Artificial intelligence and suicide prevention: a systematic review

Evaluating the ability of artificial intelligence to predict suicide: A systematic review of reviews

INTRODUCTION

Suicide remains a critical global public health issue, with approximately 800,000 deaths annually. Despite various prevention efforts, suicide rates are rising, highlighting the need for more effective strategies. Traditional suicide risk assessment methods often fall short in accuracy and predictive capability. This has driven interest in artificial intelligence (AI), particularly machine learning (ML), as a potential solution. This paper reviews systematic evaluations of AI's effectiveness in predicting suicide risk, aiming to explore AI's potential while addressing its challenges and limitations.

METHODOLOGY

A meta-research approach was used to review existing systematic reviews on AI's role in suicide risk prediction. Following PRISMA guidelines, a comprehensive search was conducted in PubMed and Web of Science for publications from 2004 to 2024. Relevant studies were selected based on specific inclusion criteria, and data were extracted on review characteristics, AI techniques, outcomes, and methodological quality. The review focuses on AI/ML models predicting suicidal ideation (SI), suicide attempts (SA), and suicide deaths (SD) separately, excluding non-suicidal self-injury.

RESULTS

Out of 96 initial articles, 23 met the inclusion criteria for full-text review. Most studies focused on developing ML models to identify suicide risk, showing promising results in enhancing accuracy and effectiveness. These models utilize various data sources and analytical techniques. However, challenges remain, including high bias risk and issues with interpretability, which necessitate further validation and refinement of AI-driven methods.

CONCLUSION

The review underscores the significant potential of AI, especially ML, in predicting suicide risk and attempts. Although ML models show promise, challenges like data limitations, bias, and interpretability issues need addressing. Continued research and ethical scrutiny are crucial to fully realize AI's potential in suicide prevention.

AI-based personalized real-time risk prediction for behavioral management in psychiatric wards using multimodal data

BACKGROUND

Suicide is a major global health issue, with approximately 700,000 deaths annually (WHO). In psychiatric wards, managing harmful behaviors such as suicide, self-harm, and aggression is essential to ensure patient and staff safety. However, psychiatric wards in South Korea face challenges due to high patient-to-psychiatrist ratios and heavy workloads. Current models relying on demographic data struggle to provide real-time predictions. This study introduces the Temporal Fusion Transformer (TFT) model to address these limitations by integrating sensor, location, and clinical data for predicting harmful behaviors. The TFT model's advanced features, such as Variable Selection Networks and temporal attention mechanisms, make it particularly suitable for capturing complex time-series patterns and providing interpretable results in psychiatric settings.

METHODS

Data from 145 patients across three hospitals were collected using wearable devices that tracked heart rate, movement, and location. The data were aggregated hourly, preprocessed to handle missing values, and standardized. A binary classification model using TFT was developed and evaluated with accuracy, recall, F1 score, and AUC. Bayesian optimization was employed for hyperparameter tuning, and 5-fold cross-validation was performed to ensure generalizability.

RESULTS

The TFT model outperformed Multi-LSTM and Multi-GRU models, achieving 95.1% accuracy, 74.9% recall, an F1 score of 78.1, and an AUC of 0.863. The Variable Selection Network effectively identified key predictive factors, such as daily entropy and heart rate variability, improving interpretability. Incorporating location and biometric data enhanced prediction accuracy and enabled real-time risk assessments.

CONCLUSION

This study is the first to use the TFT model for predicting behavioral risks in psychiatric wards. The model's ability to integrate diverse data sources, prioritize cirtical variables, and capture temporal dependencies make it highly suitable for psychiatric environments. While the TFT model performed well, challenges remain with recall due to the limited dataset. Future research will focus on expanding datasets, optimizing variable selection, and standardizing data through a multimodal Common Data Model (CDM) to further improve performance and clinical utility.

The Application of AI to Ecological Momentary Assessment Data in Suicide Research: Systematic Review

BACKGROUND

Ecological momentary assessment (EMA) captures dynamic processes suitable to the study of suicidal ideation and behaviors. Artificial intelligence (AI) has increasingly been applied to EMA data in the study of suicidal processes.

OBJECTIVE

This review aims to (1) synthesize empirical research applying AI strategies to EMA data in the study of suicidal ideation and behaviors; (2) identify methodologies and data collection procedures used, suicide outcomes studied, AI applied, and results reported; and (3) develop a standardized reporting framework for researchers applying AI to EMA data in the future.

