
Objectives: This scoping review maps peer-reviewed evaluations of suicide prevention trainings by mapping study (e.g., design, geographic location) and training characteristics (e.g., content, duration, modality, pedagogical methods, populations), and outcomes assessed using Kirkpatrick's four-level evaluation model (Level 1: reaction; Level 2: learning; Level 3: behavior change; Level 4: results). Methods: Systematic searches in MEDLINE, PsycINFO, Embase, and CINAHL (2014-2024) identified peer-reviewed studies published in English or French that quantitatively examined suicide prevention training outcomes. Reviewers independently screened titles/abstracts and full texts, with conflicts resolved by consensus. Study and training characteristics were charted using a structured guide. Outcomes were categorized using Kirkpatrick's four-level model, and descriptive statistics were used to summarize data. We followed the framework developed by Arksey and O'Malley and adhered to the PRISMA Extension for Scoping Reviews guidelines. Results: Across 11,312 screened records, 326 studies met inclusion criteria, with most conducted in high-income countries (91%) and relying on nonrandomized designs (79%). Trainings were primarily delivered in-person (61%), brief (≤4 h; 53%), targeted clinicians (58%), and focused on supporting clinical populations (37%). Outcomes most frequently assessed Level 2 variables (e.g., knowledge, attitudes, skills; 88%), whereas far fewer studies measured level 3 (29%) or level 4 outcomes (9%), such as increased use of skills in real-world practice or reduced suicide-related distress. Conclusion: Suicide prevention trainings are widely implemented, and although many studies have attempted to evaluate their outcomes, this review identifies major gaps in geographic representation, study design, and outcome measurement. Future research should prioritize higher-level outcomes, with longer-term follow-up and evaluations conducted across diverse contexts to strengthen the evidence base and inform implementation.
The Suicide Observatory in Cork and Kerry was established to address the urgent need for real-time suicide data in Ireland to enable timely suicide prevention. This study aimed to assess the performance and efficiency of the Suicide Observatory to inform decisions about its upscaling and wider implementation. We designed a mixed-methods approach and used Centers for Disease Control and Prevention evaluation guidelines to assess key attributes of the Observatory: sensitivity, positive predictive value, data quality, timeliness, simplicity, accessibility, acceptability, and the overall usefulness. We compared data obtained from an audit of the Observatory for 2019-2020 with coroner-determined suicide cases and national suicide mortality data, and conducted semistructured interviews with stakeholders. We found that the Suicide Observatory in Cork shows a high level of data quality (sensitivity: 83.72%; 95% CI = 74.20%-90.80%), positive predictive value: 75.79% (95% CI = 74.04%-77.46%), data missingness: <2%). It applied a fortnightly case submission interval, and participants generally regarded the data collection as straightforward. The Suicide Observatory data were made available to stakeholders and decision makers. Participants reported strong acceptability, primarily due to its simplicity and accessibility. The overall usefulness of the Suicide Observatory was considered to be high as it improved the understanding of the suicide demography in the regions and enabled timely suicide prevention initiatives and support for affected communities. It was concluded that the Suicide Observatory has high levels of performance and efficiency and has been crucial for timely suicide prevention. The findings support its wider implementation in Ireland.
Background: Suicide-related media is known to influence suicide rates. Large language models (LLMs), a form of generative artificial intelligence (AI), are increasingly being used as a writing tool. However, the quality of LLM-generated suicide-related content has yet to be assessed. Aims: We aimed to examine suicide-related outputs from three LLMs (GPT-4, Grok, ERNIE) to characterize output quality. Methods: We provided the LLMs with 11 prompts to write different types of suicide-related content common in public media, each in five writing/content styles (broadsheet news report, tabloid news report, adult fiction, teen fiction, social media influencer). We used 3 × 2 chi square tests to compare characteristics of the outputs particularly related to adherence to responsible media guidelines for suicide reporting and overarching narratives. Results: AI outputs were generated from March 12 to July 11, 2024. A total of 147, 263, and 143 responses from GPT-4, Grok, and ERNIE were analyzed, respectively. Willingness to respond to suicide-related prompts and narrative content varied substantially across the three LLMs, with Grok producing more responses than the other two (GPT-4 = 53%, Grok = 96%, ERNIE = 52%). GPT-4 was more likely to generate emotionally supportive and antistigma messaging. Grok more frequently included harmful details such as suicide methods and romanticized portrayals. ERNIE emphasized male suicide, social support, and societal/community solutions. The proportion of outputs communicating stories of hope and recovery from a suicide crisis was both relatively low and consistent across the LLMs (16-22%). Limitations: This study examined specific prompts posed to three specific LLMs during a single epoch of time. Conclusions: LLMs produce a variety of suicide-related story content when prompted. Some stories, particularly those generated by Grok, were inconsistent with responsible media guidelines and only a minority of stories focused on hope and recovery. Engagement with AI companies to promote safer and more accurate suicide-related content is warranted as the use of LLMs continues to expand.
