Development and Validation of a Community-Based Prediction Model for Depression in Elderly Patients with Diabetes: A Cross-Sectional Study

Shanshan Li,Le Zhang,Boyi Yang,Yi Huang,Yuqi Guan,Nanbo Huang,Yingnan Wu,Wenshuo Wang,Qing Wang,Haochen Cai,Yong Sun,Zijun Xu,Qin Wu
DOI: https://doi.org/10.2147/DMSO.S465052
2024-07-01
Abstract:Background: In elderly diabetic patients, depression is often overlooked because professional evaluation requires psychiatrists, but such specialists are lacking in the community. Therefore, we aimed to create a simple depression screening model that allows earlier detection of depressive disorders in elderly diabetic patients by community health workers. Methods: The prediction model was developed in a primary cohort that consisted of 210 patients with diabetes, and data were gathered from December 2022 to February 2023. The independent validation cohort included 99 consecutive patients from February 2023 to March 2023. Multivariable logistic regression analysis was used to develop the predictive model. We incorporated common demographic characteristics, diabetes-specific factors, family structure characteristics, the self-perceived burden scale (SPBS) score, and the family APGAR (adaptation, partnership, growth, affection, resolution) score. The performance of the nomogram was assessed with respect to its calibration (calibration curve, the Hosmer-Lemeshow test), discrimination (the area under the curve (AUC)), and clinical usefulness (Decision curve analysis (DCA)). Results: The prediction nomogram incorporated 5 crucial factors such as glucose monitoring status, exercise status, monthly income, sleep disorder status, and the SPBS score. The model demonstrated strong discrimination in the primary cohort, with an AUC of 0.839 (95% CI, 0.781-0.897). This discriminative ability was further validated in the validation cohort, with an AUC of 0.857 (95% CI, 0.779-0.935). Moreover, the nomogram exhibited satisfactory calibration. DCA suggested that the prediction of depression in elderly patients with diabetes mellitus was of great clinical value. Conclusion: The prediction model provides precise and user-friendly guidance for community health workers in preliminary screenings for depression among elderly patients with diabetes.
What problem does this paper attempt to address?