Development and validation of a predictive model for readmission of major depressive disorder patients with suicide attempts

The study aimed to investigate the determinants associated with readmission in first-hospitalised major depressive disorder (MDD) patients with suicide attempts. The study included a cohort of 3977 patients diagnosed with MDD, as retrieved from the hospital information system (HIS). The data were split into a training set and a validation set using R software. To develop a stepwise regression model, the multivariate logistic analysis included significant factors (p threshold <0.2 in univariate analysis).(1) There was a significant difference in the incidence of readmission between first-hospitalised MDD patients with suicide attempts and those without; (2) Length of stay, course of disease, somatic symptoms, psychosis symptoms, anxiety symptoms, alcohol/tobacco dependence, family history of psychiatric disease, continued use of antidepressants combined with antipsychotics or not after discharge, and total bilirubin (TBIL) were included in these variables to build a predictive model in R software; (3) The area under the curve (AUC) was 0.783 for the training set and 0.778 for the validation set. The Hosmer-Lemesho (H-L) test for calibration was <i>p</i> = 0.3521 for the training set and <i>p</i> = 0.8497 for the validation set. The study revealed a significantly increased risk of readmission in first-hospitalised MDD patients with suicide attempts compared to their counterparts.