Suicide Risk Detection Research Articles

Detection of suicidality from medical text using privacy-preserving large language models.

Attempts to use artificial intelligence (AI) in psychiatric disorders show moderate success, highlighting the potential of incorporating information from clinical assessments to improve the models. This study focuses on using large language models (LLMs) to detect suicide risk from medical text in psychiatric care. To extract information about suicidality status from the admission notes in electronic health records (EHRs) using privacy-sensitive, locally hosted LLMs, specifically evaluating the efficacy of Llama-2 models. We compared the performance of several variants of the open source LLM Llama-2 in extracting suicidality status from 100 psychiatric reports against a ground truth defined by human experts, assessing accuracy, sensitivity, specificity and F1 score across different prompting strategies. A German fine-tuned Llama-2 model showed the highest accuracy (87.5%), sensitivity (83.0%) and specificity (91.8%) in identifying suicidality, with significant improvements in sensitivity and specificity across various prompt designs. The study demonstrates the capability of LLMs, particularly Llama-2, in accurately extracting information on suicidality from psychiatric records while preserving data privacy. This suggests their application in surveillance systems for psychiatric emergencies and improving the clinical management of suicidality by improving systematic quality control and research.

Suicide Ideation and Risk Detection from Social Media Using GPT Models

Suicide Ideation and Risk Detection from Social Media Using GPT Models

Detecting suicide risk among U.S. servicemembers and veterans: a deep learning approach using social media data.

Military Servicemembers and Veterans are at elevated risk for suicide, but rarely self-identify to their leaders or clinicians regarding their experience of suicidal thoughts. We developed an algorithm to identify posts containing suicide-related content on a military-specific social media platform. Publicly-shared social media posts (n = 8449) from a military-specific social media platform were reviewed and labeled by our team for the presence/absence of suicidal thoughts and behaviors and used to train several machine learning models to identify such posts. The best performing model was a deep learning (RoBERTa) model that incorporated post text and metadata and detected the presence of suicidal posts with relatively high sensitivity (0.85), specificity (0.96), precision (0.64), F1 score (0.73), and an area under the precision-recall curve of 0.84. Compared to non-suicidal posts, suicidal posts were more likely to contain explicit mentions of suicide, descriptions of risk factors (e.g. depression, PTSD) and help-seeking, and first-person singular pronouns. Our results demonstrate the feasibility and potential promise of using social media posts to identify at-risk Servicemembers and Veterans. Future work will use this approach to deliver targeted interventions to social media users at risk for suicide.

Tree hole rescue: an AI approach for suicide risk detection and online suicide intervention.

Adolescent suicide has become an important social issue of general concern. Many young people express their suicidal feelings and intentions through online social media, e.g., Twitter, Microblog. The "tree hole" is the Chinese name for places on the Web where people post secrets. It provides the possibility of using Artificial Intelligence and big data technology to detect the posts where someone express the suicidal signal from those "tree hole" social media. We have developed the Web-based intelligent agents (i.e., AI-based programs) which can monitor the "tree hole" websites in Microblog every day by using knowledge graph technology. We have organized Tree-hole Rescue Team, which consists of more than 1000 volunteers, to carry out suicide rescue intervention according to the daily monitoring notifications. From 2018 to 2023, Tree-hole Rescue Team has prevented more than 6600 suicides. A few thousands of people have been saved within those 6 years. In this paper, we present the basic technology of Web-based Tree Hole intelligent agents and elaborate how the intelligent agents can discover suicide attempts and issue corresponding monitoring notifications and how the volunteers of Tree Hole Rescue Team can conduct online suicide intervention. This research also shows that the knowledge graph approach can be used for the semantic analysis on social media.

Facial image analysis for automated suicide risk detection with deep neural networks

Accurately assessing suicide risk is a critical concern in mental health care. Traditional methods, which often rely on self-reporting and clinical interviews, are limited by their subjective nature and may overlook non-verbal cues. This study introduces an innovative approach to suicide risk assessment using facial image analysis. The Suicidal Visual Indicators Prediction (SVIP) Framework leverages EfficientNetb0 and ResNet architectures, enhanced through Bayesian optimization techniques, to detect nuanced facial expressions indicating mental state. The models’ interpretability is improved using GRADCAM, Occlusion Sensitivity, and LIME, which highlight significant facial regions for predictions. Using datasets DB1 and DB2, which consist of full and cropped facial images from social media profiles of individuals with known suicide outcomes, the method achieved 67.93% accuracy with EfficientNetb0 on DB1 and up to 76.6% accuracy with a Bayesian-optimized Support Vector Machine model using ResNet18 features on DB2. This approach provides a less intrusive, accessible alternative to video-based methods and demonstrates the substantial potential for early detection and intervention in mental health care.

