Machine learning prediction of suicide attempts in major depression: Feature selection and model development using...

Area of Science:

• Psychiatry
• Clinical Psychology
• Epidemiology

Background:

• Major depressive disorder (MDD) is a significant public health concern associated with increased risk of suicidal attempts (SA).
• Identifying reliable risk factors for SA in MDD patients is crucial for effective prevention and intervention strategies.

Purpose of the Study:

• To explore and identify independent risk factors associated with suicidal attempts (SA) in patients diagnosed with major depressive disorder (MDD).
• To develop a predictive model for assessing suicide risk in MDD patients.

Main Methods:

• A cross-sectional analysis was performed on 3247 MDD patients from the National Survey on Symptomatology of Depression (NSSD).
• Boruta's Algorithm and Lasso regression were employed for predictor screening.
• Logistic regression and nomogram were utilized to identify independent risk factors and visualize their impact on SA risk.

Main Results:

• A history of SA was reported by 12.1% of the patients.
• Twenty significant risk factors for SA were identified, with self-harm and a sense of decreased ability being particularly prominent.
• Independent predictors of increased SA risk included hospitalization frequency, SSRI and antipsychotic use, psychosis, self-harm, gastrointestinal complaints, weight gain, helplessness, unhappiness, and derealization.

Conclusions:

• The study identified several key predictors of suicidal attempts in MDD patients, including self-harm and feelings of helplessness.
• These findings can inform the development of clinical assessment tools for suicide risk stratification in MDD.
• The identified factors may also guide future research into the underlying mechanisms of suicidal behavior in depression.