), clinician response, and suicide mortality at a comprehensive cancer treatment center.

Methods

We conducted a naturalistic, retrospective cohort study of patients attending the Princess Margaret Cancer Centre, where routine screening for suicidal intent within the Distress Assessment and Response Tool (DART-SRS) was implemented in 2010. Inverse probability of treatment weighting was used to evaluate the impact of DART-SRS completion on suicide mortality from 2005 to 2014. Chart audits were conducted for clinician response to suicidality, and crude suicide rates over the study period were analyzed. All statistical tests were 2-sided.

Results

Among 78 650 cancer patients, 89 (0.1%) died by suicide, of whom only 4 (4.5%) had completed DART-SRS. Among DART-SRS completers (n = 14 517), 69 (0.5%) reported suicidal intent, none of whom died by suicide. DART-SRS completion was associated with increased clinician response to suicidality (17.4% vs 6.7%, P = .04), more psychosocial service usage (30.5% vs 18.3%, P < .001), and lower suicide mortality (hazard ratio = 0.29, 95% confidence interval = 0.28 to 0.31). Crude suicide rates at the Princess Margaret Cancer Centre were lower in patients whose first contact year was after DART-SRS implementation.

Conclusion

DART-SRS completion is associated with lower suicide mortality and increased access to psychosocial care, but patients who did not complete DART-SRS were at highest suicide risk. Further research is needed to identify mechanisms to ensure psychosocial and suicidality assessment in cancer patients who do not complete SRS.

Accuracy Requirements for Cost-effective Suicide Risk Prediction Among Primary Care Patients in the US

Importance

Several statistical models for predicting suicide risk have been developed, but how accurate such models must be to warrant implementation in clinical practice is not known.

Objective

To identify threshold values of sensitivity, specificity, and positive predictive value that a suicide risk prediction method must attain to cost-effectively target a suicide risk reduction intervention to high-risk individuals.

Design, Setting, and Participants

This economic evaluation incorporated published data on suicide epidemiology, the health care and societal costs of suicide, and the costs and efficacy of suicide risk reduction interventions into a novel decision analytic model. The model projected suicide-related health economic outcomes over a lifetime horizon among a population of US adults with a primary care physician. Data analysis was performed from September 19, 2019, to July 5, 2020.

Interventions

Two possible interventions were delivered to individuals at high predicted risk: active contact and follow-up (ACF; relative risk of suicide attempt, 0.83; annual health care cost, $96) and cognitive behavioral therapy (CBT; relative risk of suicide attempt, 0.47; annual health care cost, $1088).

Main Outcomes and Measures

Fatal and nonfatal suicide attempts, quality-adjusted life-years (QALYs), health care sector costs and societal costs (in 2016 US dollars), and incremental cost-effectiveness ratios (ICERs) (with ICERs ≤$150 000 per QALY designated cost-effective).

Results

With a specificity of 95% and a sensitivity of 25%, primary care-based suicide risk prediction could reduce suicide death rates by 0.5 per 100 000 person-years (if used to target ACF) or 1.6 per 100 000 person-years (if used to target CBT) from a baseline of 15.3 per 100 000 person-years. To be cost-effective from a health care sector perspective at a specificity of 95%, a risk prediction method would need to have a sensitivity of 17.0% or greater (95% CI, 7.4%-37.3%) if used to target ACF and 35.7% or greater (95% CI, 23.1%-60.3%) if used to target CBT. To achieve cost-effectiveness, ACF required positive predictive values of 0.8% for predicting suicide attempt and 0.07% for predicting suicide death; CBT required values of 1.7% for suicide attempt and 0.2% for suicide death.

Conclusions and Relevance

These findings suggest that with sufficient accuracy, statistical suicide risk prediction models can provide good health economic value in the US. Several existing suicide risk prediction models exceed the accuracy thresholds identified in this analysis and thus may warrant pilot implementation in US health care systems.

Columbia-Suicide Severity Rating Scale Screen Version: initial screening for suicide risk in a psychiatric emergency department

BACKGROUND

Suicide screening is routine practice in psychiatric emergency (PE) departments, but evidence for screening instruments is sparse. Improved identification of nascent suicide risk is important for suicide prevention. The aim of the current study was to evaluate the association between the novel Colombia Suicide Severity Rating Scale Screen Version (C-SSRS Screen) and subsequent clinical management and suicide within 1 week, 1 month and 1 year from screening.

METHODS

Consecutive patients (N = 18 684) attending a PE department in Stockholm, Sweden between 1 May 2016 and 31 December 2017 were assessed with the C-SSRS Screen. All patients (52.1% women; mean age = 39.7, s.d. = 16.9) were followed-up in the National Cause of Death Register. Logistic regression and receiver operating characteristic curves analyses were conducted. Optimal cut-offs and accuracy statistics were calculated.

