Identifying the relative importance of non-suicidal self-injury features in classifying suicidal ideation, plans, and behavior using exploratory data mining

Quantifying the Importance of Non-Suicidal Self-Injury Characteristics in Predicting Different Clinical Outcomes: Using Random Forest Model

Existing research on non-suicidal self-injury (NSSI) among adolescents has primarily concentrated on general risk factors, leaving a significant gap in understanding the specific NSSI characteristics that predict diverse psychopathological outcomes. This study aims to address this gap by using Random Forests to discern the significant predictors of different clinical outcomes. The study tracked 348 adolescents (64.7% girls; mean age = 13.31, SD = 0.91) over 6 months. Initially, 46 characteristics of NSSI were evaluated for their potential to predict the repetition of NSSI, as well as depression, anxiety, and suicidal risks at a follow-up (T2). The findings revealed distinct predictors for each psychopathology. Specifically, psychological pain was identified as a significant predictor for depression, anxiety, and suicidal risks, while the perceived effectiveness of NSSI was crucial in forecasting its repetition. These findings imply that it is feasible to identify high-risk individuals by assessing key NSSI characteristics, and also highlight the importance of considering diverse NSSI characteristics when working with self-injurers.

Psychological pain and sociodemographic factors classified suicide attempt and non-suicidal self-injury in adolescents

This study aimed to utilize machine learning to explore the psychological similarities and differences between suicide attempt (SA) and non-suicidal self-injury (NSSI), with a particular focus on the role of psychological pain. A total of 2385 middle school students were recruited using cluster sampling. The random forest algorithm was used with 25 predictors to develop classification models of SA and NSSI, respectively, and to estimate the importance scores of each predictor. Based on these scores and related theories, shared risk factors (control feature set) and distinct risk factors (distinction feature set) were selected and tested to distinguish between NSSI and SA. The machine learning algorithm exhibited fair to good performance in classifying SA history [Area Under Curves (AUCs): 0.65-0.87] and poor performance in classifying NSSI history (AUC: 0.61-0.68). The distinction feature set comprised pain avoidance, family togetherness, and deviant peer affiliation, while the control feature set included pain arousal, painful feelings, and crisis events. The distinction feature set slightly but stably outperformed the control feature set in classifying SA from NSSI. The three-dimensional psychological pain model, especially pain avoidance, might play a dominant role in understanding the similarities and differences between SA and NSSI.

Using recursive partitioning to predict presence and severity of suicidal ideation amongst college students

OBJECTIVE

Predicting the presence and severity of suicidal ideation in college students is important, as deaths by suicide amongst young adults have increased in the past 20 years.

PARTICIPANTS

We recruited college students (N = 5494) from ten universities across eight states.

METHOD

Participants answered three questionnaires related to lifetime and past month suicidal ideation, and an indicator of suicidal ideation in a DSM-5 symptom measure. We used recursive partitioning to predict the presence, absence, and severity, of suicidal ideation.

RESULTS

Recursive partitioning models varied in their accuracy and performance. The best-performing model consisted of predictors and outcomes measured by the DSM-5 Level 1 Cross-Cutting Symptom Measure. Sexual orientation was also an important predictor in most models.

CONCLUSIONS

A single measure of DSM-5 symptom severity may help universities understand suicide severity to promote targeted interventions. Though further work is needed, as similar scaling amongst predictors could have influenced the model.

The use of machine learning on administrative and survey data to predict suicidal thoughts and behaviors: a systematic review

Background

Machine learning is a promising tool in the area of suicide prevention due to its ability to combine the effects of multiple risk factors and complex interactions. The power of machine learning has led to an influx of studies on suicide prediction, as well as a few recent reviews. Our study distinguished between data sources and reported the most important predictors of suicide outcomes identified in the literature.

Objective

Our study aimed to identify studies that applied machine learning techniques to administrative and survey data, summarize performance metrics reported in those studies, and enumerate the important risk factors of suicidal thoughts and behaviors identified.

Methods

A systematic literature search of PubMed, Medline, Embase, PsycINFO, Web of Science, Cumulative Index to Nursing and Allied Health Literature (CINAHL), and Allied and Complementary Medicine Database (AMED) to identify all studies that have used machine learning to predict suicidal thoughts and behaviors using administrative and survey data was performed. The search was conducted for articles published between January 1, 2019 and May 11, 2022. In addition, all articles identified in three recently published systematic reviews (the last of which included studies up until January 1, 2019) were retained if they met our inclusion criteria. The predictive power of machine learning methods in predicting suicidal thoughts and behaviors was explored using box plots to summarize the distribution of the area under the receiver operating characteristic curve (AUC) values by machine learning method and suicide outcome (i.e., suicidal thoughts, suicide attempt, and death by suicide). Mean AUCs with 95% confidence intervals (CIs) were computed for each suicide outcome by study design, data source, total sample size, sample size of cases, and machine learning methods employed. The most important risk factors were listed.

Results

The search strategy identified 2,200 unique records, of which 104 articles met the inclusion criteria. Machine learning algorithms achieved good prediction of suicidal thoughts and behaviors (i.e., an AUC between 0.80 and 0.89); however, their predictive power appears to differ across suicide outcomes. The boosting algorithms achieved good prediction of suicidal thoughts, death by suicide, and all suicide outcomes combined, while neural network algorithms achieved good prediction of suicide attempts. The risk factors for suicidal thoughts and behaviors differed depending on the data source and the population under study.

Conclusion

The predictive utility of machine learning for suicidal thoughts and behaviors largely depends on the approach used. The findings of the current review should prove helpful in preparing future machine learning models using administrative and survey data.

Systematic review registration

https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42022333454 identifier CRD42022333454.

