The Differences Between Individuals Engaging in Nonsuicidal Self-Injury and Suicide Attempt Are Complex (vs. Complicated or Simple)

Abstract

Background: Why do some people engage in nonsuicidal self-injury (NSSI) while others attempt suicide? One way to advance knowledge about this question is to shed light on the differences between people who engage in NSSI and people who attempt suicide. These groups could differ in three broad ways. First, these two groups may differ in a simple way, such that one or a small set of factors is both necessary and sufficient to accurately distinguish the two groups. Second, they might differ in a complicated way, meaning that a specific set of a large number of factors is both necessary and sufficient to accurately classify them. Third, they might differ in a complex way, with no necessary factor combinations and potentially no sufficient factor combinations. In this scenario, at the group level, complicated algorithms would either be insufficient (i.e., no complicated algorithm produces good accuracy) or unnecessary (i.e., many complicated algorithms produce good accuracy) to distinguish between groups. This study directly tested these three possibilities in a sample of people with a history of NSSI and/or suicide attempt. Method: A total of 954 participants who have either engaged in NSSI and/or suicide attempt in their lifetime were recruited from online forums. Participants completed a series of measures on factors commonly associated with NSSI and suicide attempt. To test for simple differences, univariate logistic regressions were conducted. One theoretically informed multiple logistic regression model with suicidal desire, capability for suicide, and their interaction term was considered as well. To examine complicated and complex differences, multiple logistic regression and machine learning analyses were conducted. Results: No simple algorithm (i.e., single factor or small set of factors) accurately distinguished between groups. Complicated algorithms constructed with crossvalidation methods produced fair accuracy; complicated algorithms constructed with bootstrap optimism methods produced good accuracy, but multiple different algorithms with this method produced similar results. Conclusions: Findings were consistent with complex differences between people who engage in NSSI and suicide attempts. Specific complicated algorithms were either

Predicting self-harm within six months after initial presentation to youth mental health services: A machine learning study

“…The problem extends to other biological [15] or clinical [16,17] risk factors which have similarly weak predictive value. Selfharm is likely to be driven by the complex interplay between a broad range of social, biological, psychological, and contextual factors rather than any one or simple set of factors [18]. Further, the influence these factors have on self-harm is probably dynamic over time.…”

Evidence of Inflated Prediction Performance: A Commentary on Machine Learning and Suicide Research

“…Walsh et al (2017) found that random forests, a machine-learning algorithm, outperformed standard logistic regression models (as indexed by the area under the receiver-operator-characteristic curve [AUC]) when used to predict risk of suicide attempts (i.e., random-forests AUCs were .80-.84 vs. .66-.68 for multiple logistic regression models; for more details, see Walsh et al, 2017). Other studies compared random forests and logistic regression for predicting suicidal thoughts and behaviors (Ribeiro et al, 2019), predicting nonsuicidal selfinjury over short time windows (within 2-4 weeks; Fox et al, 2019), identifying variables that distinguish those who think about suicide from those who attempt suicide (Huang et al, 2020b), and differentiating nonsuicidal selfinjury and suicide (Huang et al, 2020a). In each of these studies, the authors found that random forests outperformed single and multiple logistic regression and concluded that results supported the idea that the outcome of interest was too complex to be explained through simple (e.g., linear) modeling approaches.…”

“…There are a multitude of approaches researchers use to assess and thus prevent overfitting, with the most common being crossvalidation (CV) and bootstrap sampling. Of particular interest, given its use in a number of applications that have found support for the superiority of machinelearning models over traditional statistical models (e.g., Huang et al, 2020aHuang et al, , 2020bMatsuki et al, 2016;Ribeiro et al, 2019;Walsh et al, 2017), is the optimism-corrected bootstrap.…”

Validation of the German capability for suicide questionnaire (GCSQ) in a high-risk sample of suicidal inpatients

“…[ 6 , 12 , 13 ] (Forkmann T, Glaesmer H, Paashaus L, Rath D, Schönfelder A, Juckel G, Assion J, Stengler K, Teismann T: Interpersonal-psychological theory of suicide: a prospective examination, submitted)). This is all the more true since two recent studies have comprehensively shown that no simple algorithm (i.e., a single factor such as capability for suicide) can accurately distinguish between suicide attempters and suicide ideators [ 44 ] or between (lifetime) suicide attempters and individuals engaging in non-suicidal self-injury [ 45 ]. These findings imply that recent ideation-to-action models do not sufficiently take into account the complexity of suicidal behaviour.…”