RESULTS

Both suicidal ideation and behaviour were prevalent at screening. In total, 107 patients died by suicide during follow-up. Both C-SSRS Screen Ideation Severity and Behaviour Scales were associated with death by suicide within 1-week, 1-month and 1-year follow-up. The optimal cut-off for the ideation severity scale was associated with at least four times the odds of dying by suicide within 1 week (adjusted OR 4.7, 95% confidence interval 1.5-14.8). Both scales were also associated with short-term clinical management.

CONCLUSIONS

The C-SSRS Screen may be feasible to use in the actual management setting as an initial step before the clinical assessment of suicide risk. Future research may investigate the utility of combining the C-SSRS Screen with a more thorough assessment.

[Comparison of the Prediction Model of Adolescents' Suicide Attempt Using Logistic Regression and Decision Tree: Secondary Data Analysis of the 2019 Youth Health Risk Behavior Web-Based Survey]

PURPOSE

The purpose of this study was to develop and compare the prediction model for suicide attempts by Korean adolescents using logistic regression and decision tree analysis.

METHODS

This study utilized secondary data drawn from the 2019 Youth Health Risk Behavior web-based survey. A total of 20 items were selected as the explanatory variables (5 of sociodemographic characteristics, 10 of health-related behaviors, and 5 of psychosocial characteristics). For data analysis, descriptive statistics and logistic regression with complex samples and decision tree analysis were performed using IBM SPSS ver. 25.0 and Stata ver. 16.0.

RESULTS

A total of 1,731 participants (3.0%) out of 57,303 responded that they had attempted suicide. The most significant predictors of suicide attempts as determined using the logistic regression model were experience of sadness and hopelessness, substance abuse, and violent victimization. Girls who have experience of sadness and hopelessness, and experience of substance abuse have been identified as the most vulnerable group in suicide attempts in the decision tree model.

CONCLUSION

Experiences of sadness and hopelessness, experiences of substance abuse, and experiences of violent victimization are the common major predictors of suicide attempts in both logistic regression and decision tree models, and the predict rates of both models were similar. We suggest to provide programs considering combination of high-risk predictors for adolescents to prevent suicide attempt.

Violence and mental disorders: a structured review of associations by individual diagnoses, risk factors, and risk assessment

In this Review, we summarise evidence on the association between different mental disorders and violence, with emphasis on high quality designs and replicated findings. Relative risks are typically increased for all violent outcomes in most diagnosed psychiatric disorders compared with people without psychiatric disorders, with increased odds in the range of 2-4 after adjustment for familial and other sources of confounding. Absolute rates of violent crime over 5-10 years are typically below 5% in people with mental illness (excluding personality disorders, schizophrenia, and substance misuse), which increases to 6-10% in personality disorders and schizophrenia spectrum disorders, and to more than 10% in substance misuse. Past criminality and comorbid substance misuse are strongly predictive of future violence in many individual disorders. We reviewed national clinical practice guidelines, which vary in content and require updating to reflect the present epidemiological evidence. Standardised and clinically feasible approaches to the assessment and management of violence risk in general psychiatric settings need to be developed.

Akut suizidal-affektive Störung: Ein systematisches Review

<b><i>Hintergrund:</i></b> Die Arbeitsgruppe um Thomas Joiner postuliert ein als <i>acute suicidal affective disturbance</i> (ASAD) bezeichnetes Syndrom, welches durch vier Symptomgruppen (Suizidabsicht, Entfremdung, Hoffnungslosigkeit, Übererregung) definiert sein soll. Ziel des vorliegenden Artikels ist, die Literatur zum ASAD-Syndrom zusammenfassend darzustellen und eine Einschätzung vorzunehmen, inwieweit tatsächlich von einem einheitlichen Syndrom ausgegangen werden kann. <b><i>Methoden:</i></b> Im Rahmen einer Literaturrecherche konnten neun Artikel identifiziert werden, die im Zeitraum von 2016 bis 2020 zum Thema publiziert wurden. <b><i>Ergebnisse und Schlussfolgerungen:</i></b> Die Befundlage unterstützt den einheitlichen Störungscharakter und die Ab­grenzbarkeit der akut suizidal-affektiven Störung von anderen Störungsbildern. Die Aussagekraft der Befundlage ist dadurch eingeschränkt, dass bislang ausschließlich Querschnittsuntersuchungen durchgeführt wurden und keinerlei Befunde dazu vorliegen, ob ASAD tatsächlich suizidalem Verhalten vorausgeht.