Characteristics and Functions of Non-Suicidal Self-Injury That Inform Suicide Risk

OBJECTIVE

Although non-suicidal self-injury (i.e., NSSI) has been suggested as a robust risk factor of suicide, NSSI related information that is most related to suicide risk remains unclear. Commonly studied NSSI characteristics are its frequency and the number of methods endorsed. However, it may not be merely how frequent or how many different methods that matters, but "why," which alludes to the importance of NSSI functions (or why individuals engage in NSSI). Thus, this study examined how the interactions between NSSI characteristics and functions are associated with suicide risk.

METHODS

Undergraduate students (n = 820) with a lifetime history of NSSI, filled out self-report measures on NSSI and suicide risk. A hierarchical regression analysis was conducted to examine the moderation effects between four 2-way interactions (i.e., method X intrapersonal; method X interpersonal; frequency X intrapersonal; frequency X interpersonal) on suicide risk.

RESULTS

Main effects of all four independent variables were statistically significant. In terms of interactions, the intrapersonal function moderated both the effects of NSSI frequency and methods on suicide risk, whereas the interpersonal function moderated the effects of NSSI frequency on suicide risk.

CONCLUSION

Our results highlight that some NSSI related information than others are more indicative of suicide risk. In particular, the combination of NSSI functions, along with its frequency and number of methods, holds promise when assessing for current and lifetime suicidal thoughts and behaviors.

The Inter-Relationships Between Depressive Symptoms and Suicidality Among Macau Residents After the "Relatively Static Management" COVID-19 Strategy: A Perspective of Network Analysis

Background

Suicidality is a global public health problem which has increased considerably during the coronavirus disease 2019 (COVID-19) pandemic. This study examined the inter-relationships between depressive symptoms and suicidality using network analysis among Macau residents after the "relatively static management" COVID-19 strategy.

Methods

An assessment of suicidal ideation (SI), suicide plan (SP), suicide attempt (SA) and depressive symptoms was conducted with the use of individual binary response items (yes/no) and Patient Health Questionnaire (PHQ-9). In the network analysis, central and bridge symptoms were identified in the network through "Expected Influence" and "Bridge Expected Influence", and specific symptoms that were directly associated with suicidality were identified via the flow function. Network Comparison Tests (NCT) were conducted to examine the gender differences in network characteristics.

Results

The study sample included a total of 1008 Macau residents. The prevalence of depressive symptoms and suicidality were 62.50% (95% CI = 59.4-65.5%) and 8.9% (95% CI = 7.2-10.9%), respectively. A network analysis of the sample identified SI ("Suicidal ideation") as the most central symptom, followed by SP ("Suicide plan") and PHQ4 ("Fatigue"). SI ("Suicidal ideation") and PHQ6 ("Guilt") were bridge nodes connecting depressive symptoms and suicidality. A flow network revealed that the strongest connection was between S ("Suicidality") and PHQ6 ("Guilt"), followed by S ("Suicidality") and PHQ 7 ("Concentration"), and S ("Suicidality") and PHQ3 ("Sleep").

Conclusion

The findings indicated that reduction of specific depressive symptoms and suicidal thoughts may be relevant in decreasing suicidality among adults. Further, suicide assessment and prevention measures should address the central and bridge symptoms identified in this study.

Non-suicidal self-injury and suicidal ideation among adolescents: the chain-mediating role of rumination and decentering

Objective

To explore the relationship between non-suicidal self-injury and suicidal ideation in adolescents and examine the roles of rumination and decentering in that relationship.

Method

By means of a questionnaire, 175 adolescent patients in a psychiatric hospital in Fujian Province were given the Functional Assessment of Self-Mutilation: Chinese Version, Positive and Negative Suicide Ideation, Ruminative Response Scale: Chinese Version, and Experiences Questionnaire: Decentering Scale.

Results

(1) Adolescent non-suicidal self-injury was significantly positively related to suicidal ideation and rumination and significantly negatively related to decentering. Suicidal ideation was significantly positively related to rumination and significantly negatively related to decentering. Rumination was significantly negatively related to decentering. (2) Rumination and decentering played a complete chain-mediating role between non-suicidal self-injury and suicidal ideation. Non-suicidal self-injury was found to indirectly affect suicidal ideation along three pathways: the independent mediating role of rumination (the mediating effect accounted for 40.166%), independent mediating role of decentering (the mediating effect accounted for 41.274%), and chain-mediating role of rumination and decentering (the mediating effect accounted for 14.958%).

Conclusion

Adolescent non-suicidal self-injury can indirectly affect suicidal ideation through rumination and decentering. In the future, mindfulness and other methods should be used to improve individuals' levels of decentering and cultivate emotional regulation abilities, so as to reduce the incidence of non-suicidal self-injury and suicide in adolescents.

The relationship between emotional neglect and non-suicidal self-injury among middle school students in China: the mediating role of social anxiety symptoms and insomnia

 BACKGROUND: Non-suicidal self-injury (NSSI) is a vital public concern around the world, and it often starts in adolescence. Emotional neglect (EN) has been considered a distal risk factor for NSSI, but the effects of social anxiety symptoms (SA) and insomnia on this relationship have remained unclear. This study aimed to investigate the potential pathways from EN to NSSI, examining the role of SA and insomnia in this association.

METHODS

One thousand three hundred thirty seven Chinese middle school students (M_age = 13.040, SD = 0.981, 50.2% males) in China were enrolled in this cross-sectional study. Participants completed the Emotional Neglect sub-scale of Childhood Trauma Questionnaire (CTQ-SF), the Social Anxiety Scale for Adolescent (SAS-A), Athens Insomnia Scale (AIS) and non-suicidal self-injury assessment. Structural equation modelling (SEM) was used to test the possible mediation model among these variables.

RESULTS

231(17.3%) students reported NSSI history during last year and 322 (24.1%) participants reported experiences of EN. Students who experienced EN have higher rates of NSSI compared to students without EN history (29.2% vs 13.5%). EN, SA, insomnia and NSSI were positively related to each other. Furthermore, both SA and insomnia played a mediating role in the relationship between EN and NSSI, the series mediating effect of SA and insomnia on this association was also significant after controlling for demographics. Indirect effects accounted for 58.26% of the total effects (EN → NSSI).