METHODS

PsycINFO, PubMed, Scopus, and Embase were searched for published articles applying AI to EMA data in the investigation of suicide outcomes. The PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines were used to identify studies while minimizing bias. Quality appraisal was performed using CREMAS (adapted STROBE [Strengthening the Reporting of Observational Studies in Epidemiology] Checklist for Reporting Ecological Momentary Assessment Studies).

RESULTS

In total, 1201 records were identified across databases. After a full-text review, 12 (1%) articles, comprising 4398 participants, were included. In the application of AI to EMA data to predict suicidal ideation, studies reported mean area under the curve (0.74-0.86), sensitivity (0.64-0.81), specificity (0.73-0.86), and positive predictive values (0.72-0.77). Studies met between 4 and 13 of the 16 recommended CREMAS reporting standards, with an average of 7 items met across studies. Studies performed poorly in reporting EMA training procedures and treatment of missing data.

CONCLUSIONS

Findings indicate the promise of AI applied to self-report EMA in the prediction of near-term suicidal ideation. The application of AI to EMA data within suicide research is a burgeoning area hampered by variations in data collection and reporting procedures. The development of an adapted reporting framework by the research team aims to address this.

TRIAL REGISTRATION

Open Science Framework (OSF); https://doi.org/10.17605/OSF.IO/NZWUJ and PROSPERO CRD42023440218; https://www.crd.york.ac.uk/PROSPERO/view/CRD42023440218.

An Explainable Artificial Intelligence Text Classifier for Suicidality Prediction in Youth Crisis Text Line Users: Development and Validation Study

BACKGROUND

Suicide represents a critical public health concern, and machine learning (ML) models offer the potential for identifying at-risk individuals. Recent studies using benchmark datasets and real-world social media data have demonstrated the capability of pretrained large language models in predicting suicidal ideation and behaviors (SIB) in speech and text.

OBJECTIVE

This study aimed to (1) develop and implement ML methods for predicting SIBs in a real-world crisis helpline dataset, using transformer-based pretrained models as a foundation; (2) evaluate, cross-validate, and benchmark the model against traditional text classification approaches; and (3) train an explainable model to highlight relevant risk-associated features.

METHODS

We analyzed chat protocols from adolescents and young adults (aged 14-25 years) seeking assistance from a German crisis helpline. An ML model was developed using a transformer-based language model architecture with pretrained weights and long short-term memory layers. The model predicted suicidal ideation (SI) and advanced suicidal engagement (ASE), as indicated by composite Columbia-Suicide Severity Rating Scale scores. We compared model performance against a classical word-vector-based ML model. We subsequently computed discrimination, calibration, clinical utility, and explainability information using a Shapley Additive Explanations value-based post hoc estimation model.

RESULTS

The dataset comprised 1348 help-seeking encounters (1011 for training and 337 for testing). The transformer-based classifier achieved a macroaveraged area under the curve (AUC) receiver operating characteristic (ROC) of 0.89 (95% CI 0.81-0.91) and an overall accuracy of 0.79 (95% CI 0.73-0.99). This performance surpassed the word-vector-based baseline model (AUC-ROC=0.77, 95% CI 0.64-0.90; accuracy=0.61, 95% CI 0.61-0.80). The transformer model demonstrated excellent prediction for nonsuicidal sessions (AUC-ROC=0.96, 95% CI 0.96-0.99) and good prediction for SI and ASE, with AUC-ROCs of 0.85 (95% CI 0.97-0.86) and 0.87 (95% CI 0.81-0.88), respectively. The Brier Skill Score indicated a 44% improvement in classification performance over the baseline model. The Shapley Additive Explanations model identified language features predictive of SIBs, including self-reference, negation, expressions of low self-esteem, and absolutist language.

CONCLUSIONS

Neural networks using large language model-based transfer learning can accurately identify SI and ASE. The post hoc explainer model revealed language features associated with SI and ASE. Such models may potentially support clinical decision-making in suicide prevention services. Future research should explore multimodal input features and temporal aspects of suicide risk.

Applications of Large Language Models in the Field of Suicide Prevention: Scoping Review

BACKGROUND

Prevention of suicide is a global health priority. Approximately 800,000 individuals die by suicide yearly, and for every suicide death, there are another 20 estimated suicide attempts. Large language models (LLMs) hold the potential to enhance scalable, accessible, and affordable digital services for suicide prevention and self-harm interventions. However, their use also raises clinical and ethical questions that require careful consideration.