Predicting suicide attempt or suicide death following a visit to psychiatric specialty care: A machine learning study using Swedish national registry data

BACKGROUND

Suicide is a major public health concern globally. Accurately predicting suicidal behavior remains challenging. This study aimed to use machine learning approaches to examine the potential of the Swedish national registry data for prediction of suicidal behavior.

METHODS AND FINDINGS

The study sample consisted of 541,300 inpatient and outpatient visits by 126,205 Sweden-born patients (54% female and 46% male) aged 18 to 39 (mean age at the visit: 27.3) years to psychiatric specialty care in Sweden between January 1, 2011 and December 31, 2012. The most common psychiatric diagnoses at the visit were anxiety disorders (20.0%), major depressive disorder (16.9%), and substance use disorders (13.6%). A total of 425 candidate predictors covering demographic characteristics, socioeconomic status (SES), electronic medical records, criminality, as well as family history of disease and crime were extracted from the Swedish registry data. The sample was randomly split into an 80% training set containing 433,024 visits and a 20% test set containing 108,276 visits. Models were trained separately for suicide attempt/death within 90 and 30 days following a visit using multiple machine learning algorithms. Model discrimination and calibration were both evaluated. Among all eligible visits, 3.5% (18,682) were followed by a suicide attempt/death within 90 days and 1.7% (9,099) within 30 days. The final models were based on ensemble learning that combined predictions from elastic net penalized logistic regression, random forest, gradient boosting, and a neural network. The area under the receiver operating characteristic (ROC) curves (AUCs) on the test set were 0.88 (95% confidence interval [CI] = 0.87-0.89) and 0.89 (95% CI = 0.88-0.90) for the outcome within 90 days and 30 days, respectively, both being significantly better than chance (i.e., AUC = 0.50) (p < 0.01). Sensitivity, specificity, and predictive values were reported at different risk thresholds. A limitation of our study is that our models have not yet been externally validated, and thus, the generalizability of the models to other populations remains unknown.

CONCLUSIONS

By combining the ensemble method of multiple machine learning algorithms and high-quality data solely from the Swedish registers, we developed prognostic models to predict short-term suicide attempt/death with good discrimination and calibration. Whether novel predictors can improve predictive performance requires further investigation.

Exposure to suicidality in professionals working with oncology patients: An online survey

OBJECTIVE

To explore and describe exposure to suicidality in healthcare providers (HCP) working with oncological patients. Special emphasis was put on five central aspects from the HCPs perspective: Exposure, Confidence, Expertise, Distress, and Education.

METHODS

A 48-item online questionnaire was developed and distributed to HCPs working with cancer patients. Three hundred fifty-four answered questionnaires were analyzed.

RESULTS

Overall 83.3% of HCPs reported to have encountered at least one suicidal patient in the last year. Feeling confident in talking about suicidality was reported by 72.1% of HCPs, with 71.2% of nurses reporting feeling insecure compared with only 5.1% of psychotherapists. Similarly, 22.3% of HCPs felt overwhelmed when confronted with a patient who substantiated his suicidality during consultation. A lack of personal knowledge concerning suicidality in general and in oncological patients in particular, was reported by 39.6% and 49.8%, respectively. In total, 88.1% of HCPs reported feeling distressed when confronted with suicidality, while 81.1% of participants wanted further education regarding suicidality in cancer patients despite that 73.2% had already received some sort of psycho-oncology education.

CONCLUSIONS

Despite the well-documented fact of elevated suicide rates in cancer patients, there remain deficits in knowledge, which induce feelings of insecurity and helplessness in HCPs. There is a demand for further education concerning the treatment of suicidal cancer patients. Therefore, special curricula addressing this topic should be devised. A general debate about suicidality in cancer patients could help raise awareness of this problem and generate means of prevention.

What would cardiology do? Lessons from other medical specialties should help guide suicide prevention research

Suicide is among the most important causes of mortality in medicine as it is the most common cause of death due to illness from the teenage years into middle age. Yet our approach to mental health research aimed at suicide prevention has often diverged from accepted practices in other areas of medicine. This includes the exclusion of those at highest risk of suicide from clinical trials and the recent emphasis on prediction. In this Viewpoint, we propose that comparing our approach to that of other medical specialties would help us to avoid strategic errors and discuss the implications for the field of suicide prevention.

Identifying Predictors of Suicide in Severe Mental Illness: A Feasibility Study of a Clinical Prediction Rule (Oxford Mental Illness and Suicide Tool or OxMIS)

BACKGROUND

Oxford Mental Illness and Suicide tool (OxMIS) is a brief, scalable, freely available, structured risk assessment tool to assess suicide risk in patients with severe mental illness (schizophrenia-spectrum disorders or bipolar disorder). OxMIS requires further external validation, but a lack of large-scale cohorts with relevant variables makes this challenging. Electronic health records provide possible data sources for external validation of risk prediction tools. However, they contain large amounts of information within free-text that is not readily extractable. In this study, we examined the feasibility of identifying suicide predictors needed to validate OxMIS in routinely collected electronic health records.

