Prevalence and factors associated with suicidal ideation based on the Malaysia National Health and Morbidity Survey:...

This study estimated the prevalence of past-two-week suicidal ideation and examined its demographic, clinical, and health-related correlates in a nationally representative Malaysian sample using machine learning.

The National Health and Morbidity Survey (NHMS) 2019 in Malaysia comprised a nationally representative sample of 11,674 adults aged 18 and above. Suicidal ideation was assessed using Item 9 of the Patient Health Questionnaire-9 (PHQ-9), "Thoughts that you would be better off dead or of hurting yourself in some way" during the past two weeks. Exposure variables included depressive symptoms (PHQ-9 Items 1-8), self-rated health, acute physical health problems, chronic bodily pain, and medical check-up history. Three machine learning models, Random Forest, Least Absolute Shrinkage and Selection Operator (LASSO) logistic regression, and K-Nearest Neighbours (KNN), were applied to identify significant predictors of suicidal ideation, with feature importance assessed using permutation importance and SHapley Additive exPlanations (SHAP) analyses.

The estimated prevalence of past-two-week suicidal ideation was 2.3%. Depressive symptoms, especially low mood, worthlessness, appetite loss, fatigue, and concentration difficulties, were the strongest correlates. Among models, LASSO logistic regression showed the best balance of performance (Area Under the Curve [AUC] 0.80, accuracy 98%, sensitivity 100%), followed by Random Forest (AUC 0.88) and KNN (AUC 0.74). Key predictors included depressive symptoms, health-related factors, and sociodemographic variables such as being male, older, married, less educated, or having a lower income.

The results underscore the need for targeted suicide prevention strategies that address both mental health and social determinants within the Malaysian context.