Prevalence, risk factors and clinical characteristics of suicidal ideation in Chinese patients with depression

Background: Suicide risk is greatly increased in depression. Detection of those at risk is clinically important. Hence, this study aimed to evaluate the prevalence and identify independent risk factors associated with suicidal ideation (SI) in a widespread symptomatology within and outside DSM framework. Methods: This study was part of the National Survey on Symptomatology of Depression (NSSD) which was designed to investigate the magnitude of symptoms of current major depressive episode in China. Stepwise multivariable logistic regression was performed to examine the independent risk factors for SI, including variables that are statistically significant in univariate analysis. Receiver operating characteristic (ROC) was used to evaluate the performance of the regression model. Results: A total of 3275 patients (1293 males and 1982 females) were included in our analysis. Of these, 1750 patients (53.4%) had SI. Independent risk predictors included crying (P = 0.000; odds ratio = 1.827), helplessness (P = 0.000; odds ratio = 1.514), worthlessness (P = 0.001; odds ratio = 1.359), hopelessness (P = 0.000; odds ratio = 1.805), unusually restless (P = 0.005; odds ratio = 1.276), self-harm (P = 0.000; odds ratio = 3.385), mood-incongruent psychosis (P = 0.000; odds ratio = 2.782), feeling losing control of oneself (P = 0.009; odds ratio = 1.352), hypersomnia (P = 0.000; odds ratio = 1.805), sensory system complaints (P = 0.000; odds ratio = 1.546), derealization (P = 0.006; odds ratio = 1.580), guilt (P = 0.002; odds ratio = 1.332), suicidal attempts (P = 0.000; odds ratio = 2.841), male gender (P = 0.001; odds ratio = 0.756), the total course of depression (P = 0.010; odds ratio = 1.003) in the regression model. In addition, the areas under the curve of the ROC and the accuracy for the regression model were 0.80 and 0.76, respectively. Conclusions: This study provided an effective risk model for SI in MDD and indicated that all these factors in our model allow better the employment of preventative measures.