Background: Clozapine is associated with decreased insomnia and suicidal ideation and behavior (SIB), both of which are prevalent in psychiatric disorders. Aims: To consider effects on insomnia underlying the antisuicidal properties of clozapine, we investigated the reported odds of suicidal ideation, suicide attempt and/or suicide death with and without concurrent insomnia as spontaneously reported psychiatric adverse drug reactions (ADRs) in adults using clozapine compared to other second-generation antipsychotics (SGAs) for schizophrenia or schizoaffective disorder. Methods: We searched the US FDA Adverse Event Reporting System from inception through September 2024 for which an SGA used for schizophrenia or schizoaffective disorder was the suspected agent of a psychiatric ADR. Results: Compared to clozapine, the unadjusted reported odds of concurrent insomnia and suicidal ideation (rORs = 1.39-3.44) were significantly increased for all other SGAs. Similarly, compared to clozapine, the unadjusted reported odds of concurrent insomnia and suicide attempt/death (rORs = 2.32-4.76) were significantly increased for all other SGAs. Controlling for age and sex did not change the pattern of findings. Our findings are consistent with existing evidence for decreased insomnia and SIB in individuals taking clozapine for schizophrenia or schizoaffective disorder. Limitations: Rigorous longitudinal studies with established measures of sleep and suicide are warranted to investigate this mechanism. Conclusion: Compared to clozapine, all other SGAs were associated with increased reported odds of concurrent insomnia SIB. Findings provide additional evidence for improved sleep as a potential pathway underlying the antisuicidal properties of clozapine.
Background: The suicide crisis syndrome (SCS) is a proposed diagnostic construct under consideration for inclusion in DSM as a suicide-specific syndrome. The SCS is an acute affective-cognitive state, characterized by entrapment (Criterion A), affective disturbances (Criterion B1), loss of cognitive control (B2), hyperarousal (B3), and social withdrawal (B4). Suicidal ideation is not a criterion. While psychometrically valid self-report measures of SCS are well-established, the need for a psychometrically valid interview remains. A clinician-administered tool is essential for broader clinical use and dissemination. Aims: This study assesses the convergent validity of the Suicide Crisis Syndrome Checklist (SCS-C), a semistructured interview, relative to the Revised Suicide Crisis Inventory (SCI-2), a validated dimensional self-report measure. The goals of this study are: (1) to evaluate the convergent validity of the Suicide Crisis Syndrome Checklist (SCS-C) against SCI-2 and (2) to evaluate classification configurations of SCS-C criteria based on sensitivity and specificity. Method: In a sample of 217 adult psychiatric inpatients, receiver operating characteristic curve (ROC) analyses investigated each individual SCS-C criterion and configurations of criteria (e.g., inclusion of one, two, three, or all four B criteria) for concordance with SCI-2. Sensitivity and specificity values corresponded to the score with the highest Youden's J. Results: The first four SCS-C criteria showed good concordance with corresponding subscale scores (AUC = .794 - .866) while the final criterion, B4 Social Withdrawal, showed weaker alignment (AUC = .716). A configuration of Criterion A plus three B Criteria demonstrated particularly strong concordance (AUC = .800, sensitivity = .861, specificity = .654). Limitations: Recruiting from an acute inpatient setting and administering via research assistants, potentially affecting generalizability, and not conducting further construct validity tests. Conclusions: Findings support the convergent validity of the SCS-C and suggest at present the SCS-C corresponds most closely with the SCI-2 when utilizing four of the five criteria.