A systematic review of suicide risk management strategies in primary care settings.

Suicide is a major public health concern. Recently, suicide rates have increased among traditionally low-risk groups (e.g., white, middle-aged males). Suicide risk assessments and prevention strategies should be tailored to specific at-risk populations. This systematic review examines suicide risk detection and management in primary care, focusing on treatments to reduce suicide rates and improve prevention efforts. A systematic review was conducted following PRISMA guidelines. Literature was collected and analyzed using Boolean operators with relevant keywords in databases (e.g., PubMed, Google Scholar, PsycINFO) to identify randomized and non-randomized studies focusing on suicide risk factors and management strategies in primary care, published in the past 10 years. The risk of bias 2.0 and Newcastle Ottawa scale was used to assess risk of bias, and data from moderate-quality studies were synthesized. Thirteen moderate-quality studies were reviewed. Key findings include the need for assessing modifiable risk factors like substance use and mental health. General practitioner (GP) engagement post-suicide attempt (SA) improves outcomes and reduces repeat SAs. Effective strategies include comprehensive risk assessments, collaborative treatment, and enhanced GP support. Barriers to effective suicide prevention include insufficient information, judgmental communication, lack of positive therapeutic relationships, and inadequate holistic assessments. These findings highlight the need for tailored suicide prevention strategies in primary care. However, the evidence sample size is small with reduced statistical power that limits generalizability. The included studies were also regional examinations, which restrict their broader relevance. Significant risk factors, barriers, and effective strategies for suicide prevention were identified. For children aged 12 or younger, preexisting psychiatric, developmental, or behavioral disorders, impulsive behaviors, aggressiveness, and significant stressful life events within the family were critical. For adults, loneliness, gaps in depression treatment, and social factors are significant. Barriers to suicide prevention included insufficient information, judgmental communication, lack of positive therapeutic relationships, inadequate holistic risk assessments, lack of individualized care, insufficient tangible support and resources, inconsistent follow-up procedures, variability in risk assessment, poor communication, stigma, and negative attitudes. Effective methods include the Postvention Assisting Bereaved by Suicide training program, continued education, comprehensive clinical assessments, individualized care, and community-based interventions like the SUPRANET program. https://www.crd.york.ac.uk/PROSPERO, identifier CRD42024550904.

Resilience and social support as protective factors against suicidal ideation among tertiary students during COVID-19: a cross-sectional study

BackgroundSuicidal ideation is a depression symptom which represents a key (cognitive) component of suicidality and plays an important role in suicide risk detection, intervention, and prevention. Despite existing research showing the importance of certain factors of depression symptoms and suicidal ideation, less is known about the interaction between the various risk and protective factors. The aim of the study was to examine whether living conditions characteristics and personal circumstances during the COVID-19 pandemic predicted the presence of depression symptoms and suicidal ideation among tertiary students and whether resilience and social support can mitigate the detrimental effects of difficult life circumstances.MethodA large online cross-sectional study was conducted in March 2021 among 4,645 Slovenian tertiary students. Hierarchical multiple regression and hierarchical logistic regression methods were used to assess and compare the effect of life circumstances variables, as opposed to resilience and social support, on depression symptoms and suicidal ideation.ResultsFemale gender, single relationship status, living alone, a higher degree of household conflict, having a history of mental illness and chronic disease diagnosis were significant predictors of depression scores. All but gender were also predictors of suicidal ideation. Household conflict and a history of mental illness were the factors showing the strongest effect in both cases. On the other hand, social support and, in particular, resilience proved to be strong protective factors against depression symptoms and suicidal ideation. After accounting for one’s resilience and social support, the explained variance in depression scores was more than doubled, while the harmful effect of household conflict and history of mental illness significantly decreased.ConclusionsThe findings stress the importance of one’s resilience and social support and explain why some people manage to maintain mental well-being despite finding themselves in difficult life circumstances, which was the case for many tertiary students during the COVID-19 pandemic. These insights may inform preventive efforts against developing suicidal ideation and may be used as support for the design and implementation of interventions for improving resilience and social support from childhood onward.