OBJECTIVE

This scoping review aims to identify emergent trends in LLM applications in the field of suicide prevention and self-harm research. In addition, it summarizes key clinical and ethical considerations relevant to this nascent area of research.

METHODS

Searches were conducted in 4 databases (PsycINFO, Embase, PubMed, and IEEE Xplore) in February 2024. Eligible studies described the application of LLMs for suicide or self-harm prevention, detection, or management. English-language peer-reviewed articles and conference proceedings were included, without date restrictions. Narrative synthesis was used to synthesize study characteristics, objectives, models, data sources, proposed clinical applications, and ethical considerations. This review adhered to the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) standards.

RESULTS

Of the 533 studies identified, 36 (6.8%) met the inclusion criteria. An additional 7 studies were identified through citation chaining, resulting in 43 studies for review. The studies showed a bifurcation of publication fields, with varying publication norms between computer science and mental health. While most of the studies (33/43, 77%) focused on identifying suicide risk, newer applications leveraging generative functions (eg, support, education, and training) are emerging. Social media was the most common source of LLM training data. Bidirectional Encoder Representations from Transformers (BERT) was the predominant model used, although generative pretrained transformers (GPTs) featured prominently in generative applications. Clinical LLM applications were reported in 60% (26/43) of the studies, often for suicide risk detection or as clinical assistance tools. Ethical considerations were reported in 33% (14/43) of the studies, with privacy, confidentiality, and consent strongly represented.

CONCLUSIONS

This evolving research area, bridging computer science and mental health, demands a multidisciplinary approach. While open access models and datasets will likely shape the field of suicide prevention, documenting their limitations and potential biases is crucial. High-quality training data are essential for refining these models and mitigating unwanted biases. Policies that address ethical concerns-particularly those related to privacy and security when using social media data-are imperative. Limitations include high variability across disciplines in how LLMs and study methodology are reported. The emergence of generative artificial intelligence signals a shift in approach, particularly in applications related to care, support, and education, such as improved crisis care and gatekeeper training methods, clinician copilot models, and improved educational practices. Ongoing human oversight-through human-in-the-loop testing or expert external validation-is essential for responsible development and use.

TRIAL REGISTRATION

OSF Registries osf.io/nckq7; https://osf.io/nckq7.

Harnessing digital health data for suicide prevention and care: A rapid review

Background and aim

Suicide is a global public health issue disproportionately impacting equity-deserving groups. Recent advances in Artificial Intelligence and increased access to a variety of digital data sources have enabled the development of novel and personalized suicide prevention strategies. However, standards on how to harness these data in a comprehensive and equitable way remain unclear. The primary aim of this study is to identify considerations for the collection and use of digital health data for suicide prevention and care. The results will inform the development of a data governance framework for a multinational suicide prevention mHealth platform.

Method

We used a modified Cochrane Rapid Reviews Method. Inclusion criteria focused on primary studies published in English from 2007 to the present that referenced the use of digital health data in the context of suicide prevention and care. Screening and data extraction was performed independently by multiple reviewers, with disagreements resolved through discussion. Qualitative and quantitative synthesis methods were employed to identify emergent themes.

Results

Our search identified 2453 potential articles, with 70 meeting inclusion criteria. We found little consensus on best practices for the collection and use of digital health data for suicide prevention and care. Issues of data quality, fairness and equity persist, compounded by inadequate consideration of key governance issues including privacy and trust, especially in multinational initiatives.

Conclusions

Recommendations for future research and practice include prioritizing engagement with knowledge users, establishing robust data governance frameworks aligned with clinical guidelines, and leveraging advanced analytics, such as natural language processing, to improve the quality of health equity data.

Artificial intelligence and machine learning techniques for suicide prediction: Integrating dietary patterns and environmental contaminants

Background

Suicide remains a leading cause of death globally, with nearly 800,000 deaths annually, particularly among young adults in regions like Europe, Australia, and the Middle East, highlighting the urgent need for innovative intervention strategies beyond conventional methods.

Objectives

This review aims to explore the transformative role of artificial intelligence (AI) and machine learning (ML) in enhancing suicide risk prediction and developing effective prevention strategies, examining how these technologies integrate complex risk factors, including psychiatric, socio-economic, dietary, and environmental influences.

