RESUMO
BACKGROUND: The predictive accuracy of suicidal behaviour has not improved over the last decades. We aimed to explore the potential of machine learning to predict future suicidal behaviour using population-based longitudinal data. METHOD: Baseline risk data assessed within the Scottish wellbeing study, in which 3508 young adults (18-34 years) completed a battery of psychological measures, were used to predict both suicide ideation and suicide attempts at one-year follow-up. The performance of the following algorithms was compared: regular logistic regression, K-nearest neighbors, classification tree, random forests, gradient boosting and support vector machine. RESULTS: At one year follow up, 2428 respondents (71%) finished the second assessment. 336 respondents (14%) reported suicide ideation between baseline and follow up, and 50 (2%) reported a suicide attempt. All performance metrics were highly similar across methods. The random forest algorithm was the best algorithm to predict suicide ideation (AUC 0.83, PPV 0.52, BA 0.74) and the gradient boosting to predict suicide attempt (AUC 0.80, PPV 0.10, BA 0.69). LIMITATIONS: The number of respondents with suicidal behaviour at follow up was small. We only had data on psychological risk factors, limiting the potential of the more complex machine learning algorithms to outperform regular logistical regression. CONCLUSIONS: When applied to population-based longitudinal data containing multiple psychological measurements, machine learning techniques did not significantly improve the predictive accuracy of suicidal behaviour. Adding more detailed data on for example employment, education or previous health care uptake, might result in better performance of machine learning over regular logistical regression.
Assuntos
Ideação Suicida , Tentativa de Suicídio , Humanos , Modelos Logísticos , Estudos Longitudinais , Aprendizado de Máquina , Adulto JovemRESUMO
Evidence suggests that suicidal behaviour arises from one's attempt to escape from unbearable situations or unbearable thoughts and feelings. These feelings of entrapment are usually assessed via the 16-item Entrapment Scale, but this is too long for routine use in clinical practice. The aim of this study was to develop a brief version of the full scale that reliably assesses entrapment. We used data collected from a clinical sample (n = 497) of patients following hospital-treated self-harm and a population-based sample (n = 3457) of young adults. Four items were selected that had both the highest factor loading and discriminatory parameters and that covered the theoretical constructs of internal and external entrapment. Correlations between the 4-item short-form and the 16-item full scale were nearly perfect (0.94 for the clinical sample, 0.97 for the population-based sample). When comparing the correlations between the short-form and the full scale with other clinical and psychological scales, the correlations were nearly identical. The 4-item Entrapment Scale Short-Form (E-SF) will provide very comparable information about entrapment for each respondent as the full scale will do. However, its brevity will increase the likelihood that the assessment of entrapment will be implemented into everyday clinical practice.