Evaluation of a Model to Target High-risk Psychiatric Inpatients for an Intensive Postdischarge Suicide Prevention Intervention

Design of a multicenter randomized controlled trial of a post-discharge suicide prevention intervention for high-risk psychiatric inpatients: The Veterans Coordinated Community Care Study

BACKGROUND

The period after psychiatric hospital discharge is one of elevated risk for suicide-related behaviors (SRBs). Post-discharge clinical outreach, although potentially effective in preventing SRBs, would be more cost-effective if targeted at high-risk patients. To this end, a machine learning model was developed to predict post-discharge suicides among Veterans Health Administration (VHA) psychiatric inpatients and target a high-risk preventive intervention.

METHODS

The Veterans Coordinated Community Care (3C) Study is a multicenter randomized controlled trial using this model to identify high-risk VHA psychiatric inpatients (n = 850) randomized with equal allocation to either the Coping Long Term with Active Suicide Program (CLASP) post-discharge clinical outreach intervention or treatment-as-usual (TAU). The primary outcome is SRBs over a 6-month follow-up. We will estimate average treatment effects adjusted for loss to follow-up and investigate the possibility of heterogeneity of treatment effects.

RESULTS

Recruitment is underway and will end September 2024. Six-month follow-up will end and analysis will begin in Summer 2025.

CONCLUSION

Results will provide information about the effectiveness of CLASP versus TAU in reducing post-discharge SRBs and provide guidance to VHA clinicians and policymakers about the implications of targeted use of CLASP among high-risk psychiatric inpatients in the months after hospital discharge.

CLINICAL TRIALS REGISTRATION

ClinicalTrials.Gov identifier: NCT05272176 (https://www.

CLINICALTRIALS

gov/ct2/show/NCT05272176).

Performance of Machine Learning Suicide Risk Models in an American Indian Population

Importance

Few suicide risk identification tools have been developed specifically for American Indian and Alaska Native populations, even though these populations face the starkest suicide-related inequities.

Objective

To examine the accuracy of existing machine learning models in a majority American Indian population.

Design, Setting, and Participants

This prognostic study used secondary data analysis of electronic health record data collected from January 1, 2017, to December 31, 2021. Existing models from the Mental Health Research Network (MHRN) and Vanderbilt University (VU) were fitted. Models were compared with an augmented screening indicator that included any previous attempt, recent suicidal ideation, or a recent positive suicide risk screen result. The comparison was based on the area under the receiver operating characteristic curve (AUROC). The study was performed in partnership with a tribe and local Indian Health Service (IHS) in the Southwest. All patients were 18 years or older with at least 1 encounter with the IHS unit during the study period. Data were analyzed between October 6, 2022, and July 29, 2024.

Exposures

Suicide attempts or deaths within 90 days.

Main Outcomes and Measures

Model performance was compared based on the ability to distinguish between those with a suicide attempt or death within 90 days of their last IHS visit with those without this outcome.

Results

Of 16 835 patients (mean [SD] age, 40.0 [17.5] years; 8660 [51.4%] female; 14 251 [84.7%] American Indian), 324 patients (1.9%) had at least 1 suicide attempt, and 37 patients (0.2%) died by suicide. The MHRN model had an AUROC value of 0.81 (95% CI, 0.77-0.85) for 90-day suicide attempts, whereas the VU model had an AUROC value of 0.68 (95% CI, 0.64-0.72), and the augmented screening indicator had an AUROC value of 0.66 (95% CI, 0.63-0.70). Calibration was poor for both models but improved after recalibration.

Conclusion and Relevance

This prognostic study found that existing risk identification models for suicide prevention held promise when applied to new contexts and performed better than relying on a combined indictor of a positive suicide risk screen result, history of attempt, and recent suicidal ideation.