Methods

A comprehensive review of literature from databases such as PubMed and Web of Science was conducted, focusing on studies that utilize AI and ML technologies. The review assessed the efficacy of various models, including Random Forest, neural networks, and others, in analyzing data from electronic health records, social media, and digital behaviors. Additionally, it evaluated a broad spectrum of dietary factors and their influence on suicidal behaviors, as well as the impact of environmental contaminants like lithium, arsenic, fluoride, mercury, and organophosphorus pesticides.

Conclusions

AI and ML are revolutionizing suicide prevention strategies, with models achieving nearly 90 % predictive accuracy by integrating diverse data sources. Our findings highlight the need for geographically and demographically tailored public health interventions and comprehensive AI models that address the multifactorial nature of suicide risk. However, the deployment of these technologies must address critical ethical and privacy concerns, ensuring compliance with regulations and the development of transparent, ethically guided AI systems. AI-driven tools, such as virtual therapists and chatbots, are essential for immediate support, particularly in underserved regions.

Evaluating of BERT-based and Large Language Mod for Suicide Detection, Prevention, and Risk Assessment: A Systematic Review

Suicide constitutes a public health issue of major concern. Ongoing progress in the field of artificial intelligence, particularly in the domain of large language models, has played a significant role in the detection, risk assessment, and prevention of suicide. The purpose of this review was to explore the use of LLM tools in various aspects of suicide prevention. PubMed, Embase, Web of Science, Scopus, APA PsycNet, Cochrane Library, and IEEE Xplore-for studies published were systematically searched for articles published between January 1, 2018, until April 2024. The 29 reviewed studies utilized LLMs such as GPT, Llama, and BERT. We categorized the studies into three main tasks: detecting suicidal ideation or behaviors, assessing the risk of suicidal ideation, and preventing suicide by predicting attempts. Most of the studies demonstrated that these models are highly efficient, often outperforming mental health professionals in early detection and prediction capabilities. Large language models demonstrate significant potential for identifying and detecting suicidal behaviors and for saving lives. Nevertheless, ethical problems still need to be examined and cooperation with skilled professionals is essential.

Prevalence, Patterns, and Nomogram Model of Intimate Partner Violence Against Mothers During the Child-Rearing Stage: A Family System Analysis in China for Targeted Prevention Strategies

Intimate partner violence (IPV) is a significant public health issue affecting many women worldwide. While extensive research exists on IPV during pregnancy and postpartum, there is limited information on IPV against mothers during the critical child-rearing stage, specifically the first three years following childbirth. This study examines the prevalence and patterns of IPV among mothers in China during this stage, identifying associated factors across four family subsystems: individual, husband-and-wife, mother-child, and family context, to guide the development of tailored prevention strategies. This study involved 1,099 Chinese mothers, surveyed within the first three years postpartum, through purposive sampling. The revised Conflict Tactics Scale (CTS2) was utilized to evaluate IPV, while a comprehensive questionnaire gathered data on potential risk and protective factors within the four family subsystems. Chi-square tests and lasso regression analyses were used to identify significant independent risk factors, which were used to construct nomograms of IPV among mothers during the critical child-rearing stage. The nomogram's discrimination, calibration, clinical applicability, and generalizability were evaluated using receiver operating characteristic curves (ROC), calibration curves, decision curve analysis (DCA), and internal validation. Approximately 30% of mothers had experienced IPV within three years postpartum, with psychological violence being the most common. Three main patterns of IPV were identified, with multiple forms of violence often co-occurring. Significant risk factors for IPV included age at childbirth, attachment styles, marital issues, marital stability, feeding choices, maternal sense of parenting competence, support from friends, and family stress events. A nomogram model was developed to identify associated factors of IPV, demonstrating good performance. This model integrates factors from individual, spousal, mother-child, and family context subsystems, providing a comprehensive approach to understanding and preventing IPV during the critical child-rearing stage. The high prevalence of IPV underscores the urgent need for targeted prevention strategies to support mothers during this vulnerable period.

