Choice of suicide method can strongly influence the outcome of suicidal behaviour, and is an important aspect of the process and planning involved in a suicide attempt. Yet, the reasons why individuals consider, choose or discard particular methods are not well understood.

This is the first study to explore method choices among people with a history of suicidal behaviour and individuals who have experienced, but not enacted, suicidal thoughts.

Via an online survey, we gathered open-ended data about choice of methods in relation to suicidal thoughts and behaviours, including reasons for and against specific means of harm.

A total of 712 respondents had attempted suicide, and a further 686 experienced suicidal thoughts (but not acted on them). Self-poisoning was the most commonly contemplated and used method of suicide, but most respondents had considered multiple methods. Method choices when contemplating suicide included a broader range of means than those used in actual attempts, and more unusual methods, particularly if perceived to be lethal, ‘easy’, quick, accessible and/or painless. Methods used in suicide attempts were, above all, described as having been accessible at the time, and were more commonly said to have been chosen impulsively. Key deterrents against the use of specific methods were the presence of and impact on other people, especially loved ones, and fears of injury and survival.

Exploration of method choices can offer novel insights into the transition from suicidal ideation to behaviour. Results underscore the need for preventative measures to restrict access to means and delay impulsive behaviour.

The processes and planning involved in choosing and attempting to die by a particular method of suicide are not well understood. Accounts from those who have thought about or attempted suicide using a specific method might allow us to better understand the ways in which people come to think about, plan and enact a suicide attempt.

To understand from first-person accounts the processes and planning involved in a suicide attempt on the railway.

Thematic analysis was conducted of qualitative interviews (N = 34) undertaken with individuals who had contemplated or attempted suicide by train.

Participants explained how they decided upon a particular method, time and place for a suicide attempt. Plans were described as being contingent on a number of elements (including the likelihood of being seen or interrupted), rather than being fixed in advance. Participants mentally rehearsed and evaluated a particular method, which would sometimes involve imagining in detail what would happen before, during and after an attempt. The extent to which this involved others (train drivers, partners, friends) was striking.

By giving people free reign to describe in their own words the processes they went through in planning and undertaking a suicide attempt, and by not interpreting such accounts through a lens of deficit and pathology, we can arrive at important insights into how people come to think and feel about, plan and enact a suicide attempt. The findings have implications in terms of understanding suicide risk and prevention more broadly.

There is still little knowledge of objective suicide risk stratification.

This study aims to develop models using machine-learning approaches to predict suicide attempt (1) among survey participants in a nationally representative sample and (2) among participants with lifetime major depressive episodes. We used a cohort called the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC) that was conducted in two waves and included a nationally representative sample of the adult population in the United States. Wave 1 involved 43 093 respondents and wave 2 involved 34 653 completed face-to-face reinterviews with wave 1 participants. Predictor variables included clinical, stressful life events, and sociodemographic variables from wave 1; outcome included suicide attempt between wave 1 and wave 2.

The model built with elastic net regularization distinguished individuals who had attempted suicide from those who had not with an area under the ROC curve (AUC) of 0.89, balanced accuracy 81.86%, specificity 89.22%, and sensitivity 74.51% for the general population. For participants with lifetime major depressive episodes, AUC was 0.89, balanced accuracy 81.64%, specificity 85.86%, and sensitivity 77.42%. The most important predictor variables were a diagnosis of borderline personality disorder, post-traumatic stress disorder, and being of Asian descent for the model in all participants; and previous suicide attempt, borderline personality disorder, and overnight stay in hospital because of depressive symptoms for the model in participants with lifetime major depressive episodes. Random forest and artificial neural networks had similar performance.

Risk for suicide attempt can be estimated with high accuracy.

Suicide prediction models have been formulated in a variety of ways and are heterogeneous in the strength of their predictions. Machine learning has been a proposed as a way of improving suicide predictions by incorporating more suicide risk factors.

To determine whether machine learning and the number of suicide risk factors included in suicide prediction models are associated with the strength of the resulting predictions.

Random-effect meta-analysis of exploratory suicide prediction models constructed by combining two or more suicide risk factors or using clinical judgement (Prospero Registration CRD42017059665). Studies were located by searching for papers indexed in PubMed before 15 August 2020 with the term suicid* in the title.

In total, 86 papers reported 102 suicide prediction models and included 20 210 411 people and 106 902 suicides. The pooled odds ratio was 7.7 (95% CI 6.7–8.8) with high between-study heterogeneity (I2 = 99.5). Machine learning was associated with a non-significantly higher odds ratio of 11.6 (95% CI 6.0–22.3) and clinical judgement with a non-significantly lower odds ratio of 4.7 (95% CI 2.1–10.9). Models including a larger number of suicide risk factors had a higher odds ratio when machine-learning studies were included (P = 0.02). Among non-machine-learning studies, suicide prediction models including fewer risk factors performed just as well as those including more risk factors.

Machine learning might have the potential to improve the performance of suicide prediction models by increasing the number of included suicide risk factors but its superiority over other methods is unproven.

