Predicting depression risk in COPD patients: a model based on NHANES 2007-2012 data.

Abstract

Objective: Patients with chronic obstructive pulmonary disease (COPD) are at an elevated risk for depression. However, effective predictive tools for identifying high-risk individuals are currently lacking. This study aims to develop a nomogram for predicting depression risk in COPD patients.

Methods: A total of 1,671 participants from NHANES 2007-2012 were included in the study. The data were divided into training and testing sets in a 7:3 ratio. LASSO regression was employed to identify the optimal predictors in the training set. Subsequently, univariate and multivariate logistic regression analyses were conducted to determine independent predictors for constructing the nomogram. The model was then evaluated using the C-index, calibration curve, Hosmer-Lemeshow test, and decision curve analysis (DCA). And conducted a sensitivity analysis to assess the robustness of the model's predictive performance. Finally, the Youden index was used to determine the optimal prediction threshold.

Results: Eight predictors were selected for the model, including age, gender, marital status, poverty income ratio (PIR), body mass index(BMI), sleep disorder, work limitation, and social barriers. The C-index for the training and test sets were 0.71 and 0.72, respectively, indicating significant classification performance. All four evaluation methods demonstrated that the model has strong discriminatory ability, calibration, and clinical utility. Additionally, the threshold for predicting risk and the corresponding score from the nomogram were 0.57 and 93, respectively. The sensitivity analysis demonstrated the robustness of the results, with the model exhibiting good discrimination and calibration across different gender and age groups.

Conclusion: The nomogram has potential value in the preliminary prediction of depression risk in COPD patients, facilitating the early initiation of preventive interventions for depression. Future studies should focus on optimizing the model and validating its performance in larger, more diverse populations.