CONCLUSIONS

Our study revealed that EN was associated with NSSI, SA and insomnia play indirect roles in the association between EN and NSSI. The findings of our research may have implications for clinicians, families, and schools in their efforts to lower the risk of NSSI in adolescents.

Examining the use of decision trees in population health surveillance research: an application to youth mental health survey data in the COMPASS study

INTRODUCTION

In population health surveillance research, survey data are commonly analyzed using regression methods; however, these methods have limited ability to examine complex relationships. In contrast, decision tree models are ideally suited for segmenting populations and examining complex interactions among factors, and their use within health research is growing. This article provides a methodological overview of decision trees and their application to youth mental health survey data.

METHODS

The performance of two popular decision tree techniques, the classification and regression tree (CART) and conditional inference tree (CTREE) techniques, is compared to traditional linear and logistic regression models through an application to youth mental health outcomes in the COMPASS study. Data were collected from 74 501 students across 136 schools in Canada. Anxiety, depression and psychosocial well-being outcomes were measured along with 23 sociodemographic and health behaviour predictors. Model performance was assessed using measures of prediction accuracy, parsimony and relative variable importance.

RESULTS

Decision tree and regression models consistently identified the same sets of most important predictors for each outcome, indicating a general level of agreement between methods. Tree models had lower prediction accuracy but were more parsimonious and placed greater relative importance on key differentiating factors.

CONCLUSION

Decision trees provide a means of identifying high-risk subgroups to whom prevention and intervention efforts can be targeted, making them a useful tool to address research questions that cannot be answered by traditional regression methods.

The influence of cannabis use disorder on post-traumatic stress disorder symptoms, depression symptoms, and non-suicidal self-injury among college students

Objective: The interrelationships between cannabis use disorder (CUD), post-traumatic stress disorder (PTSD) and depressive symptoms, and non-suicidal self-injury (NSSI) were examined. Participants: Undergraduates (N = 363) who reported using cannabis within the past six months were recruited. Method: Mediation analyses was conducted to examine if CUD symptoms were indirectly associated with greater risk for engagement in current NSSI through more severe PTSD and depression symptoms. A factor comprised of PTSD-depression symptoms was created given PTSD and depression symptoms were highly correlated. Results: Greater CUD symptom severity indirectly increased risk for current NSSI via more severe PTSD-depression symptoms, after accounting for gender and alcohol use disorder symptoms. Conclusion: The current study provides preliminary evidence for the negative consequences of CUD on college students' mental health symptoms and engagement in NSSI, and future longitudinal research is needed to examine the dynamic relationships between CUD, PTSD-depression symptoms, and NSSI over time.

NSSI questionnaires revisited: A data mining approach to shorten the NSSI questionnaires

BACKGROUND AND OBJECTIVE

Non-suicidal self-injury (NSSI) is a psychological disorder that the sufferer consciously damages their body tissues, often too severe that requires intensive care medicine. As some individuals hide their NSSI behaviors, other people can only identify them if they catch them while injuring, or via dedicated questionnaires. However, questionnaires are long and tedious to answer, thus the answers might be inconsistent. Hence, in this study for the first time, we abstracted a larger questionnaire (of 662 items in total) to own only 22 items (questions) via data mining techniques. Then, we trained several machine learning algorithms to classify individuals based on their answers into two classes.

METHODS

Data from 277 previously-questioned participants is used in several data mining methods to select features (questions) that highly represent NSSI, then 245 different people were asked to participate in an online test to validate those features via machine learning methods.

RESULTS

The highest accuracy and F1 score of the selected features-via the Genetics algorithm-are 80.0% and 74.8% respectively for a Random Forest algorithm. Cronbach's alpha of the online test (validation on the selected features) is 0.82. Moreover, results suggest that an MLP can classify participants into two classes of NSSI Positive and NSSI Negative with 83.6% accuracy and 83.7% F1-score based on the answers to only 22 questions.

CONCLUSION

While previously psychologists used many combined questionnaires to see whether someone is involved in NSSI, via various data mining methods, the present study showed that only 22 questions are enough to predict if someone is involved or not. Then different machine learning algorithms were utilized to classify participants based on their NSSI behaviors, among which, an MLP with 10 hidden layers had the best performance.

Prediction of Mental Health Problem Using Annual Student Health Survey: A Machine Learning Approach (Preprint)

BACKGROUND

One of the reasons why students go to counseling is being called on based on self-reported health survey results. However, there is no concordant standard for such calls.

OBJECTIVE

This study aims to develop a machine learning (ML) model to predict students' mental health problems in 1 year and the following year using the health survey's content and answering time (response time, response time stamp, and answer date).

METHODS

Data were obtained from the responses of 3561 (62.58%) of 5690 undergraduate students from University A in Japan (a national university) who completed the health survey in 2020 and 2021. We performed 2 analyses; in analysis 1, a mental health problem in 2020 was predicted from demographics, answers for the health survey, and answering time in the same year, and in analysis 2, a mental health problem in 2021 was predicted from the same input variables as in analysis 1. We compared the results from different ML models, such as logistic regression, elastic net, random forest, XGBoost, and LightGBM. The results with and without answering time conditions were compared using the adopted model.

RESULTS

On the basis of the comparison of the models, we adopted the LightGBM model. In this model, both analyses and conditions achieved adequate performance (eg, Matthews correlation coefficient [MCC] of with answering time condition in analysis 1 was 0.970 and MCC of without answering time condition in analysis 1 was 0.976; MCC of with answering time condition in analysis 2 was 0.986 and that of without answering time condition in analysis 2 was 0.971). In both analyses and in both conditions, the response to the questions about campus life (eg, anxiety and future) had the highest impact (Gain 0.131-0.216; Shapley additive explanations 0.018-0.028). Shapley additive explanations of 5 to 6 input variables from questions about campus life were included in the top 10. In contrast to our expectation, the inclusion of answering time-related variables did not exhibit substantial improvement in the prediction of students' mental health problems. However, certain variables generated based on the answering time are apparently helpful in improving the prediction and affecting the prediction probability.

