Construction of a nomogram risk prediction model for depressive symptoms in elderly breast cancer patients

Abstract

The primary objective of this study was to identify the factors associated with the development of depressive symptoms in elderly breast cancer (BC) patients and to construct a nomogram model for predicting these symptoms. We recruited 409 patients undergoing BC treatment in the breast departments of two tertiary-level hospitals in Jiangsu Province from November 2023 to April 2024 as our study cohort. Participants were categorized into depressed and non-depressed groups based on their clinical outcomes. Univariate and multivariate logistic regression analyses were employed to identify independent risk factors for depression among BC patients. Multivariate analysis revealed that monthly income, pain score, family support score, and physical activity score significantly influenced the onset of depression in older BC patients ( P < 0.05).The risk prediction model, constructed using these identified factors, demonstrated excellent discriminatory power, as evidenced by an area under the ROC curve (AUC) of 0.824. The maximum Youden index was 0.627, with a sensitivity of 90.60%, specificity of 72.10%, and a diagnostic threshold value of 1.501. The results of the Hosmer-Lemeshow goodness-of-fit test (χ² = 3.181, P = 0.923) indicated that the model fit the data well. The calibration curve for the model closely followed the ideal curve, suggesting a strong fit and high predictive accuracy. Our nomogram model exhibited superior predictive performance, enabling healthcare professionals to identify high-risk patients early and implement preventative measures to mitigate the development of depressive symptoms. This study is a cross-sectional study that lacks longitudinal data and has a small sample size. Future research could involve larger samples, multicenter studies, and prospective designs to build better clinical predictive models.