Using Natural Language Processing to develop risk-tier specific suicide prediction models for Veterans Affairs patients

Suicide is a leading cause of death. Suicide rates are particularly elevated among Department of Veterans Affairs (VA) patients. While VA has made impactful suicide prevention advances, efforts primarily target high-risk patients with documented suicide risk. This high-risk population accounts for less than 10% of VA patient suicide deaths. We previously evaluated epidemiological patterns among VA patients that had lower classified suicide risk and derived moderate- and low-risk groupings. Expanding upon VA's leading suicide prediction model, this study uses national VA data to refine high-, moderate-, and low-risk specific suicide prediction methods. We selected all VA patients who died by suicide in 2017 or 2018 (n = 4584), matching each case with five controls who remained alive during treatment year and shared suicide risk percentiles. We extracted all sample unstructured electronic health record notes, analyzed them using natural language processing, and applied machine-learning classification algorithms to develop risk-tier-specific predictive models. We calculated area under the curve (AUC) and suicide risk concentration to evaluate predictive accuracy and analyzed derived words. RESULTS: Our high-risk model (AUC = 0.621 (95% CI: 0.55-0.68)), moderate-risk (AUC = 0.669 (95% CI: 0.64-0.71)), and low-risk (AUC = 0.673 (95% CI: 0.63-0.72)) models offered significant predictive accuracy over VA's leading suicide prediction algorithm. Derived words varied considerably, the high-risk model including chronic condition service words, moderate-risk model including outpatient care, and low-risk model including acute condition care. Study suggests benefit of leveraging unstructured electronic health records and expands prediction resources for non-high-risk suicide decedents, an historically underserved population.

Predicting all-cause mortality and premature death using interpretable machine learning among a middle-aged and elderly Chinese population

Objective

To develop machine learning-based prediction models for all-cause and premature mortality among the middle-aged and elderly population in China.

Method

Adults aged 45 years or older at baseline of 2011 from the China Health and Retirement Longitudinal Study (CHARLS) were included. The stacked ensemble model was built utilizing five selected machine learning algorithms. These models underwent training and testing using the CHARLS 2011-2015 cohort (derivation cohort) and subsequently underwent external validation using the CHARLS 2015-2018 cohort (validation cohort). SHapley Additive exPlanations (SHAP) was introduced to quantify the importance of risk factors and explain machine learning algorithms.

Result

In derivation cohort, a total of 10,677 subjects were included, 478 died during the follow-up. The stacked ensemble model demonstrated the highest efficacy in terms of its discrimination capability for predicting all-cause mortality and premature death, with an AUC[95 % CI] of 0.826[0.792-0.859] and 0.773[0.725-0.821], respectively. In validation cohort, the corresponding AUC[95 % CI] were 0.803[0.743-0.864] and 0.791[0.719-0.863], respectively. Risk factors including age, sex, self-reported health, activities of daily living, cognitive function, ever smoker, levels of systolic blood pressure, Cystatin C and low density lipoprotein were strong predictors for both all-cause mortality and premature death.

Conclusion

Stacked ensemble models performed well in predicting all-cause and premature death in this Chinese cohort. Interpretable techniques can aid in identifying significant risk factors and non-linear relationships between predictors and mortality.

A Secondary Analysis of the Complex Interplay between Psychopathology, Cognitive Functions, Brain Derived Neurotrophic Factor Levels, and Suicide in Psychotic Disorders: Data from a 2-Year Longitudinal Study

Identifying phenotypes at high risk of suicidal behaviour is a relevant objective of clinical and translational research and can facilitate the identification of possible candidate biomarkers. We probed the potential association and eventual stability of neuropsychological profiles and serum BDNF concentrations with lifetime suicide ideation and attempts (LSI and LSA, respectively) in individuals with schizophrenia (SCZ) and schizoaffective disorder (SCA) in a 2-year follow-up study. A secondary analysis was conducted on a convenience sample of previously recruited subjects from a single outpatient clinic. Retrospectively assessed LSI and LSA were recorded by analysing the available longitudinal clinical health records. LSI + LSA subjects consistently exhibited lower PANSS-defined negative symptoms and better performance in the BACS-letter fluency subtask. There was no significant association between BDNF levels and either LSI or LSA. We found a relatively stable pattern of lower negative symptoms over two years among patients with LSI and LSA. No significant difference in serum BDNF concentrations was detected. The translational viability of using neuropsychological profiles as a possible avenue for the identification of populations at risk for suicide behaviours rather than the categorical diagnosis represents a promising option but requires further confirmation.