CONCLUSIONS

These results demonstrate the possibility of predicting mental health across years using health survey data. Demographic and behavioral data, including answering time, were effective as well as self-rating items. This model demonstrates the possibility of synergistically using the characteristics of health surveys and advantages of ML. These findings can improve health survey items and calling criteria.

Examining Nonsuicidal Self-Injury Features as Motivational Moderators in the Relationship Between Hopelessness and Suicide Ideation

OBJECTIVE

Nonsuicidal self-injury (NSSI) is a robust predictor of suicide ideation and attempts, but it is not clear how and why this connection is so strong. Using the Integrated Motivational-Volitional Model of suicide as a framework, select features of NSSI were examined as motivational moderators between hopelessness and suicide ideation.

METHOD

Data were collected from 420 emerging adults (mean age = 18.9; 84% female, 92% white), all of whom had past-year NSSI. Participants completed self-report measures that assessed NSSI and suicide history, effectiveness of NSSI in achieving functions, and hopelessness; they also completed the self-injury Implicit Association Test (IAT).

RESULTS

Moderation analyses revealed that none of the interactions were significant. Additional analyses tested unconditional effects of all predictor variables and found hopelessness, self-rated future likelihood of engaging in NSSI, effectiveness of NSSI in achieving intrapersonal functions, and self-injury IAT scores were each significantly associated with suicide ideation.

CONCLUSIONS

These results reinforce the important role of hopelessness, perceived effectiveness of NSSI in achieving intrapersonal functions, acknowledgement of future NSSI, and self-identification with NSSI as potential indicators of increased suicide risk, while also emphasizing further study of other possible moderating factors.HIGHLIGHTSFeatures of NSSI did not moderate the hopelessness-suicide ideation relationship.Features of NSSI did associate with increased suicide ideation frequency.Future likelihood of, implicit association with, and intrapersonal effectiveness of NSSI were important.

A Longitudinal Study of Correlates, Discharge Disposition, and Rate of Re-presentation to Emergency Services of Adults who Engage in Non-suicidal Self-injury

OBJECTIVE

In DSM-5, non-suicidal self-injury (NSSI) and suicide attempts (SA) are classified as distinct disorders in the section of conditions for further study. However, some have questioned the validity of distinguishing NSSI from SA. The objective of this study was to longitudinally examine the correlates, discharge disposition, and rate of re-presentation to emergency services of adults who engaged in NSSI and compare them to (a) adults who engage in SA and (b) adults with no self-harm or suicidal ideation (SI).

METHOD

Data came from 4,772 presentations to adult psychiatric services in the emergency departments of tertiary care hospitals in Winnipeg, Canada between January 2009 and June 2012. Chart reviews were conducted for all presentations with NSSI (n = 158), and a sample of those with SA (n = 172) and no SH or SI (n = 173).

RESULTS

Among the adults who returned to emergency services, those who originally presented with SA re-presented significantly sooner than those who presented with NSSI. (χ²(1) = 7.457, p = 0.006). Those who originally presented with NSSI that returned to hospital did not return with repeat NSSI, but instead the majority re-presented with suicidal thoughts and SA. Further, those who re-presented with NSSI and SA were less likely to be hospitalized or to receive a referral to mental health services, and more likely to be discharged to usual care at time of initial presentation.

CONCLUSIONS

Overall, these findings indicate a trajectory of escalation of self-harm behavior for certain people who engage in NSSI, especially those who re-present to emergency services.HIGHLIGHTSA quarter of people with self-harm re-present to emergency services within four-and-a-half years.Those with suicide attempts re-presented significantly sooner than non-suicidal self-injury.There is a need for increased interventions in emergency services for those with self-harm.

Association between violence and mental distress, self-harm and suicidal ideation and attempts among young people in Malawi

BACKGROUND

Mental health problems ranging from depression to more severe acts such as self-harm or suicidal behaviours are a serious problem among adolescents and young adults. Exposure to violence during the life of young people can increase mental health issues for youth. This study examines the relationship between exposure to violence and mental health issues among youth using a nationally representative study in Malawi.

METHODS

We analysed data from the nationally representative Violence Against Children Survey from Malawi (2013) to quantify the association between exposures to violence (physical, sexual and emotional) and their relationship with mental distress, self-harm behaviours and suicidal ideation and attempts among youth aged 13-24 years. We evaluated the association of exposures to violence against children with reported mental health conditions among women and men. We used ordinal logistic regression models with appropriate survey weights to assess exposures to violence and the three outcomes of interest.

RESULTS

Children and youth aged 13-24 years exposed to violence in childhood reported higher levels of adverse mental health effects, including mental distress, self-harm behaviours and suicidal ideation and attempts. The odds of reporting these outcomes increased as the number of violence types increased.

CONCLUSIONS

Understanding the risks based on different combinations of exposures to violence in Malawi can help identify populations at higher risk and optimise violence prevention strategies.

Self-Injuries and Their Functions with Respect to Suicide Risk in Adolescents with Conduct Disorder: Findings from a Path Analysis

Non-suicidal self-injuries (NSSIs) have been identified as one of the most predictive factors of suicidal behaviours in adolescents. However, it remains unknown whether certain functions of NSSIs are associated with suicide risk, and what are the underlying mechanisms. Therefore, we aimed to investigate the association between functions of NSSIs and suicide risk in adolescents with conduct disorder (CD), which shares some common characteristics with NSSIs. Participants were 215 adolescents (155 females, 72.1%) with CD. Functions of NSSIs, depressive symptoms, the levels of impulsivity, anxiety, self-esteem and aggression were examined. There were 77 adolescents with lifetime history of NSSIs (35.8%). Among them, adolescents with lifetime history of suicide attempt were significantly more likely to report anti-dissociation and anti-suicide function of NSSIs. They had significantly higher levels of anxiety as well as significantly lower self-esteem. Higher lifetime number of NSSIs was associated with higher odds of reporting anti-dissociation and anti-suicide functions. Moreover, these two functions fully mediated the association between lifetime number of NSSIs and suicide risk after co-varying for depressive and anxiety symptoms as well as self-esteem. The present findings indicate that anti-suicide and anti-dissociation functions of NSSIs might be crucial predictors of suicide risk in adolescents with CD.

