The construction and validation of the novel nomograms for the risk prediction of prenatal depression: a cross-sectional study

Abstract

BackgroundNomograms are superior to traditional multivariate regression models in the competence of quantifying an individual’s personalized risk of having a given condition. To date, no literature has been found to report a quantified risk prediction model for prenatal depression. Therefore, this study was conducted to investigate the prevalence and associated factors of prenatal depression. Moreover, two novel nomograms were constructed for the quantitative risk prediction.MethodsIn this cross-sectional study, the participants were recruited using convenience sampling and administered with the research questionnaires. The prevalence of prenatal depression was calculated with a cutoff point of ≥ 10 in the 8-item Patient Health Questionnaire. Univariate and multivariate binomial logistic regression models were subsequently employed to identify the associated factors of prenatal depression. Two nomograms for the risk prediction were constructed and multiple diagnostic parameters were used to examine their performances.ResultsThe prevalence of prenatal depression was 9.5%. Multivariate binomial logistic regression model based on sociodemographic, health-related, and pregnancy-related variables (model I) suggested that unemployment, poor relationship with partners, antecedent history of gynecologic diseases, unplanned pregnancy, an earlier stage of pregnancy, and more severe vomiting symptoms were associated with increased risk of prenatal depression. In the regression model that further included psychosocial indicators (model II), unemployment, antecedent history of gynecologic diseases, unplanned pregnancy, an earlier stage of pregnancy, and a higher total score in the Pregnancy Stress Rating Scale were found to be associated with prenatal depression. The diagnostic parameters suggested that both nomograms for the risk prediction of prenatal depression have satisfactory discriminative and predictive efficiency and clinical utility. The nomogram based on model II tended to have superior performances and a broader estimating range and that based on model I could be advantageous in its ease of use.ConclusionsThe prevalence of prenatal depression was considerably high. Risk factors associated with prenatal depression included unemployment, poor relationship with partners, antecedent history of gynecologic diseases, unplanned pregnancy, an earlier stage of pregnancy, more severe vomiting symptoms, and prenatal stress. The risk prediction model I could be used for fasting screening, while model II could generate more precise risk estimations.