Predicting Suicidal Behavior Without Asking About Suicidal Ideation: Machine Learning and the Role of Borderline Personality Disorder Criteria

OBJECTIVE

Identifying predictors contributing to suicide risk could help prevent suicides via targeted interventions. However, using only known risk factors may not yield accurate enough results. Furthermore, risk models typically rely on suicidal ideation, even though people often withhold this information.

METHOD

This study examined the contribution of various predictors to the accuracy of six machine learning models for identifying suicidal behavior in a prison population (n = 353), including borderline personality disorder (BPD) and antisocial personality disorder (APD) criteria, and compared how excluding data about suicidal ideation affects accuracy.

RESULTS

Results revealed that gradient tree boosting accurately identified individuals with suicidal behavior, even without relying on questions about suicidal ideation (AUC = 0.875, F1 = 0.846). Furthermore, the model maintained this accuracy with only 29 predictors. Meeting five or more diagnostic criteria of BPD was an important risk factor for suicidal behavior. APD criteria, in the presence of other predictors, did not substantially improve accuracy. Additionally, it may be possible to implement a decision tree model to assess individuals at risk of suicide, without focusing upon suicidal ideation.

CONCLUSIONS

These findings highlight that modern classification algorithms do not necessarily require information about suicidal ideation for modeling suicide and self-harm behavior.

Chinese versions of the Interpersonal Needs Questionnaire: Psychometric properties, measurement invariance across gender and cultures

Associations between unmet interpersonal needs and different aspects of suicide have been observed in both Western and non-Western cultures using the Interpersonal Needs Questionnaire (INQ). However, measurement invariance is a prerequisite for comparing differences between culturally different groups, and to date, no studies have examined measurement invariance of INQ across cultures. This study aimed to (a) validate Chinese versions of the INQ, (b) assess measurement invariance across gender for the Chinese INQ, (c) assess measurement invariance across Australian and Chinese cultures for the INQ, and (d) comprehensively assess the association of interpersonal needs with suicide ideation. A sample of 469 Australian undergraduates and a sample of 854 Chinese undergraduates were used in this study. For testing measurement invariance across gender, the sample of Chinese undergraduates was split by gender into the Chinese male and Chinese female samples. Five versions of INQ (10-, 12-, 15-, 18- and 25-item) were tested. The 10- and 15-item Chinese INQ demonstrated adequate psychometric properties through various analyses (i.e., reliability, confirmatory factor analysis, and structural equation modeling) and also demonstrated measurement invariance across gender via multigroup confirmatory factor analysis. The 10-item INQ demonstrated measurement invariance across Australian and Chinese cultures. Of the two interpersonal factors, only perceived burdensomeness was significantly associated with suicide ideation. Multigroup structural equation modeling demonstrated that perceived burdensomeness may be a greater risk factor of suicide among Australian undergraduates than among Chinese undergraduates. Practical and theoretical contributions of this study are discussed.