Addressing suicidality in autistic youth: Implications for school mental health professionals

AbstractSuicide is a complex, multifaceted crisis affecting children and adolescents, and has become a major public health concern. While there is a plethora of research on this topic among neurotypically developing youth, much less is known about the nature of suicidality in autistic individuals. Autistic youth show an increased vulnerability for suicidal tendencies that are poorly understood among school mental health (SMH) professionals, who report a range of competency and self‐efficacy to address suicidality in school settings (Debski et al., 2007; O'Neill et al., 2020). Co‐occurring mental health difficulties pose significant challenges in the early detection of suicide risk in autistic youth. Given that suicide is the leading cause of death among autistic individuals, the incidence of suicidality among autistic youth is of particular concern. This article provides a practice‐friendly review of the research outlining innate vulnerabilities that present risk factors for suicide among autistic youth, along with assessment considerations that help inform the management of suicide risk in autistic youth. We conclude with recommendations for integrating a multi‐tiered neurodiverse‐affirming approach with sensitivity for student intersecting identities and system‐level influences on student mental health that enables SMH professionals to create supportive school environments for autistic youth that prevent further harm.

Leveraging Psychiatric Scale for Suicide Risk Detection on Social Media

Leveraging Psychiatric Scale for Suicide Risk Detection on Social Media

Innovative Use of Self-Attention-Based Ensemble Deep Learning for Suicide Risk Detection in Social Media Posts

Suicidal ideation constitutes a critical concern in mental health, adversely affecting individuals and society at large. The early detection of such ideation is vital for providing timely support to individuals and mitigating its societal impact. With social media serving as a platform for self-expression, it offers a rich source of data that can reveal early symptoms of mental health issues. This paper introduces an innovative ensemble learning method named LSTM-Attention-BiTCN, which fuses LSTM and BiTCN models with a self-attention mechanism to detect signs of suicidality in social media posts. Our LSTM-Attention-BiTCN model demonstrated superior performance in comparison to baseline models in the realm of classification and suicidal ideation detection, boasting an accuracy of 0.9405, a precision of 0.9385, a recall of 0.9424, and an F1-score of 0.9405. Our proposed model can aid healthcare professionals in recognizing suicidal tendencies among social media users accurately, thereby contributing to efforts to reduce suicide rates.

Characteristics of Multi-Class Suicide Risks Tweets Through Feature Extraction and Machine Learning Techniques

This paper presents a detailed analysis of the linguistic characteristics connected to specific levels of suicide risks, providing insight into the impact of the feature extraction techniques on the effectiveness of the predictive models of suicide ideation. Prevalent initiatives of research works had been observed in the detection of suicide ideation from social media posts through feature extraction and machine learning techniques but scarcely on the multiclass classification of suicide risks and analysis of linguistic characteristics' impact on predictability. To address this issue, this paper proposes the implementation of a machine learning framework that is capable of analyzing multiclass classification of suicide risks from social media posts with extended analysis of linguistic characteristics that contribute to suicide risk detection. A total of 552 samples of a supervised dataset of Twitter posts were manually annotated for suicide risk modeling. Feature extraction was done through a combination of feature extraction techniques of term frequency-inverse document frequency (TF-IDF), Part-of-Speech (PoS) tagging, and valence-aware dictionary for sentiment reasoning (VADER). Data training and modeling were conducted through the Random Forest technique. Testing of 138 samples with scenarios of detections in real-time data for the performance evaluation yielded 86.23% accuracy, 86.71% precision, and 86.23% recall, an improved result with a combination of feature extraction techniques rather than data modeling techniques. An extended analysis of linguistic characteristics showed that a sentence's context is the main contributor to suicide risk classification accuracy, while grammatical tags and strong conclusive terms were not.