How do machine learning models perform in the detection of depression, anxiety, and stress among undergraduate students? A systematic review

Undergraduate students are often impacted by depression, anxiety, and stress. In this context, machine learning may support mental health assessment. Based on the following research question: "How do machine learning models perform in the detection of depression, anxiety, and stress among undergraduate students?", we aimed to evaluate the performance of these models. PubMed, Embase, PsycINFO, and Web of Science databases were searched, aiming at studies meeting the following criteria: publication in English; targeting undergraduate university students; empirical studies; having been published in a scientific journal; and predicting anxiety, depression, or stress outcomes via machine learning. The certainty of evidence was analyzed using the GRADE. As of January 2024, 2,304 articles were found, and 48 studies met the inclusion criteria. Different types of data were identified, including behavioral, physiological, internet usage, neurocerebral, blood markers, mixed data, as well as demographic and mobility data. Among the 33 studies that provided accuracy assessment, 30 reported values that exceeded 70%. Accuracy in detecting stress ranged from 63% to 100%, anxiety from 53.69% to 97.9%, and depression from 73.5% to 99.1%. Although most models present adequate performance, it should be noted that 47 of them only performed internal validation, which may overstate the performance data. Moreover, the GRADE checklist suggested that the quality of the evidence was very low. These findings indicate that machine learning algorithms hold promise in Public Health; however, it is crucial to scrutinize their practical applicability. Further studies should invest mainly in external validation of the machine learning models.

Suicide among lymphoma patients

BACKGROUND

Higher suicide rates were observed in patients diagnosed with lymphoma. In this study, we accurately identified patients with high-risk lymphoma for suicide by constructing a nomogram with a view to effective interventions and reducing the risk of suicide.

METHODS

235,806 patients diagnosed with lymphoma between 2000 and 2020 were picked from the Surveillance, Epidemiology, and End Results (SEER) database and randomly divided into training (N = 165,064) and validation set (N = 70,742). A combination of the Least absolute shrinkage and selection operator (LASSO) and Cox proportional hazards regression identified the predictors that constructed the nomogram. To assess the discrimination, calibration, clinical applicability, and generalization of this nomogram, we implemented receiver operating characteristic curves (ROC), calibration curves, decision curve analysis (DCA), and internal validation. The robustness of the results was assessed by the competing risks regression model.

RESULTS

Age at diagnosis, gender, ethnicity, marital status, stage, surgery, radiotherapy, and annual household income were key predictors of suicide in lymphoma patients. A nomogram was created to visualize the risk of suicide after a lymphoma diagnosis. The c-index for the training set was 0.773, and the validation set was 0.777. The calibration curve for the nomogram fitted well with the diagonal and the clinical decision curve indicated its clinical benefit.

LIMITATION

The effects of unmeasured and unnoticed biases and confounders were difficult to eliminate due to retrospective studies.

CONCLUSION

A convenient and reliable model has been constructed that will help to individualize and accurately quantify the risk of suicide in patients diagnosed with lymphoma.

Can profiles of behaviors occurring within 48 h of a suicide attempt predict future severity of suicidal thoughts and reattempt?: An examination of hospitalized patients 12 Months post-discharge

Behavioral warning signs (WS) are near-term changes within individuals, which aid in determining imminent risk for suicide attempts. However, those who attempt suicide differ in their engagement of WS, and it is unclear if these differences relate to future risk of suicidal behavior. Using a sample of 132 adults presenting to a hospital following a suicide attempt, the current study sought to determine if differences in engagement in WS for the index attempt prospectively predicted suicide attempt, frequency of ideation, and intensity of suicide ideation 12 months post discharge. Latent class analyses (LCAs) conducted on 6 behaviors (i.e., alcohol use, nightmares, interpersonal negative life events, suicide communication, risky behavior, low sleep, and high sleep) found a 5-class solution optimally fit the data. One identified class, characterized by engagement in risky behaviors the hours before an attempt differed from other identified classes in terms of risk for future suicidal ideation and behaviors. More specifically, participants in "High Risky Behavior" class had higher rates of 12-month suicide reattempt, significantly more frequent suicide ideation, and significantly worse intensity of suicide ideation during the 12 months following their index attempt compared to participants endorsing typical patterns of WS. These results held when adjusting for various traditional baseline covariates (e.g., depressive symptoms). The current study demonstrates that patterns of behavioral WS may be utilized as their own prognostic indicator of future suicidal ideation and behaviors among high-risk individuals reporting a recent suicide attempt, which can inform post-discharge clinical intervention and prevention efforts.

Big Data in Health Care: An Interprofessional Course

BACKGROUND

The widespread adoption of the electronic health record (EHR) has resulted in vast repositories of EHR big data that are being used to identify patterns and correlations that translate into data-informed health care decision making.

PROBLEM

Health care professionals need the skills necessary to navigate a digitized, data-rich health care environment as big data plays an increasingly integral role in health care.

APPROACH

Faculty incorporated the concept of big data in an asynchronous online course allowing an interprofessional mix of students to analyze EHR big data on over a million patients.