Short-term associations between nonsuicidal and suicidal thoughts and behaviors: A daily diary study with high-risk adolescents

BACKGROUND

In this daily dairy study of adolescents at elevated suicide risk, we examined proximal associations between nonsuicidal self-injury (NSSI) and suicidal thoughts as well as behaviors. We also investigated the prominence of the anti-suicide function underlying NSSI engagement, relative to intrapersonal and interpersonal motives.

METHODS

Seventy-eight adolescents (67.9% female; ages 13-17) hospitalized due to suicide risk completed daily surveys assessing NSSI and suicidal thoughts for four weeks after discharge (n=1621 observations). Suicidal behavior (actual, aborted, interrupted suicide attempts) was assessed at 1-month follow-up.

RESULTS

Over and above lifetime NSSI, adolescents who generally experienced more enduring (OR=2.54, p=<.001) and intense (OR=1.87, p=.002) suicidal ideation were more likely to engage in NSSI on a given day. Moreover, NSSI likelihood increased when adolescents experienced more enduring (OR=1.99, p<.001) and intense (OR=1.66, p<.001) ideation relative to their typical levels. This pattern was consistent for those with recent NSSI. The anti-suicide function of NSSI was frequently endorsed at hospitalization and when NSSI occurred daily (65.6% of the time), alongside the intrapersonal-negative motive (to avoid aversive states). Exploratory analyses suggest adolescents with suicidal behavior within the month after discharge experienced higher NSSI levels reported daily over the same period (Hedge's g=1.26, p=<.001).

LIMITATIONS

Daily-level associations were examined concurrently and generalizability of results is limited by sample characteristics.

CONCLUSIONS

The notable proximal associations between NSSI and suicidal thoughts and behaviors, as well as the prominence of the anti-suicide function, point to the importance of intervention efforts targeting these intersecting phenomena among adolescents at elevated suicide risk.

AI enabled suicide prediction tools: a qualitative narrative review

Background: Suicide poses a significant health burden worldwide. In many cases, people at risk of suicide do not engage with their doctor or community due to concerns about stigmatisation and forced medical treatment; worse still, people with mental illness (who form a majority of people who die from suicide) may have poor insight into their mental state, and not self-identify as being at risk. These issues are exacerbated by the fact that doctors have difficulty in identifying those at risk of suicide when they do present to medical services. Advances in artificial intelligence (AI) present opportunities for the development of novel tools for predicting suicide.

Method: We searched Google Scholar and PubMed for articles relating to suicide prediction using artificial intelligence from 2017 onwards.

Conclusions: This paper presents a qualitative narrative review of research focusing on two categories of suicide prediction tools: medical suicide prediction and social suicide prediction. Initial evidence is promising: AI-driven suicide prediction could improve our capacity to identify those at risk of suicide, and, potentially, save lives. Medical suicide prediction may be relatively uncontroversial when it pays respect to ethical and legal principles; however, further research is required to determine the validity of these tools in different contexts. Social suicide prediction offers an exciting opportunity to help identify suicide risk among those who do not engage with traditional health services. Yet, efforts by private companies such as Facebook to use online data for suicide prediction should be the subject of independent review and oversight to confirm safety, effectiveness and ethical permissibility.

Nonsuicidal Self-injury, Suicide Planning, and Suicide Attempts Among High-risk Adolescents Prior to Psychiatric Hospitalization

The purpose of this study was to understand the trajectories of nonsuicidal self-injury (NSSI) and suicide plans (SP) in the 90 days prior to inpatient hospitalization, understand the role of NSSI and SP in predicting suicide attempts (SA) on a given day, and to test the interaction between NSSI and SP in predicting same-day SA. Participants included 69 adolescents (77% female, 65% white, 77% Non-Hispanic/Latinx, M_age = 15.77 SD_age = 1.00) from an inpatient psychiatric unit. Past 90 day NSSI, SP, and SA were measured using the Columbia Suicide Severity Rating Scale and Timeline Follow Back. First, mixed effect models were conducted to assess trajectories of NSSI and SP leading up to inpatient hospitalization. The odds of NSSI remained relatively stable prior to hospitalization (OR = 1.01, 95% CI [1.00,1.02]). The odds of SP increased in the 90 days prior to hospitalization (OR = 1.04, 95% CI [1.02,1.05]) with each day associated with a 4% increase in the odds of making a SP. Second, random effect models were conducted to predict the odds of same-day SA from NSSI and SP. When adolescents endorsed either NSSI (OR = 2.99, p < .001) or a SP (OR = 77.13, p < .001) there was elevated odds of same-day SA. However, the presence of both NSSI and SP on a given day did not increase risk of SA on that same day. For this high-risk clinical sample of suicidal adolescents who drink alcohol, odds of SP increased in the days leading up to psychiatric hospitalization, but NSSI remained stable. On days when adolescents reported NSSI or SP, they had an increased odds of same-day SA. These results underscore the importance of frequent monitoring of NSSI and SP among high-risk adolescents who drink alcohol to prevent suicide attempts.

Nonsuicidal self-injury thoughts and behavioural characteristics: Associations with suicidal thoughts and behaviours among community adolescents

BACKGROUND

We consider whether nonsuicidal self-injury (NSSI) thoughts - in the absence of any NSSI behaviour - are associated with suicidal thoughts and behaviours among adolescents, before examining whether characteristics of NSSI behaviour are associated with greater suicidal thoughts and behaviours.