Characteristics of Multi-Class Suicide Risks Tweets Through Feature Extraction and Machine Learning Techniques

This paper presents a detailed analysis of the linguistic characteristics connected to specific levels of suicide risks, providing insight into the impact of the feature extraction techniques on the effectiveness of the predictive models of suicide ideation. Prevalent initiatives of research works had been observed in the detection of suicide ideation from social media posts through feature extraction and machine learning techniques but scarcely on the multiclass classification of suicide risks and analysis of linguistic characteristics' impact on predictability. To address this issue, this paper proposes the implementation of a machine learning framework that is capable of analyzing multiclass classification of suicide risks from social media posts with extended analysis of linguistic characteristics that contribute to suicide risk detection. A total of 552 samples of a supervised dataset of Twitter posts were manually annotated for suicide risk modeling. Feature extraction was done through a combination of feature extraction techniques of term frequency-inverse document frequency (TF-IDF), Part-of-Speech (PoS) tagging, and valence-aware dictionary for sentiment reasoning (VADER). Data training and modeling were conducted through the Random Forest technique. Testing of 138 samples with scenarios of detections in real-time data for the performance evaluation yielded 86.23% accuracy, 86.71% precision, and 86.23% recall, an improved result with a combination of feature extraction techniques rather than data modeling techniques. An extended analysis of linguistic characteristics showed that a sentence's context is the main contributor to suicide risk classification accuracy, while grammatical tags and strong conclusive terms were not.

Characteristics of Multi-Class Suicide Risks Tweets Through Feature Extraction and Machine Learning Techniques

This paper presents a detailed analysis of the linguistic characteristics connected to specific levels of suicide risks, providing insight into the impact of the feature extraction techniques on the effectiveness of the predictive models of suicide ideation. Prevalent initiatives of research works had been observed in the detection of suicide ideation from social media posts through feature extraction and machine learning techniques but scarcely on the multiclass classification of suicide risks and analysis of linguistic characteristics' impact on predictability. To address this issue, this paper proposes the implementation of a machine learning framework that is capable of analyzing multiclass classification of suicide risks from social media posts with extended analysis of linguistic characteristics that contribute to suicide risk detection. A total of 552 samples of a supervised dataset of Twitter posts were manually annotated for suicide risk modeling. Feature extraction was done through a combination of feature extraction techniques of term frequency-inverse document frequency (TF-IDF), Part-of-Speech (PoS) tagging, and valence-aware dictionary for sentiment reasoning (VADER). Data training and modeling were conducted through the Random Forest technique. Testing of 138 samples with scenarios of detections in real-time data for the performance evaluation yielded 86.23% accuracy, 86.71% precision, and 86.23% recall, an improved result with a combination of feature extraction techniques rather than data modeling techniques. An extended analysis of linguistic characteristics showed that a sentence's context is the main contributor to suicide risk classification accuracy, while grammatical tags and strong conclusive terms were not.

ASSESSING THE PSYCHOMETRIC PROPERTIES OF THE GERIATRIC SUICIDE IDEATION SCALE IN MIDDLE-AGED AND OLDER MEN

Abstract Middle-aged and older men have high rates of suicide, necessitating risk detection. We developed the Geriatric Suicide Ideation Scale (GSIS; Heisel & Flett, 2006) as an age-specific, multidimensional suicide risk assessment tool. The GSIS has shown strong psychometric properties in clinical, community, and residential samples (see Heisel & Flett, 2016), yet research has lagged investigating its utility with middle-aged and older men. The purpose of the present study was thus to assess the psychometric properties of the GSIS administered to 82 men, 55 years and older (M=63.3, SD=4.6 years), who participated in a meaning-centered psychological intervention group for those concerned about or struggling with the transition to retirement (Heisel et al., 2020). Psychometric analyses included investigation of participant response characteristics, internal consistency, and construct validity. Findings demonstrated acceptable internal consistency for GSIS totals (α =.88) and for its Suicide Ideation, Death Ideation, Loss of Personal and Social Worth, and Perceived Meaning in Life subscales (α =.62-.81). Positive associations between the GSIS and negative psychological factors (depression, anxiety, hopelessness, loneliness, perceived lack of mattering to others, and history of suicidal behavior; r =.30 to .51) and negative associations with positive factors (life satisfaction, psychological well-being, perceived support, and meaning in life; r = -.21 to -.51) supported its construct validity. These and other findings will be discussed in the broader context of upstream population level approaches to suicide risk detection and prevention. (Note: this abstract was accepted for the 2022 GSA meeting, but we had to withdraw due to family loss)

Detection of Suicidal Ideation in Clinical Interviews for Depression Using Natural Language Processing and Machine Learning: Cross-Sectional Study.