OUTCOMES

Students conducted a descriptive analysis of cohorts of patients with selected diagnoses and presented their findings.

CONCLUSIONS

Students collaborated with an interprofessional team to analyze EHR big data on selected variables. The teams used data visualization tools to describe an assigned diagnosis patient population.

Suicide prevention in Hong Kong: pushing boundaries while building bridges

Hong Kong is a natural laboratory for studying suicides-small geographic footprint, bustling economic activity, rapidly changing socio-demographic transitions, and cultural crossroads. Its qualities also intensify the challenges posed when seeking to prevent them. In this viewpoint, we showed the research and practices of suicide prevention efforts made by the Hong Kong Jockey Club Centre for Suicide Research and Prevention (CSRP), which provide the theoretical underpinning of suicide prevention and empirical evidence. CSRP adopted a multi-level public health approach (universal, selective and indicated), and has collaboratively designed, implemented, and evaluated numerous programs that have demonstrated effectiveness in suicide prevention and mental well-being promotion. The center serves as a hub and a catalyst for creating, identifying, deploying, and evaluating suicide prevention initiatives, which have the potential to reduce regional suicides rates when taken to scale and sustained.

Exploring the Potential of Artificial Intelligence in Adolescent Suicide Prevention: Current Applications, Challenges, and Future Directions

ObjectiveThe global surge in adolescent suicide necessitates the development of innovative and efficacious preventive measures. Traditionally, various approaches have been used, but with limited success. However, with the rapid advancements in artificial intelligence (AI), new possibilities have emerged. This paper reviews the potentials and challenges of integrating AI into suicide prevention strategies, focusing on adolescents. Method: This narrative review assesses the impact of AI on suicide prevention strategies, the strategies and cases of AI applications in adolescent suicide prevention, as well as the challenges faced. Through searches on the PubMed, web of science, PsycINFO, and EMBASE databases, 19 relevant articles were included in the review. Results: AI has significantly improved risk assessment and predictive modeling for identifying suicidal behavior. It has enabled the analysis of textual data through natural language processing and fostered novel intervention strategies. Although AI applications, such as chatbots and monitoring systems, show promise, they must navigate challenges like data privacy and ethical considerations. The research underscores the potential of AI to enhance future suicide prevention efforts through personalized interventions and integration with emerging technologies. Conclusion: AI possesses transformative potential for adolescent suicide prevention by offering targeted and adaptive solutions, while they also raise crucial ethical and practical considerations. Looking forward, AI can play a critical role in mitigating adolescent suicide rates, marking a new frontier in mental health care.

A shift in psychiatry through AI? Ethical challenges

The digital transformation has made its way into many areas of society, including medicine. While AI-based systems are widespread in medical disciplines, their use in psychiatry is progressing more slowly. However, they promise to revolutionize psychiatric practice in terms of prevention options, diagnostics, or even therapy. Psychiatry is in the midst of this digital transformation, so the question is no longer "whether" to use technology, but "how" we can use it to achieve goals of progress or improvement. The aim of this article is to argue that this revolution brings not only new opportunities but also new ethical challenges for psychiatry, especially with regard to safety, responsibility, autonomy, or transparency. As an example, the relationship between doctor and patient in psychiatry will be addressed, in which digitization is also leading to ethically relevant changes. Ethical reflection on the use of AI systems offers the opportunity to accompany these changes carefully in order to take advantage of the benefits that this change brings. The focus should therefore always be on balancing what is technically possible with what is ethically necessary.

Building an Integrated Data Infrastructure to Examine the Spectrum of Suicide Risk Factors in Philadelphia Medicaid