Validation of a Multivariable Model to Predict Suicide Attempt in a Mental Health Intake Sample

Importance

Given that suicide rates have been increasing over the past decade and the demand for mental health care is at an all-time high, targeted prevention efforts are needed to identify individuals seeking to initiate mental health outpatient services who are at high risk for suicide. Suicide prediction models have been developed using outpatient mental health encounters, but their performance among intake appointments has not been directly examined.

Objective

To assess the performance of a predictive model of suicide attempts among individuals seeking to initiate an episode of outpatient mental health care.

Design, Setting, and Participants

This prognostic study tested the performance of a previously developed machine learning model designed to predict suicide attempts within 90 days of any mental health outpatient visit. All mental health intake appointments scheduled between January 1, 2012, and April 1, 2022, at Kaiser Permanente Northern California, a large integrated health care delivery system serving over 4.5 million patients, were included. Data were extracted and analyzed from August 9, 2022, to July 31, 2023.

Main Outcome and Measures

Suicide attempts (including completed suicides) within 90 days of the appointment, determined by diagnostic codes and government databases. All predictors were extracted from electronic health records.

Results

The study included 1 623 232 scheduled appointments from 835 616 unique patients. There were 2800 scheduled appointments (0.17%) followed by a suicide attempt within 90 days. The mean (SD) age across appointments was 39.7 (15.8) years, and most appointments were for women (1 103 184 [68.0%]). The model had an area under the receiver operating characteristic curve of 0.77 (95% CI, 0.76-0.78), an area under the precision-recall curve of 0.02 (95% CI, 0.02-0.02), an expected calibration error of 0.0012 (95% CI, 0.0011-0.0013), and sensitivities of 37.2% (95% CI, 35.5%-38.9%) and 18.8% (95% CI, 17.3%-20.2%) at specificities of 95% and 99%, respectively. The 10% of appointments at the highest risk level accounted for 48.8% (95% CI, 47.0%-50.6%) of the appointments followed by a suicide attempt.

Conclusions and Relevance

In this prognostic study involving mental health intakes, a previously developed machine learning model of suicide attempts showed good overall classification performance. Implementation research is needed to determine appropriate thresholds and interventions for applying the model in an intake setting to target high-risk cases in a manner that is acceptable to patients and clinicians.

Premature Death, Suicide, and Nonlethal Intentional Self-Harm After Psychiatric Discharge

Importance

There is a need for representative research on serious adverse outcomes following discharge from psychiatric hospitalization.

Objective

To compare rates of premature death, suicide, and nonlethal intentional self-harm after psychiatric discharge with rates in the general population and investigate associations of these outcomes with relevant variables associated with the index psychiatric hospitalization.

Design, Setting, and Participants

This retrospective cohort study included all residents from Catalonia, Spain (7.6 million population), who had psychiatric hospitalizations between January 1, 2014, and December 31, 2018, and were older than 10 years at the index (first) hospitalization. Follow-up was until December 31, 2019. Statistical analysis was performed from December 1, 2022, through April 11, 2024.

Exposures

Socioeconomic status, psychiatric diagnoses, duration of index hospitalization, and number of previous psychiatric hospitalizations.

Main Outcomes and Measures

Postdischarge premature death (ie, all-cause death before age 70 years) and suicide (International Statistical Classification of Diseases and Related Health Problems, Tenth Revision [ICD-10] code range X60-X84), identified using mortality data, and postdischarge nonlethal intentional self-harm, identified using electronic health record and self-harm case register data. Standardized mortality ratios (SMRs) compared rates of premature death and suicide between the cohort and the general population. Fully adjusted, multivariable, cause-specific Cox proportional hazards regression models for the 3 outcomes were fitted.