Assessing patients' suicide risk is challenging, especially among those who deny suicidal ideation. Primary care providers have poor agreement in screening suicide risk. Patients' speech may provide more objective, language-based clues about their underlying suicidal ideation. Text analysis to detect suicide risk in depression is lacking in the literature. This study aimed to determine whether suicidal ideation can be detected via language features in clinical interviews for depression using natural language processing (NLP) and machine learning (ML). This cross-sectional study recruited 305 participants between October 2020 and May 2022 (mean age 53.0, SD 11.77 years; female: n=176, 57%), of which 197 had lifetime depression and 108 were healthy. This study was part of ongoing research on characterizing depression with a case-control design. In this study, 236 participants were nonsuicidal, while 56 and 13 had low and high suicide risks, respectively. The structured interview guide for the Hamilton Depression Rating Scale (HAMD) was adopted to assess suicide risk and depression severity. Suicide risk was clinician rated based on a suicide-related question (H11). The interviews were transcribed and the words in participants' verbal responses were translated into psychologically meaningful categories using Linguistic Inquiry and Word Count (LIWC). Ordinal logistic regression revealed significant suicide-related language features in participants' responses to the HAMD questions. Increased use of anger words when talking about work and activities posed the highest suicide risk (odds ratio [OR] 2.91, 95% CI 1.22-8.55; P=.02). Random forest models demonstrated that text analysis of the direct responses to H11 was effective in identifying individuals with high suicide risk (AUC 0.76-0.89; P<.001) and detecting suicide risk in general, including both low and high suicide risk (AUC 0.83-0.92; P<.001). More importantly, suicide risk can be detected with satisfactory performance even without patients' disclosure of suicidal ideation. Based on the response to the question on hypochondriasis, ML models were trained to identify individuals with high suicide risk (AUC 0.76; P<.001). This study examined the perspective of using NLP and ML to analyze the texts from clinical interviews for suicidality detection, which has the potential to provide more accurate and specific markers for suicidal ideation detection. The findings may pave the way for developing high-performance assessment of suicide risk for automated detection, including online chatbot-based interviews for universal screening.

Suicide-related disclosure patterns among culturally minoritized youth: Examining differences across race, ethnicity, gender identity, and sexual orientation.

Sharing one's suicidal thoughts and behaviors, or suicide-related disclosure, allows adolescents to recruit help from others. Despite elevated risk among culturally minoritized youth, their suicide-related disclosure remains understudied. 191 adolescents (M = 15.98, SD = 1.04, range = 13-17), including minoritized youth (38% racially, 19% ethnically, 40% gender, and 77% sexually), were recruited via social media ads and completed an anonymous online survey on suicide-related disclosure to informal support sources (e.g., family, friends). Disclosure rates, targets, and reasons for disclosure and nondisclosure were compared based on race, ethnicity, gender identity, and sexual orientation. Racially minoritized adolescents less often disclosed suicidal ideation and more strongly endorsed fear of negative reactions and resistance to intervention as reasons for nondisclosure, and reciprocity as reasons for disclosure, than White adolescents. Cisgender adolescents less often disclosed suicidal ideation and more strongly endorsed dismissal of suicide risk as reasons for nondisclosure than gender minoritized adolescents. Non-Hispanic adolescents more strongly endorsed help-seeking as reasons for disclosure than Hispanic adolescents. While adolescents overall disclosed most often to friends, heterosexual adolescents disclosed more to family than sexually minoritized adolescents. Racial and gender disparities in suicide-related disclosure may occur for distinct reasons among adolescents. Uncovering patterns of disclosure may facilitate detection of suicide risk among minoritized youth.

Suicide among Cancer Patients: Current Knowledge and Directions for Observational Research.

Cancer is a major public health concern associated with an increased risk of psychosocial distress and suicide. The reasons for this increased risk are still being characterized. The purpose of this study is to highlight existing observational studies on cancer-related suicides in the United States and identify gaps for future research. This work helps inform clinical and policy decision-making on suicide prevention interventions and ongoing research on the detection and quantification of suicide risk among cancer patients. We identified 73 peer-reviewed studies (2010-2022) that examined the intersection of cancer and suicide using searches of PubMed and Embase. Overall, the reviewed studies showed that cancer patients have an elevated risk of suicide when compared to the general population. In general, the risk was higher among White, male, and older cancer patients, as well as among patients living in rural areas and with lower socioeconomic status. Future studies should further investigate the psychosocial aspects of receiving a diagnosis of cancer on patients' mental health as well as the impact of new treatments and their availability on suicide risk and disparities among cancer patients to better inform policies.