While there are many data-driven approaches to identifying individuals at risk of suicide, they tend to focus on clinical risk factors, such as previous psychiatric hospitalizations, and rarely include risk factors that occur in nonclinical settings, such as jails or emergency shelters. A better understanding of system-level encounters by individuals at risk of suicide could help inform suicide prevention efforts. In Philadelphia, we built a community-level data infrastructure that encompassed suicide death records, behavioral health claims, incarceration episodes, emergency housing episodes, and involuntary commitment petitions to examine a broader spectrum of suicide risk factors. Here, we describe the development of the data infrastructure, present key trends in suicide deaths in Philadelphia, and, for the Medicaid-eligible population, determine whether suicide decedents were more likely to interact with the behavioral health, carceral, and housing service systems compared to Medicaid-eligible Philadelphians who did not die by suicide. Between 2003 and 2018, there was an increase in the number of annual suicide deaths among Medicaid-eligible individuals, in part due to changes in Medicaid eligibility. There were disproportionately more suicide deaths among Black and Hispanic individuals who were Medicaid-eligible, who were younger on average, compared to suicide decedents who were never Medicaid-eligible. However, when we accounted for the racial and ethnic composition of the Medicaid population at large, we found that White individuals were four times as likely to die by suicide, while Asian, Black, Hispanic, and individuals of other races were less likely to die by suicide. Overall, 58% of individuals who were Medicaid-eligible and died by suicide had at least one Medicaid-funded behavioral health claim, 10% had at least one emergency housing episode, 25% had at least one incarceration episode, and 22% had at least one involuntary commitment. By developing a data infrastructure that can incorporate a broader spectrum of risk factors for suicide, we demonstrate how communities can harness administrative data to inform suicide prevention efforts. Our findings point to the need for suicide prevention in nonclinical settings such as jails and emergency shelters, and demonstrate important trends in suicide deaths in the Medicaid population.

Artificial intelligence-based approaches for suicide prediction: Hope or hype?

Accurate prediction of suicide risk is important because it allows evidence-based interventions to be targeted to at-risk populations. Conventional approaches to prediction of suicide risk have shown suboptimal accuracy. In this context, artificial intelligence (AI)-based models for suicide prediction, with their ability to handle big data collected through low-burden methods, have gained research traction in the recent past. Preliminary evidence suggests the promise of AI-driven methods for suicide prediction and prevention. These methods may hold particular relevance for India and other low-and-middle-income countries because they may be more feasible, scalable, less stigmatizing, and potentially more resource-effective.

The Role of Virtual Triage in Improving Clinician Experience and Satisfaction: A Narrative Review

Objective

This review examines the literature on improving clinician satisfaction with a focus on what has been most effective in improving experience from the perspective of clinicians, and the potential role that virtual triage (VT) technology can play in delivering positive clinician experiences that improve clinical care, and bring value to health care delivery organizations (HDOs).

Methods

Review and synthesis of evidence on clinician satisfaction indicating a potential for VT to favorably impact clinician experience, sense of effectiveness, efficiency, and reduction of administrative task burden. Analysis considers how to conceptualize and the value of improving clinician experience, leading clinician dissatisfiers, and the potential role of VT in improving clinician experience/satisfaction.

Results

Contributors to poor clinician experience/satisfaction where VT could have a beneficial impact include better managing resource limitations, administrative workload, lack of care coordination, information overload, and payer interactions. VT can improve clinician experience through the technology's ability to leverage real-time actionable data clinicians can use, streamlining patient-clinician communications, personalizing care delivery, optimizing care coordination, and better aligning digital/virtual services with clinical practice. From an organizational perspective, improvements in clinician experience and satisfaction derive from establishing an effective digital back door, increasing the clinical impact of and satisfaction derived from telemedicine and virtual care, and enhancing clinician centricity.

Conclusions

By embracing digital transformation and implementing solutions such as VT that focus on improving patient and clinician experience, HDOs can address barriers to delivery of high-quality, efficient, and cost-effective care. VT is a digital health tool that can create a more streamlined and satisfying experience for clinicians and the patients they care for. VT is a technology solution that can help clinicians make faster more informed decisions, reduces avoidable care, improves communication with patients and within care teams, and lowers their administrative burden so they have more quality time to care for patients.

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

Importance

The months after psychiatric hospital discharge are a time of high risk for suicide. Intensive postdischarge case management, although potentially effective in suicide prevention, is likely to be cost-effective only if targeted at high-risk patients. A previously developed machine learning (ML) model showed that postdischarge suicides can be predicted from electronic health records and geospatial data, but it is unknown if prediction could be improved by adding additional information.

Objective

To determine whether model prediction could be improved by adding information extracted from clinical notes and public records.

Design, Setting, and Participants

Models were trained to predict suicides in the 12 months after Veterans Health Administration (VHA) short-term (less than 365 days) psychiatric hospitalizations between the beginning of 2010 and September 1, 2012 (299 050 hospitalizations, with 916 hospitalizations followed within 12 months by suicides) and tested in the hospitalizations from September 2, 2012, to December 31, 2013 (149 738 hospitalizations, with 393 hospitalizations followed within 12 months by suicides). Validation focused on net benefit across a range of plausible decision thresholds. Predictor importance was assessed with Shapley additive explanations (SHAP) values. Data were analyzed from January to August 2022.