Results

A total of 49 108 patients discharged from psychiatric hospitalization were included (25 833 males [52.6%]; mean [SD] age at discharge, 44.2 [18.2] years). During follow-up, 2260 patients (4.6%) died prematurely, 437 (0.9%) died by suicide, and 4752 (9.7%) had an episode of nonlethal intentional self-harm. The overall SMR for premature death was 7.5 (95% CI, 7.2-7.9). For suicide, SMR was 32.9 (95% CI, 29.9-36.0) overall and was especially high among females (47.6 [95% CI, 40.2-54.9]). In fully adjusted sex-stratified hazard models, postdischarge premature death was associated with cognitive disorders (adjusted hazard ratio [AHR], 2.89 [95% CI, 2.24-3.74] for females; 2.59 [95% CI, 2.17-3.08] for males) and alcohol-related disorders (AHR, 1.41 [95% CI, 1.18-1.70] for females; 1.22 [95% CI, 1.09-1.37] for males). Postdischarge suicide was associated with postdischarge intentional self-harm (AHR, 2.83 [95% CI, 1.97-4.05] for females; 3.29 [95% CI, 2.47-4.40] for males), with depressive disorders (AHR, 2.13 [95% CI, 1.52-2.97]) and adjustment disorders (AHR, 1.94 [95% CI, 1.32-2.83]) among males, and with bipolar disorder among females (AHR, 1.94 [95% CI, 1.21-3.09]). Postdischarge intentional self-harm was associated with index admissions for intentional self-harm (AHR, 1.95 [95% CI, 1.73-2.21] for females; 2.62 [95% CI, 2.20-3.13] for males) as well as for adjustment disorders (AHR, 1.48 [95% CI, 1.33-1.65] for females; 1.99 [95% CI, 1.74-2.27] for males), anxiety disorders (AHR, 1.24 [95% CI, 1.10-1.39] for females; 1.36 [95% CI, 1.18-1.58] for males), depressive disorders (AHR, 1.54 [95% CI, 1.40-1.69] for females; 1.80 [95% CI, 1.58-2.04] for males), and personality disorders (AHR, 1.59 [95% CI, 1.46-1.73] for females; 1.43 [95% CI, 1.28-1.60] for males).

Conclusions and Relevance

In this cohort study of patients discharged from psychiatric hospitalization, risk for premature death and suicide was significantly higher compared with the general population, suggesting individuals discharged from psychiatric inpatient care are a vulnerable population for premature death and suicidal behavior.

Variations in Suicide Risk and Risk Factors After Hospitalization for Depression in Finland, 1996-2017

Importance

Although incidence of suicide in depression varies remarkably temporally, risk factors have been modeled as constant and remain uncharted in the short term. How effectively factors measured at one point in time predict risk at different time points is unknown.

Objective

To examine the absolute risk and risk factors for suicide in hospitalized patients with depression starting from the first days after discharge up to 2 years and to evaluate whether the size of relative risk by factor displays temporal patterns over consecutive phases of follow-up.

Design, Setting, and Participants

This population-based study using Finnish registers (hospital discharge, population, and cause of death registers) included all hospitalizations for depression as the principal diagnosis in Finland from 1996 to 2017, with a maximum follow-up of 2 years. Data were analyzed from January 2022 to November 2023.

Main Outcomes and Measures

Incidence rate (IR), IR ratios, hazard functions, and hazard ratios for suicide by consecutive time periods (0 to 3 days, 4 to 7 days, 7 to 30 days, 31 to 90 days, 91 to 365 days, and 1 to 2 years) since discharge.

Results

This study included 193 197 hospitalizations among 91 161 individuals, of whom 51 197 (56.2%) were female, and the mean (SD) age was 44.0 (17.3) years. Altogether, patients were followed up to 226 615 person-years. A total of 1219 men and 757 women died of suicide. Incidence of suicide was extremely high during the first days after discharge (IR of 6062 [95% CI, 4963-7404] per 100 000 on days 0 to 3; IR of 3884 [95% CI, 3119-4835] per 100 000 on days 4 to 7) and declined thereafter. Several factors were associated with risk of suicide over the first days after discharge. Current suicide attempt by hanging or firearms increased the risk of suicide most on days 0 to 3 (IR ratio, 18.9; 95% CI, 3.1-59.8) and on days 0 to 7 (IR ratio, 10.1; 95% CI, 1.7-31.5). Temporal patterns of the size of the relative risk diverged over time, being constant, declining, or increasing. Clinical factors had the strongest association immediately. Relative risk remained constant among men and even increased among those with alcohol or substance use disorder.

Conclusions and Relevance

In this study, patients hospitalized for depression had extremely high risk of suicide during the first days after discharge. Thereafter, incidence declined steeply but remained high. Within the periods of the highest risk of suicide, several factors increased overall risk manyfold. Risk factors' observed potencies varied over time and had characteristic temporal patterns.