Establishment of a risk prediction model for suicide attempts in first-episode and drug naïve patients with major depressive disorder

BackgroundSuicide is common in patients with major depressive disorder (MDD) and has serious consequences for individuals and families. This study aims to establish a risk prediction model for suicide attempts in MDD patients to make the detection of suicide risk more accurate and effective. MethodsA cross-sectional survey, clinical examination, and biochemical indicator tests were performed on 1718 first-episode and drug naïve patients with major depressive disorder. We used Machine Learning to establish a risk prediction model for suicide attempts in FEDN patients with MDD. ResultsFive predictors were identified by LASSO regression analysis from a total of 20 variables studied, namely psychotic symptoms, anxiety symptoms, thyroid peroxidase antibodies (ATPO), total cholesterol (TC), and high-density lipoprotein-cholesterol (HDL-C). The model constructed using the five predictors displayed moderate predictive ability, with an area under the ROC of 0.771 in the training set and 0.720 in the validation set. The DCA curve showed that the nomogram could be applied clinically if the risk threshold was between 22 % and 60 %. The risk threshold was found to be between 20 % and 60 % in external validation. ConclusionIntroducing psychotic symptoms, anxiety symptoms, ATPO, TC, and HDL-C to the risk nomogram increased its usefulness for predicting suicide risk in patients with MDD. It may be useful in clinical decision-making or in discussions with patients, especially in crisis interventions.

Suicide risk detection using artificial intelligence: the promise of creating a benchmark dataset for research on the detection of suicide risk.

Suicide is a leading cause of death that demands cross-disciplinary research efforts to develop and deploy suicide risk screening tools. Such tools, partly informed by influential suicide theories, can help identify individuals at the greatest risk of suicide and should be able to predict the transition from suicidal thoughts to suicide attempts. Advances in artificial intelligence have revolutionized the development of suicide screening tools and suicide risk detection systems. Thus, various types of AI systems, including text-based systems, have been proposed to identify individuals at risk of suicide. Although these systems have shown acceptable performance, most of them have not incorporated suicide theories in their design. Furthermore, directly applying suicide theories may be difficult because of the diversity and complexity of these theories. To address these challenges, we propose an approach to develop speech- and language-based suicide risk detection systems. We highlight the promise of establishing a benchmark textual and vocal dataset using a standardized speech and language assessment procedure, and research designs that distinguish between the risk factors for suicide attempt above and beyond those for suicidal ideation alone. The benchmark dataset could be used to develop trustworthy machine learning or deep learning-based suicide risk detection systems, ultimately constructing a foundation for vocal and textual-based suicide risk detection systems.

Korean adolescent suicide and search volume for "self-injury" on internet search engines.

Many adolescents with suicidal ideation receive support through the Internet. However, they also find ways to attempt suicide or strengthen their suicidal ideation through this medium. This study analyzed the association between the search volume of suicide-related terms and the number of suicides among Korean adolescents. We also analyzed the correlations between the search volumes of suicide-related terms. We selected seven words (suicide, self-injury, depression, academic score, school violence, outcasts, and family trouble) related to adolescent suicide. A dataset was constructed by combining data from the most commonly used search engine in Korea (Naver Datalab) and the daily number of adolescent suicides in school settings (n = 347) from January 1, 2016 to December 31, 2018, collected from the Ministry of Education. Poisson regression and Pearson correlation analyses were performed. Significant associations were found between suicide attempts and search term volumes, which differed according to sex and time interval. Among the search terms, "self-injury" was most strongly associated with suicide, and this association was significant at all time intervals (daily, weekly, and monthly) in female adolescents and in the total population. Further, the association was strongest in the daily suicide data. More search term volumes were related to suicide in the daily and weekly data than in the monthly data. There were positive correlations between "suicide," "self-injury," and "depression" search volumes. Further studies with larger sample sizes, more search terms, and analysis of time intervals between suicide-related term search and suicide death are required. These studies can contribute to the establishment of an online suicide prevention system to detect suicide risk in adolescents and provide interventions.