Purpose: Suicide is a serious complication for patients with major depressive disorder (MDD). However, there has not been a broadly accepted scoring system for the assessment of suicidal risk in these patients. This study developed a new model to estimate the risk of suicide in patients with MDD through clinical and sociodemographic factors.
Methods: A total of 288 patients with MDD who made their first visit to our outpatient department were enrolled. Objective variables were thoroughly assessed and evaluated. Suicidal ideation was used as the dependent variable. Univariable logistic regression was used for the selection of variables. Multivariable logistic regression was performed for the selected variables to build our model, and an ROC curve was generated for analysis.
Results: Past suicidal attempt (p<0.001), number of failed antidepressant categories (p=0.001), disease onset earlier than 25 years old (p=0.011), high disease severity (p=0.002), previous psychiatric ward admission (p=0.001), and recurrent MDD (p=0.045) were found to lead to higher suicide risks. A six-item model including these factors was then derived from the multivariable logistic regression (area under curve = 0.768, p < 0.001).
Conclusion: Six clinical variables were included in our model. This model can serve as a valuable reference to prompt the modification and invention of alternative rating scales in the future.