Predicting Suicide Among US Veterans Using Natural Language Processing-enriched Social and Behavioral Determinants of Health

Despite recognizing the critical association between social and behavioral determinants of health (SBDH) and suicide risk, SBDHs from unstructured electronic health record (EHR) notes for suicide predictive modeling remain underutilized. This study investigates the impact of SBDH, identified from both structured and unstructured data utilizing a natural language processing (NLP) system, on suicide prediction within 7, 30, 90, and 180 days of discharge. Using EHR data of 2,987,006 Veterans between October 1, 2009, and September 30, 2015, from the US Veterans Health Administration (VHA), we designed a case-control study that demonstrates that incorporating structured and NLP-extracted SBDH significantly enhances the performance of three architecturally distinct suicide predictive models - elastic-net logistic regression, random forest (RF), and multilayer perceptron. For example, RF achieved notable improvements in suicide prediction within 180 days of discharge, with an increase in the area under the receiver operating characteristic curve from 83.57-84.25% (95% CI = 0.63%-0.98%, p-val < 0.001) and the area under the precision recall curve from 57.38-59.87% (95% CI = 3.86%-4.82%, p-val < 0.001) after integrating NLP-extracted SBDH. These findings underscore the potential of NLP-extracted SBDH in enhancing suicide prediction across various prediction timeframes, offering valuable insights for healthcare practitioners and policymakers.

Evaluating the clinical utility of an easily applicable prediction model of suicide attempts, newly developed and validated with a general community sample of adults

BACKGROUND

A suicide attempt (SA) is a clinically serious action. Researchers have argued that reducing long-term SA risk may be possible, provided that at-risk individuals are identified and receive adequate treatment. Algorithms may accurately identify at-risk individuals. However, the clinical utility of algorithmically estimated long-term SA risk has never been the predominant focus of any study.

METHODS

The data of this report stem from CoLaus|PsyCoLaus, a prospective longitudinal study of general community adults from Lausanne, Switzerland. Participants (N = 4,097; Mage = 54 years, range: 36-86; 54% female) were assessed up to four times, starting in 2003, approximately every 4-5 years. Long-term individual SA risk was prospectively predicted, using logistic regression. This algorithm's clinical utility was assessed by net benefit (NB). Clinical utility expresses a tool's benefit after having taken this tool's potential harm into account. Net benefit is obtained, first, by weighing the false positives, e.g., 400 individuals, at the risk threshold, e.g., 1%, using its odds (odds of 1% yields 1/(100-1) = 1/99), then by subtracting the result (400*1/99 = 4.04) from the true positives, e.g., 5 individuals (5-4.04), and by dividing the result (0.96) by the sample size, e.g., 800 (0.96/800). All results are based on 100 internal cross-validations. The predictors used in this study were: lifetime SA, any lifetime mental disorder, sex, and age.

RESULTS

SA at any of the three follow-up study assessments was reported by 1.2%. For a range of seven a priori selected threshold probabilities, ranging between 0.5% and 2%, logistic regression showed highest overall NB in 97.4% of all 700 internal cross-validations (100 for each selected threshold probability).

CONCLUSION

Despite the strong class imbalance of the outcome (98.8% no, 1.2% yes) and only four predictors, clinical utility was observed. That is, using the logistic regression model for clinical decision making provided the most true positives, without an increase of false positives, compared to all competing decision strategies. Clinical utility is one among several important prerequisites of implementing an algorithm in routine practice, and may possibly guide a clinicians' treatment decision making to reduce long-term individual SA risk. The novel metric NB may become a standard performance measure, because the a priori invested clinical considerations enable clinicians to interpret the results directly.

Predicting suicide death after emergency department visits with mental health or self-harm diagnoses

OBJECTIVE

Use health records data to predict suicide death following emergency department visits.

METHODS

Electronic health records and insurance claims from seven health systems were used to: identify emergency department visits with mental health or self-harm diagnoses by members aged 11 or older; extract approximately 2500 potential predictors including demographic, historical, and baseline clinical characteristics; and ascertain subsequent deaths by self-harm. Logistic regression with lasso and random forest models predicted self-harm death over 90 days after each visit.