METHODS

Adolescents (n = 2,057, M age =15.56) recruited from secondary schools reported their lifetime history of NSSI, suicidal thoughts and behaviours, NSSI characteristics, and NSSI functions. Receiver Operating Characteristic analysis test whether NSSI status and characteristics of NSSI behaviour are diagnostic of clinically elevated suicidal thoughts and behaviours (a score of 7 or higher on the Suicide Behaviours Questionnaire-Revised). Regression models test whether NSSI functions predict greater suicidality.

RESULTS

Adolescents with NSSI thoughts and those with NSSI behaviour were more likely to report lifetime suicidal ideation and past-year suicide plan(s) than adolescents with no history of NSSI. In addition, adolescents with a history of NSSI were more likely to report a lifetime history of suicide attempt(s) as well as past-year suicide attempt(s) than adolescents with no history of NSSI. A greater number of NSSI methods, requiring medical assistance for NSSI injuries, and engaging in NSSI for self-punishment, anti-suicide, and sensation-seeking functions were associated with greater severity of suicidal thoughts and behaviours.

LIMITATIONS

Data are cross-sectional, limiting inferences about causality.

CONCLUSIONS

Findings emphasize the importance of NSSI thoughts and characteristics of NSSI behaviour in understanding the complex relationship between NSSI and suicidal thoughts and behaviours among adolescents.

Using ordinal regression for advancing the understanding of distinct suicide outcomes

OBJECTIVE

As recent advances in suicide research have underscored the importance of studying distinct suicide outcomes (i.e., suicidal thinking vs. behavior), there is a need to consider the theoretical meaningfulness of our statistical approach(es). As an alternative to more popular statistical methods, we introduce ordinal regression, detailing specific forms that are well-aligned to examine outcomes specific to suicide research.

METHOD

Ordinal regression models allow for assessment of the influences of covariates on the experience of lower (i.e., suicidal ideation) to higher (i.e., suicidal planning) suicide risk outcomes.

RESULTS

As an empirical application, we fit a sequential ordinal regression model with 17 theoretically selected covariates and modeled category specific effects for each covariate.

CONCLUSIONS

Results detailed from depression and presence of nonsuicidal self-injury demonstrate the utility of ordinal regression in consideration of transitions across suicide outcomes. Ordinal regression models may be particularly informative in identifying risk factors unique to each suicide outcome, which has the potential to meaningfully inform theoretical models of suicide and suicide risk prediction.

Prevalence of Suicidality in Major Depressive Disorder: A Systematic Review and Meta-Analysis of Comparative Studies

Background: Suicidality is common in major depressive disorder (MDD), but there has been no systematic review published about all aspects of suicidality. This meta-analysis and systematic review compared the prevalence of the whole range of suicidality comprising suicidal ideation (SI), suicide plan (SP), suicide attempt (SA), and completed suicide (CS), between patients with MDD and non-MDD controls. Methods: Major international (PubMed, PsycINFO, Web of Science, EMBASE) and Chinese (Chinese Nation Knowledge Infrastructure and WANFANG) databases were systematically and independently searched from their inception until January 12, 2021. Results: Fifteen studies covering 85,768 patients (12,668 in the MDD group and 73,100 in the non-MDD group) were included in the analyses. Compared to non-MDD controls, the odds ratios (ORs) for lifetime, past month, past year, and 2-week prevalence of SI in MDD were 2.88 [95% confidence interval (CI) = 0.30-27.22, p = 0.36], 49.88 (95% CI = 2-8.63, p < 0.001), 13.97 (95% CI = 12.67-15.41, p < 0.001), and 24.81 (95% CI = 15.70-39.22, p < 0.001), respectively. Compared to non-MDD controls, the OR for lifetime SP in MDD was 9.51 (95% CI = 7.62-11.88, p < 0.001). Compared to non-MDD controls, the ORs of lifetime and past-year prevalence of SA were 3.45 (95% CI = 1.58-7.52, p = 0.002), and 7.34 (95% CI = 2.14-25.16, p = 0.002), respectively, in MDD patients. No difference in the prevalence of CS between MDD and controls was found (OR = 0.69, 95% CI = 0.23-2.02, p = 0.50). Conclusions: MDD patients are at a higher risk of suicidality, compared to non-MDD controls. Routine screening for a range of suicidality should be included in the management of MDD, followed by timely treatment for suicidal patients. Systematic Review Registration: Identifier [INPLASY202120078].

Artificial Intelligence and Suicide Prevention: A Systematic Review of Machine Learning Investigations

Suicide is a leading cause of death that defies prediction and challenges prevention efforts worldwide. Artificial intelligence (AI) and machine learning (ML) have emerged as a means of investigating large datasets to enhance risk detection. A systematic review of ML investigations evaluating suicidal behaviors was conducted using PubMed/MEDLINE, PsychInfo, Web-of-Science, and EMBASE, employing search strings and MeSH terms relevant to suicide and AI. Databases were supplemented by hand-search techniques and Google Scholar. Inclusion criteria: (1) journal article, available in English, (2) original investigation, (3) employment of AI/ML, (4) evaluation of a suicide risk outcome. N = 594 records were identified based on abstract search, and 25 hand-searched reports. N = 461 reports remained after duplicates were removed, n = 316 were excluded after abstract screening. Of n = 149 full-text articles assessed for eligibility, n = 87 were included for quantitative synthesis, grouped according to suicide behavior outcome. Reports varied widely in methodology and outcomes. Results suggest high levels of risk classification accuracy (>90%) and Area Under the Curve (AUC) in the prediction of suicidal behaviors. We report key findings and central limitations in the use of AI/ML frameworks to guide additional research, which hold the potential to impact suicide on broad scale.