It remains unclear whether the coronavirus disease 2019 (COVID-19) pandemic is having an impact on suicide rates (SR). Economic insecurity and mental disorders are risk factors for suicide, which may increase during the pandemic.

Data on suicide events in a major city in Germany, and the corresponding life years (LY) were provided by the local authorities. For the year 2020, periods without restrictions on freedom of movement and social contact were compared with periods of moderate and severe COVID-19 restrictions. To avoid distortions due to seasonal fluctuations and linear time trends, suicide risk during the COVID-19 pandemic was compared with data from 2010 to 2019 using an interrupted time series analysis.

A total of 643 suicides were registered and 6 032 690 LY were spent between 2010 and 2020. Of these, 53 suicides and 450 429 LY accounted for the year 2020.

In 2020, SR (suicides per 100 000 LY) were lower in periods with severe COVID-19 restrictions (SR = 7.2, χ2 = 4.033, p = 0.045) compared with periods without restrictions (SR = 16.8). A comparison with previous years showed that this difference was caused by unusually high SR before the imposition of restrictions, while SR during the pandemic were within the trend corridor of previous years (expected suicides = 32.3, observed suicides = 35; IRR = 1.084, p = 0.682).

SR during COVID-19 pandemic are in line with the trend in previous years. Careful monitoring of SR in the further course of the COVID-19 crisis is urgently needed. The findings have regional reference and should not be over-generalised.

Older adults receiving support services are a population at risk for self-harm due to physical illness and functional impairment, which are known risk factors. This study aims to investigate the relative importance of predictive factors of nonfatal self-harm among older adults assessed for support services in New Zealand.

interRAI-Home Care (HC) national data of older adults (aged ≥ 60) were linked to mortality and hospital discharge data between January 1, 2012 and December 31, 2016. We calculated the crude incidence of self-harm per 100,000 person-years, and gender and age-adjusted standardized incidence ratios (SIRs). The Fine and Gray competing risk regression model was fitted to estimate the hazard ratio (HR; 95% CIs) of self-harm associated with various demographic, psychosocial, clinical factors, and summary scales.

A total of 93,501 older adults were included. At the end of the follow-up period, 251 (0.27%) people had at least one episode of nonfatal self-harm and 36,333 (38.86%) people died. The overall incidence of nonfatal self-harm was 160.39 (95% CI, 141.36–181.06) per 100,000 person-years and SIR was 5.12 (95% CI, 4.51–5.78), with the highest incidence in the first year of follow-up. Depression diagnosis (HR, 3.02, 2.26–4.03), at-risk alcohol use (2.38, 1.30–4.35), and bipolar disorder (2.18, 1.25–3.80) were the most significant risk factors. Protective effects were found with cancer (0.57, 0.36–0.89) and severe level of functional impairment measured by Activities of Daily Living (ADL) Hierarchy Scale (0.56, 0.35–0.89).

Psychiatric factors are the most significant predictors for nonfatal self-harm among older adults receiving support services. Our results can be used to inform healthcare professionals for timely identification of people at high risk of self-harm and the development of more efficient and targeted prevention strategies, with specific attention to individuals with depression or depressive symptoms, particularly in the first year of follow-up.

Suicide is an important, understudied public health problem in Bangladesh, where risk factors for suicide have not been investigated by case–control psychological autopsy study.

To identify the major risk factors for suicide in Dhaka, Bangladesh.

We designed a matched case–control psychological autopsy study. We conducted a semi-structured interview with the next-of-kin of 100 individuals who died by suicide and 100 living controls, matched for age, gender and area of residence. The study was conducted from July 2019 to July 2020.

The odds ratios for the risk factors were 15.33 (95% CI, 4.76–49.30) for the presence of a psychiatric disorder, 17.75 (95% CI, 6.48–48.59) for life events, 65.28 (95% CI, 0.75–5644.48) for previous attempts and 12 (95% CI, 1.56–92.29) for sexual abuse.

The presence of a psychiatric disorder, immediate life events, previous suicidal attempts and sexual abuse were found as significant risk factors for suicide in Dhaka, Bangladesh.

Suicidality is one of the most common complications of mental disorders, so that the identification of potential biomarkers may be relevant in clinical practice. To date, the role of serum lipids and neutrophil/lymphocyte ratio (NLR) has been explored albeit with conflicting results. To the best of our knowledge, no study has explored lipid levels concomitantly with NLR in relation to violent suicide attempts. Therefore, we aimed to investigate whether serum lipid levels and NLR might be associated with the violent method of suicide attempts.

The study group consisted of 163 inpatients who attempted suicide. Blood samples were collected at the beginning of hospitalization to measure total cholesterol, low-density lipoprotein (LDL), high-density lipoprotein, very-low-density lipoprotein (VLDL), triglycerides, and NLR. Descriptive analyses of the total sample were performed. The included patients were divided into two groups according to violent/nonviolent method. Groups were compared in terms of lipid (MANCOVAs).