RESULTS

Records identified 2,069,170 eligible visits, 899 followed by suicide death within 90 days. The best-fitting logistic regression with lasso model yielded an area under the receiver operating curve of 0.823 (95% CI 0.810-0.836). Visits above the 95th percentile of predicted risk included 34.8% (95% CI 31.1-38.7) of subsequent suicide deaths and had a 0.303% (95% CI 0.261-0.346) suicide death rate over the following 90 days. Model performance was similar across subgroups defined by age, sex, race, and ethnicity.

CONCLUSIONS

Machine learning models using coded data from health records have moderate performance in predicting suicide death following emergency department visits for mental health or self-harm diagnosis and could be used to identify patients needing more systematic follow-up.

Realizing the Potential of Social Determinants Data: A Scoping Review of Approaches for Screening, Linkage, Extraction, Analysis and Interventions

Background

Social determinants of health (SDoH) like socioeconomics and neighborhoods strongly influence outcomes, yet standardized SDoH data is lacking in electronic health records (EHR), limiting research and care quality.

Methods

We searched PubMed using keywords "SDOH" and "EHR", underwent title/abstract and full-text screening. Included records were analyzed under five domains: 1) SDoH screening and assessment approaches, 2) SDoH data collection and documentation, 3) Use of natural language processing (NLP) for extracting SDoH, 4) SDoH data and health outcomes, and 5) SDoH-driven interventions.

Results

We identified 685 articles, of which 324 underwent full review. Key findings include tailored screening instruments implemented across settings, census and claims data linkage providing contextual SDoH profiles, rule-based and neural network systems extracting SDoH from notes using NLP, connections found between SDoH data and healthcare utilization/chronic disease control, and integrated care management programs executed. However, considerable variability persists across data sources, tools, and outcomes.

Discussion

Despite progress identifying patient social needs, further development of standards, predictive models, and coordinated interventions is critical to fulfill the potential of SDoH-EHR integration. Additional database searches could strengthen this scoping review. Ultimately widespread capture, analysis, and translation of multidimensional SDoH data into clinical care is essential for promoting health equity.

Geographic Variation, Economic Activity, and Labor Market Characteristics in Trajectories of Suicide in the United States, 2008-2020

Suicide rates in the United States have increased over the past 15 years, with substantial geographic variation in these increases; yet there have been few attempts to cluster counties by the magnitude of suicide rate changes according to intercept and slope or to identify the economic precursors of increases. We used vital statistics data and growth mixture models to identify clusters of counties by their magnitude of suicide growth from 2008 to 2020 and examined associations with county economic and labor indices. Our models identified 5 clusters, each differentiated by intercept and slope magnitude, with the highest-rate cluster (4% of counties) being observed mainly in sparsely populated areas in the West and Alaska, starting the time series at 25.4 suicides per 100,000 population, and exhibiting the steepest increase in slope (0.69/100,000/year). There was no cluster for which the suicide rate was stable or declining. Counties in the highest-rate cluster were more likely to have agricultural and service economies and less likely to have urban professional economies. Given the increased burden of suicide, with no clusters of counties improving over time, additional policy and prevention efforts are needed, particularly targeted at rural areas in the West.

Is combined antidepressant medication (ADM) and psychotherapy better than either monotherapy at preventing suicide attempts and other psychiatric serious adverse events for depressed patients? A rare events meta-analysis

Antidepressant medication (ADM)-only, psychotherapy-only, and their combination are the first-line treatment options for major depressive disorder (MDD). Previous meta-analyses of randomized controlled trials (RCTs) established that psychotherapy and combined treatment were superior to ADM-only for MDD treatment remission or response. The current meta-analysis extended previous ones by determining the comparative efficacy of ADM-only, psychotherapy-only, and combined treatment on suicide attempts and other serious psychiatric adverse events (i.e. psychiatric emergency department [ED] visit, psychiatric hospitalization, and/or suicide death; SAEs). Peto odds ratios (ORs) and their 95% confidence intervals were computed from the present random-effects meta-analysis. Thirty-four relevant RCTs were included. Psychotherapy-only was stronger than combined treatment (1.9% v. 3.7%; OR 1.96 [1.20-3.20], p = 0.012) and ADM-only (3.0% v. 5.6%; OR 0.45 [0.30-0.67], p = 0.001) in decreasing the likelihood of SAEs in the primary and trim-and-fill sensitivity analyses. Combined treatment was better than ADM-only in reducing the probability of SAEs (6.0% v. 8.7%; OR 0.74 [0.56-0.96], p = 0.029), but this comparative efficacy finding was non-significant in the sensitivity analyses. Subgroup analyses revealed the advantage of psychotherapy-only over combined treatment and ADM-only for reducing SAE risk among children and adolescents and the benefit of combined treatment over ADM-only among adults. Overall, psychotherapy and combined treatment outperformed ADM-only in reducing the likelihood of SAEs, perhaps by conferring strategies to enhance reasons for living. Plausibly, psychotherapy should be prioritized for high-risk youths and combined treatment for high-risk adults with MDD.