Methods Matter: Nonsuicidal Self-Injury in the Form of Cutting is Uniquely Associated with Suicide Attempt Severity in Patients with Substance Use Disorders

OBJECTIVE

Despite the robust relation between nonsuicidal self-injury (NSSI) and suicidal behaviors, there is considerable heterogeneity in NSSI characteristics that may have implications for the strength of the NSSI-suicide attempt relation. Past research highlights the relevance of both more severe NSSI methods and NSSI in the form of cutting in particular to suicide attempts. To further explore the relations of specific NSSI methods to suicide attempts, this study examined relations between different NSSI methods (i.e., those involving cutting, burning, blunt force, or resulting in minor tissue damage) and both overall suicide attempts and two indicators of suicide attempt severity (attempts requiring medical attention and nonambivalent attempts) in 203 substance-dependent patients in residential treatment.

METHOD

Participants were administered questionnaires and semistructured interviews assessing the variables of interest.

RESULTS

Cutting was the only NSSI method to demonstrate significant relations with overall suicide attempts, nonambivalent suicide attempts, and suicide attempts requiring medical attention. Results remained when considering relevant covariates.

CONCLUSION

Results highlight an important link between NSSI methods involving cutting and suicide attempts among patients with substance use disorders. Findings also highlight the importance of more fine-grained characterizations of NSSI subgroups.

Does function predict persistence? Nonsuicidal self-injury among adolescents during and after hospitalization

Nonsuicidal self-injury (NSSI) is a prevalent, concerning behavior among adolescents. Importantly, NSSI can serve a variety of functions. Some adolescents engage in NSSI to fulfill automatic or self-oriented functions (e.g., cutting to avoid internal negative states), whereas others engage in NSSI to serve social functions (e.g., cutting to communicate with others). This study tests whether self-reported reasons for engaging in NSSI, hereafter referred to as NSSI functions, predict NSSI thoughts and behaviors during and after hospitalization among adolescent psychiatric inpatients. Endorsement of both automatic and social NSSI functions, as well as positive and negative reinforcement subtypes, was assessed at hospital admission. Results showed that endorsement of overall automatic function predicted which adolescents engaged in NSSI behavior during hospitalization. Moreover, automatic and social functions showed distinct predictive patterns, such that automatic functions corresponded to greater likelihood of NSSI-related thoughts and behaviors whereas social functions mainly corresponded to reduced likelihood of NSSI-related outcomes. Of note, NSSI functions were less predictive of NSSI-related outcomes after hospital discharge. These findings suggest that identifying adolescent inpatients' reasons for NSSI engagement may meaningfully distinguish those at higher risk (and those at lower risk) of NSSI persistence during their hospital stay.

Anti-Suicide Function of Nonsuicidal Self-Injury in Female Inpatient Adolescents

BACKGROUND

There have been numerous studies investigating the relationship between nonsuicidal self-injury (NSSI) and suicidality. On the one side, NSSI is an important risk factor for suicidality, including suicidal thoughts and behaviors. On the other side, it has been suggested that one function of NSSI might be as a coping mechanism that can help individuals in the short term avoid suicide. The present study investigated the relationship between suicidality and NSSI in female inpatient adolescents, focusing on NSSI as an anti-suicide strategy.

METHODS

Subjects were 56 female adolescents, aged 12-18 years (M = 15.95 years, SD = 1.27), recruited from different inpatient child and adolescent psychiatric units. All participants fulfilled the DSM-5 research criteria for nonsuicidal self-injury disorder (NSSI-D). To assess suicidality, NSSI-D, and current and past diagnoses, a structured clinical interview was conducted.

RESULTS

NSSI as a short-term coping strategy for avoiding suicide was indicated by one third (32.1%) of the participants. Before participants engaged in NSSI, the anti-suicide function was reported more frequently than reducing interpersonal problems (d = -.59). Getting relief from negative emotions and inducing positive feelings were reported at the same frequency as avoiding suicide before NSSI. Participants engaging in NSSI to avoid suicide and those reporting other functions did not significantly differ regarding other NSSI characteristics, suicidality, or psychopathology. Results indicate that the anti-suicide function significantly predicts the duration of current suicidal ideation (β = .557).

CONCLUSION

This study provides preliminary support for the idea that NSSI is frequently used by female adolescents with NSSI-D to avoid suicide. Given the high co-occurrence of NSSI and suicidality, our results underline the importance of clinical assessment of suicidality and several NSSI functions, including the anti-suicide function, in adolescents with NSSI.

Daily patterns in nonsuicidal self-injury and coping among recently hospitalized youth at risk for suicide

Among adolescents at high suicide risk, using a daily diary design, this study examined: (1) the co-occurrence between nonsuicidal self-injury (NSSI) and suicidal ideation, (2) the link between NSSI and coping, and (3) endorsement of using NSSI to cope with suicidal ideation (anti-suicide function). Thirty-four adolescents hospitalized due to suicide risk (76% female; ages 13-17) responded to daily surveys for four weeks after discharge (n = 650 observations). NSSI was positively associated with suicidal ideation at the between- (i.e. relative to others) and within-person (relative to adolescents' own average) levels of analysis. When NSSI and suicidal thoughts co-occurred, adolescents used NSSI to cope with thoughts of suicide on nearly all occasions. While adolescents did not use less adaptive coping when they engaged in NSSI (i.e. within-person), youth who utilized more coping strategies in general (i.e. between-person) had lower probability of NSSI. The probability of NSSI also decreased when adolescents perceived coping to be helpful and for youth who generally tended to perceive coping as helpful. Findings offer fine-grained insights about the intersection of NSSI, suicidal thoughts, and coping among high-risk adolescents, adding to the body of research highlighting the benefit of broadening adolescents' coping strategies as well as assessing their perceived utility.