Main Outcomes and Measures

Suicides were defined by the National Death Index. Base model predictors included VHA electronic health records and patient residential data. The expanded predictors came from natural language processing (NLP) of clinical notes and a social determinants of health (SDOH) public records database.

Results

The model included 448 788 unique hospitalizations. Net benefit over risk horizons between 3 and 12 months was generally highest for the model that included both NLP and SDOH predictors (area under the receiver operating characteristic curve range, 0.747-0.780; area under the precision recall curve relative to the suicide rate range, 3.87-5.75). NLP and SDOH predictors also had the highest predictor class-level SHAP values (proportional SHAP = 64.0% and 49.3%, respectively), although the single highest positive variable-level SHAP value was for a count of medications classified by the US Food and Drug Administration as increasing suicide risk prescribed the year before hospitalization (proportional SHAP = 15.0%).

Conclusions and Relevance

In this study, clinical notes and public records were found to improve ML model prediction of suicide after psychiatric hospitalization. The model had positive net benefit over 3-month to 12-month risk horizons for plausible decision thresholds. Although caution is needed in inferring causality based on predictor importance, several key predictors have potential intervention implications that should be investigated in future studies.

Establishment and validation of a nomogram for suicidality in Chinese secondary school students

BACKGROUND

Accurately identifying high-risk of suicide groups and conducting appropriate interventions are important to reduce the risk of suicide. In this study, a nomogram technique was used to develop a predictive model for the suicidality of secondary school students based on four aspects: individual characteristics; health risk behaviors; family factors; and school factors.

METHODS

A total of 9338 secondary school students were surveyed using the stratified cluster sampling method, and subjects were randomly divided into a training set (n = 6366) and a validation set (n = 2728). In the former, the results of the lasso regression and random forest were combined, from which 7 optimal predictors of suicidality were determined. These were used to construct a nomogram. This nomogram's discrimination, calibration, clinical applicability, and generalization were assessed using receiver operating characteristic curves (ROC), calibration curves, decision curve analysis (DCA), and internal validation.

RESULTS

Gender, depression symptoms, self-injury, running away from home, parents' relationship, relationship with father, and academic stress were found to be significant predictors of suicidality. The area under the curve (AUC) of the training set was 0.806, while that of the validation data was 0.792. The calibration curve of the nomogram was close to the diagonal, and the DCA showed the nomogram was clinically beneficial across a range of thresholds of 9-89 %.

LIMITATIONS

Causal inference is limited due to the cross-sectional design.

CONCLUSION

An effective tool was constructed for predicting suicidality among secondary school students, which should help school healthcare personnel assess this information about students and also identify high-risk groups.

Use of social media data to diagnose and monitor psychotic disorders: systematic review and perspectives (Preprint)

Background

Schizophrenia is a disease associated with high burden, and improvement in care is necessary. Artificial intelligence (AI) has been used to diagnose several medical conditions as well as psychiatric disorders. However, this technology requires large amounts of data to be efficient. Social media data could be used to improve diagnostic capabilities.

Objective

The objective of our study is to analyze the current capabilities of AI to use social media data as a diagnostic tool for psychotic disorders.

Methods

A systematic review of the literature was conducted using several databases (PubMed, Embase, Cochrane, PsycInfo, and IEEE Xplore) using relevant keywords to search for articles published as of November 12, 2021. We used the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) criteria to identify, select, and critically assess the quality of the relevant studies while minimizing bias. We critically analyzed the methodology of the studies to detect any bias and presented the results.

Results

Among the 93 studies identified, 7 studies were included for analyses. The included studies presented encouraging results. Social media data could be used in several ways to care for patients with schizophrenia, including the monitoring of patients after the first episode of psychosis. We identified several limitations in the included studies, mainly lack of access to clinical diagnostic data, small sample size, and heterogeneity in study quality. We recommend using state-of-the-art natural language processing neural networks, called language models, to model social media activity. Combined with the synthetic minority oversampling technique, language models can tackle the imbalanced data set limitation, which is a necessary constraint to train unbiased classifiers. Furthermore, language models can be easily adapted to the classification task with a procedure called “fine-tuning.”

Conclusions

The use of social media data for the diagnosis of psychotic disorders is promising. However, most of the included studies had significant biases; we therefore could not draw conclusions about accuracy in clinical situations. Future studies need to use more accurate methodologies to obtain unbiased results.