Plasma levels of total cholesterol (F = 5.66; P = .02), LDL (F = 4.94; P = .03), VLDL (F = 5.66; P = .02), and NLR (F = 8.17; P < .01) resulted to be significantly lower in patients that used a violent method compared to patients who attempted suicide with a nonviolent method.

Low cholesterol, LDL, and VLDL levels as well as low NLR value were associated with a violent method of suicide attempt in patients with mental disorders. Further studies are needed to confirm these results.

Individuals attending emergency departments following self-harm have increased risks of future self-harm. Despite the common use of risk scales in self-harm assessment, there is growing evidence that combinations of risk factors do not accurately identify those at greatest risk of further self-harm and suicide.

To evaluate and compare predictive accuracy in prediction of repeat self-harm from clinician and patient ratings of risk, individual risk-scale items and a scale constructed with top-performing items.

We conducted secondary analysis of data from a five-hospital multicentre prospective cohort study of participants referred to psychiatric liaison services following self-harm. We tested predictive utility of items from five risk scales: Manchester Self-Harm Rule, ReACT Self-Harm Rule, SAD PERSONS, Modified SAD PERSONS, Barratt Impulsiveness Scale and clinician and patient risk estimates. Area under the curve (AUC), sensitivity, specificity, predictive values and likelihood ratios were used to evaluate predictive accuracy, with sensitivity analyses using classification-tree regression.

A total of 483 self-harm episodes were included, and 145 (30%) were followed by a repeat presentation within 6 months. AUC of individual items ranged from 0.43–0.65. Combining best performing items resulted in an AUC of 0.56. Some individual items outperformed the scale they originated from; no items were superior to clinician or patient risk estimations.

No individual or combination of items outperformed patients’ or clinicians’ ratings. This suggests there are limitations to combining risk factors to predict risk of self-harm repetition. Risk scales should have little role in the management of people who have self-harmed.

Although research has identified a wide range of risk factors for suicidal behavior in prisoners, it does not establish who is most likely to act on their suicidal thoughts while incarcerated.

Self-report data were collected from a random sample of 1,203 adult men incarcerated across 15 prisons in Belgium, who represent 12% of all male prisoners nationwide.

One-third (33%) of participants reported having suicidal thoughts during their incarceration, of whom 26% attempted suicide in prison (9% of all prisoners). Factors independently associated with suicide attempt among prisoners with suicidal ideation were violent offending (adjusted odds ratio [aOR] = 2.64, 95% confidence interval [CI] 1.33–5.23), in-prison drug use (aOR = 2.30, 95% CI 1.25–4.22), exposure to suicidal behavior (aOR = 1.96, 95% CI 1.04–3.68), and a lifetime history of nonsuicidal self-injury (aOR = 1.90, 95% CI 1.08–3.36). While related to suicidal thoughts, markers of psychiatric morbidity and aspects of the prison regime were not associated with the progression to suicide attempt.

Many prisoners who think about suicide do not attempt suicide while incarcerated. Factors associated with suicidal ideation are distinct from those that govern the transition to suicidal behavior. Our findings lend support to the hypothesis that behavioral disinhibition might act as a catalyst in the translation of suicidal thoughts into action.

This study explored the root causes of deaths by suicide among patients under the care of a mental health trust. Thematic analysis was carried out to identify themes from the serious incident reports for patients between 1 January 2017 and 31 July 2018.

In total, 48 cases were reviewed. Three main themes emerged from this study: patient-, professional- and organisation-related factors. The majority of the deaths were caused by patient-related factors, particularly exacerbation of the patient's mental health condition.

This study provides insight into perceived causes of death by suicide among mental health patients. It is hoped that this will, in turn, influence the manner in which decisions, policies and resource allocation are carried out to further prevent and reduce the incidence of suicide, particularly among mental health patients.

Little is known about the trend and predictors of 21-year mortality and suicide patterns in persons with schizophrenia.

To explore the trend and predictors of 21-year mortality and suicide in persons with schizophrenia in rural China.

This longitudinal follow-up study included 510 persons with schizophrenia who were identified in a mental health survey of individuals (≥15 years old) in 1994 in six townships of Xinjin County, Chengdu, China, and followed up in three waves until 2015. Kaplan–Meier survival analysis and Cox hazard regressions were conducted.

Of the 510 participants, 196 died (38.4% mortality) between 1994 and 2015; 13.8% of the deaths (n = 27) were due to suicide. Life expectancy was lower for men than for women (50.6 v. 58.5 years). Males consistently showed higher rates of mortality and suicide than females. Older participants had higher mortality (hazard ratio HR = 1.03, 95% CI 1.01–1.05) but lower suicide rates (HR = 0.95, 95% CI 0.93–0.98) than their younger counterparts. Poor family attitudes were associated with all-cause mortality and death due to other causes; no previous hospital admission and a history of suicide attempts independently predicted death by suicide.

Our findings suggest there is a high mortality and suicide rate in persons with schizophrenia in rural China, with different predictive factors for mortality and suicide. It is important to develop culture-specific, demographically tailored and community-based mental healthcare and to strengthen family intervention to improve the long-term outcome of persons with schizophrenia.