An examination of individual forms of nonsuicidal self-injury

Nonsuicidal self-injury (NSSI) is a growing public health concern, and there is an increasing need to better characterize and identify severe NSSI behavior. One readily accessible, yet understudied, avenue for improving the assessment of NSSI severity is through the examination of individual forms, or methods, of the behavior. The present study aimed to address this gap in the literature by investigating the relationship between 12 different NSSI methods with three NSSI severity indicators and three distinct suicidal thoughts and behaviors among 1,436 undergraduate students with a history of NSSI (70.90% female, M age = 20.69, SD = 3.32). Results across six decision tree analyses highlighted the use self-hitting / punching, in addition to cutting oneself, as the most informative NSSI methods for differentiating outcome severity. Gender differences were only found for the outcome of suicidal ideation. The present study provides preliminary evidence that the examination of individual NSSI methods may be useful in identifying individuals at risk for negative correlates of NSSI, including NSSI-related hospital visits, unintended serious injury, and suicidal behavior. Upon replication in longitudinal work, findings have important clinical utility by providing a potential marker of prognosis and the need for higher levels of care.

Does nonsuicidal self-injury prospectively predict change in depression and self-criticism?

Few studies have investigated nonsuicidal self-injury (NSSI) as a predictor of outcomes other than suicidal self-injury, severely limiting our understanding of this behavior's full range of consequences. Three independent studies were used to examine the prospective association between NSSI and two outcomes: depressive symptoms and self-criticism. Data were collected from samples of (1) adults with past-month NSSI, (2) adults with lifetime NSSI, and (3) adults with past-year NSSI. Studies included one-month and six-month follow-up periods. Results were tested in an internal meta-analysis. Results suggested that NSSI did not prospectively predict changes in self-criticism. No changes in depressive symptoms were seen over shorter follow-up periods; however, NSSI predicted increases in depressive symptoms at six-month follow-up in one sample. The internal meta-analysis indicated a null relationship between NSSI and prospective internalizing symptoms. Future research should replicate these findings and examine a broader range of outcomes of NSSI to better understand its complex relationship to psychopathology.

The use of machine learning in the study of suicidal and non-suicidal self-injurious thoughts and behaviors: A systematic review

BACKGROUND

Machine learning techniques offer promise to improve suicide risk prediction. In the current systematic review, we aimed to review the existing literature on the application of machine learning techniques to predict self-injurious thoughts and behaviors (SITBs).

METHOD

We systematically searched PsycINFO, PsycARTICLES, ERIC, CINAHL, and MEDLINE for articles published through February 2018.

RESULTS

Thirty-five articles met criteria to be included in the review. Included articles were reviewed by outcome: suicide death, suicide attempt, suicide plan, suicidal ideation, suicide risk, and non-suicidal self-injury. We observed three general aims in the use of SITB-focused machine learning analyses: (1) improving prediction accuracy, (2) identifying important model indicators (i.e., variable selection) and indicator interactions, and (3) modeling underlying subgroups. For studies with the aim of boosting predictive accuracy, we observed greater prediction accuracy of SITBs than in previous studies using traditional statistical methods. Studies using machine learning for variable selection purposes have both replicated findings of well-known SITB risk factors and identified novel variables that may augment model performance. Finally, some of these studies have allowed for subgroup identification, which in turn has helped to inform clinical cutoffs.

LIMITATIONS

Limitations of the current review include relatively low paper sample size, inconsistent reporting procedures resulting in an inability to compare model accuracy across studies, and lack of model validation on external samples.

CONCLUSIONS

We concluded that leveraging machine learning techniques to further predictive accuracy and identify novel indicators will aid in the prediction and prevention of suicide.

Balancing Performance and Interpretability: Selecting Features with Bootstrapped Ridge Regression

Informctticists sometimes attempt to predict chronic healthcare events that are not fully understood. The resulting models often incorporate copious numbers of predictors derived across diverse datasets. This approach may yield desirable performance characteristics, but it sacrifices interpretability and portability. The Bootstrapped Ridge Selector (BoRidge) offers a tool to balance performance with interpretability. Compared to two modern feature selection methods, Bootstrapped LASSO regression (BoLASSO) and a minimal-redundancy-maximal-relevance selector (mRMR), the BoRidge bested them for binary classification on artificially generated data (sensitivity: 0.83, specificity:0.72) versus BoLASSO (sensitivity: 0.1, specificity:1) and mRMR (sensitivity: 0.69, specificity: 0.69). On a dataset used to validate a published suicide risk prediction model, the BoRidge selected an equally precise model to the publication, with far fewer predictors (114 versus the 1,538 used in the published model). The BoRidge has the potential to simplify classification models for complex problems, making them easier to translate and act upon.

The associations between non-suicidal self-injury and first onset suicidal thoughts and behaviors

BACKGROUND

Theoretical and empirical literature suggests that non-suicidal self-injury (NSSI) is an important correlate of suicide risk. The present study was designed to evaluate: (a) whether NSSI is associated with increased odds of subsequent onsets of suicidal thoughts and behaviors (STB) independent of common mental disorders, (b) whether NSSI is associated with increased risk of transitioning from suicide ideation to attempt, and (c) which NSSI characteristics are associated with STB after NSSI.

METHOD

Using discrete-time survival models, based on retrospective age of onset reports from college students (n = 6,393, 56.8% female), we examined associations of temporally prior NSSI with subsequent STB (i.e., suicide ideation, plan, and attempt) controlling mental disorders (i.e., MDD, Broad Mania, GAD, Panic Disorder, and risk for Alcohol Dependence). NSSI characteristics associated with subsequent STB were examined using logistic regressions.

RESULTS

NSSI was associated with increased odds of subsequent suicide ideation (OR = 2.8), plan (OR = 3.0), and attempt (OR = 5.5) in models that controlled for the distribution of mental disorders. Further analyses revealed that NSSI was associated with increased risk of transitioning to a plan among those with ideation, as well as attempt among those with a plan (ORs = 1.7-2.1). Several NSSI characteristics (e.g., automatic positive reinforcement, earlier onset NSSI) were associated with increased odds of experiencing STB.

LIMITATIONS

Surveys relied on self-report, and thus, there is the potential for recall bias.

CONCLUSIONS

This study provides support for the conceptualization of NSSI as a risk factor for STB. Investigation of the underlying pathways accounting for these time-ordered associations is an important avenue for future research.