Factors related to the length of stay for major depressive disorder patients in China: A real-world retrospective study.

Background: As numerous patients with depression have to be hospitalized because of various reasons, the demand far exceeds the limited bed count in the psychiatry department. Controlling the length of stay (LOS) of the patient is gradually being considered an effective method to alleviate this problem. Given the lack of statistical evidence of the LOS of patients with major depressive disorder (MDD) in China and the strain on the limited psychiatric resources, the purpose of our study was to investigate the LOS of patients with MDD among in-patient samples and to analyze related factors of the LOS in China by building a regression model. Method: The data were exported from the electronic medical record system. A total of three categories of independent variables were enrolled in our study, namely, demographic, clinical, and biochemical. Univariate analysis and binominal regression analysis were applied comprehensively to find the factors related to the LOS among MDD samples. The discrimination accuracy of the model was evaluated by the receiver operating characteristic (ROC) analysis. ROC analysis indicated that the discrimination accuracy of our model was acceptable (AUC = 0.790, 95% CI = 0.714-0.865, Result: A total of 254 patients were finally brought into analysis after filtering. Regression analysis indicated that abnormal LDL was the only risk factor of long LOS (OR = 3.352, 95% CI = 1.087-10.337, Conclusion: Up to our knowledge, our research is the first study to show the potential factors related to the LOS from various domains, especially biochemical indexes, and the effect of drugs, among clinical patients with MDD in China. Our results could provide a theoretical reference for efficient psychiatry hospitalization management and prioritization of allocating medical resources. Future studies are required for updating independent variables which are potentially related to the LOS and verifying existing results in a larger sample.