Prediction models of suicide and non-fatal suicide attempt after discharge from a psychiatric inpatient stay: A machine learning approach on nationwide Danish registers

INTRODUCTION

To develop machine learning models capable of predicting suicide and non-fatal suicide attempt as separate outcomes in the first 30 days after discharge from a psychiatric inpatient stay.

METHODS

Prospective cohort study using nationwide Danish registry data. We included individuals who were 18 years or older, and all discharges from psychiatric hospitalizations in Denmark from 1995 to 2018. We trained predictive models using 10-fold cross validation on 80% of the data and did testing on the remaining 20%.

RESULTS

The best model for predicting non-fatal suicide attempt was an ensemble of predictions from gradient boosting (XGBoost) and categorical boosting (catBoost). The ROC-AUC for predicting suicide attempt was 0.85 (95% CI: 0.84-0.85). At a risk threshold of 4.36%, positive predictive value (PPV) was 11.0% and sensitivity was 47.2%. The best model for predicting suicide was an ensemble of predictions from random forest, XGBoost and catBoost. For suicide, the ROC-AUC was 0.71 (95% CI: 0.70-0.73). At a risk threshold of 0.15%, PPV was 0.34% and sensitivity was 56.0%. The most contributing predictors differed when predicting suicide and suicide attempt, indicating that separate models are needed. The ensemble model was fair across sex and age, and more so than the penalized logistic regression model.

CONCLUSIONS

We achieved good performance for predicting suicide attempts and demonstrated a clinical application of ensemble models. Our results indicate a difference in predictive performance for models predicting suicide and suicide attempt, respectively. Thus, we recommend that suicide and suicide attempt are treated as two separate endpoints, in particular for clinical application. We demonstrated that the ensemble model is fairer across sex and age compared with a penalized logistic regression, and therefore we recommend the use of well-tested ensembles despite a more complex explainability.

Probing the Association between Cognition, Suicidal Behavior and Tryptophan Metabolism in a Sample of Individuals Living with Bipolar Disorder: A Secondary Analysis

Background and Objectives: Alterations in hot cognition and in the tryptophan metabolism through serotonin (5-HT) and kynurenine (KYN) pathways have been associated with an increased risk of suicidal behavior. Here, we aim at probing the association between Stroop test performances and tryptophan pathway components in a sample of individuals with bipolar disorder (BD). Materials and Methods: We explored the association between the Emotion Inhibition Subtask (EIS) performances of the Brief Assessment of Cognition for Affective Disorders (BAC-A) and plasmatic levels of 5-hydroxytriptophan (5-HTP), 5-HT, KYN, 3-hydroxykynurenine (3-HK), quinolinic acid (QA), and kynurenic acid (KYNA) among subjects reporting lifetime suicide ideation (LSI) vs. non-LSI and subjects reporting lifetime suicide attempts (LSA) vs. non-LSA. Results: In a sample of 45 subjects with BD, we found a statistically significant different performance for LSA vs. non-LSA in the color naming (CN) and neutral words (NW) EIS subtasks. There was a significant association between CN performances and plasma 5-HTP levels among LSI and LSA subjects but not among non-LSI or non-LSA. Conclusions: In our sample, patients with LSA and LSI presented lower performances on some EIS subtasks compared to non-LSA and non-LSI. Moreover, we found an inverse correlation between plasma 5-HTP concentration and some EIS performances in LSA and LSI but not among non-LSA or non-LSI. This may represent an interesting avenue for future studies probing this complex association.