METHODS

In study 1, we manually reviewed electronic health records of 57 patients with severe mental illness to calculate OxMIS risk scores. In study 2, we examined the feasibility of using natural language processing to scale up this process. We used anonymized free-text documents from the Clinical Record Interactive Search database to train a named entity recognition model, a machine learning technique which recognizes concepts in free-text. The model identified eight concepts relevant for suicide risk assessment: medication (antidepressant/antipsychotic treatment), violence, education, self-harm, benefits receipt, drug/alcohol use disorder, suicide, and psychiatric admission. We assessed model performance in terms of precision (similar to positive predictive value), recall (similar to sensitivity) and F1 statistic (an overall performance measure).

RESULTS

In study 1, we estimated suicide risk for all patients using the OxMIS calculator, giving a range of 12 month risk estimates from 0.1-3.4%. For 13 out of 17 predictors, there was no missing information in electronic health records. For the remaining 4 predictors missingness ranged from 7-26%; to account for these missing variables, it was possible for OxMIS to estimate suicide risk using a range of scores. In study 2, the named entity recognition model had an overall precision of 0.77, recall of 0.90 and F1 score of 0.83. The concept with the best precision and recall was medication (precision 0.84, recall 0.96) and the weakest were suicide (precision 0.37), and drug/alcohol use disorder (recall 0.61).

CONCLUSIONS

It is feasible to estimate suicide risk with the OxMIS tool using predictors identified in routine clinical records. Predictors could be extracted using natural language processing. However, electronic health records differ from other data sources, particularly for family history variables, which creates methodological challenges.

Antidepressants in Children and Adolescents: Meta-Review of Efficacy, Tolerability and Suicidality in Acute Treatment

Antidepressants are prescribed for the treatment of a number of psychiatric disorders in children and adolescents, however there is still controversy about whether they should be used in this population. This meta-review aimed to assess the effects of antidepressants for the acute treatment of attention-deficit/hyperactivity disorder (ADHD), anxiety disorders (ADs), autistic spectrum disorder (ASD), enuresis, major depressive disorder (MDD), obsessive-compulsive disorder (OCD), and posttraumatic stress disorder (PTSD) in children and adolescents. Efficacy was measured as response to treatment (either as mean overall change in symptoms or as a dichotomous outcome) and tolerability was measured as the proportion of patients discontinuing treatment due to adverse events. Suicidality was measured as suicidal ideation, behavior (including suicide attempts) and completed suicide. PubMed, EMBASE, and Web of Science were systematically searched (until 31 October 2019) for existing systematic reviews and/or meta-analyses of double-blind randomized controlled trials. The quality of the included reviews was appraised using AMSTAR-2. Our meta-review included nine systematic reviews/meta-analyses (2 on ADHD; 1 on AD; 2 on ASD; 1 on enuresis; 1 on MDD, 1 on OCD and 1 on PTSD). In terms of efficacy this review found that, compared to placebo: fluoxetine was more efficacious in the treatment of MDD, fluvoxamine and paroxetine were better in the treatment of AD; fluoxetine and sertraline were more efficacious in the treatment of OCD; bupropion and desipramine improved clinician and teacher-rated ADHD symptoms; clomipramine and tianeptine were superior on some of the core symptoms of ASD; and no antidepressant was more efficacious for PTSD and enuresis. With regard to tolerability: imipramine, venlafaxine, and duloxetine were less well tolerated in MDD; no differences were found for any of the antidepressants in the treatment of anxiety disorders (ADs), ADHD, and PTSD; tianeptine and citalopram, but not clomipramine, were less well tolerated in children and adolescents with ASD. For suicidal behavior/ideation, venlafaxine (in MDD) and paroxetine (in AD) were associated with a significantly increased risk; by contrast, sertraline (in AD) was associated with a reduced risk. The majority of included systematic reviews/meta-analyses were rated as being of high or moderate in quality by the AMSTAR-2 critical appraisal tool (one and five, respectively). One included study was of low quality and two were of critically low quality. Compared to placebo, selected antidepressants can be efficacious in the acute treatment of some common psychiatric disorders, although statistically significant differences do not always translate into clinically significant results. Little information was available about tolerability of antidepressants in RCTs of OCD and in the treatment of ADHD, ASD, MDD, and PTSD. There is a paucity of data on suicidal ideation/behavior, but paroxetine may increase the risk of suicidality in the treatment of AD and venlafaxine for MDD. Findings from this review must be considered in light of potential limitations, such as the lack of comparative information about many antidepressants, the short-term outcomes and the quality of the available evidence.