Development and validation of nursing-oriented risk prediction models for anxiety and depression in hospitalized patients with chronic kidney disease: a retrospective cross-sectional study in Southwest China

TL;DR

This study develops and validates models to help nurses identify anxiety and depression in hospitalized chronic kidney disease patients using clinical and psychosocial data.

Contribution

New nursing-oriented predictive models for anxiety and depression in CKD patients using routinely available data.

Findings

33.2% of patients showed anxiety symptoms and 35.8% showed depressive symptoms upon admission.

Models achieved strong discrimination with AUC scores of 0.830 for anxiety and 0.829 for depression.

Key predictors included age, income, hospitalization frequency, and psychosocial factors like sleep disturbance and family accompaniment.

Abstract

Psychological disorders such as anxiety and depression are common but often underrecognized among patients with chronic kidney disease (CKD), posing challenges for inpatient nursing care. This study aimed to identify key risk factors and develop predictive models to assist clinical nurses in early psychological risk identification and intervention planning.

This retrospective cross-sectional study for model development included 1,420 adult inpatients with CKD stages 1–5 admitted to a tertiary hospital in Southwest China from March 2023 to March 2025. Anxiety and depression were assessed using the Generalized Anxiety Disorder 7-item scale (GAD-7) and Patient Health Questionnaire 9-item scale (PHQ-9) within 48 hours of admission. Nursing-relevant demographic, clinical, and psychosocial data were extracted from electronic health records. Multivariate logistic regression was used to identify predictors. Two nomograms were developed, and model performance was assessed via ROC curves, calibration plots, and 1,000 bootstrap validations.

Screening-positive anxiety and depressive symptoms (GAD-7/PHQ-9 ≥5) were observed in 33.2% and 35.8% of patients, respectively. Significant predictors of anxiety included younger age, female sex, low income, frequent hospitalizations, hypoalbuminemia, sleep disturbance, diabetes, and absence of family accompaniment. Similar predictors were found for depression, along with low education and dialysis status. Both models demonstrated strong discrimination (AUC = 0.830 for anxiety; 0.829 for depression) and good calibration. The nomograms allow bedside nurses to estimate psychological risk using routinely available data.

Anxiety and depression are highly prevalent among hospitalized CKD patients in Southwest China and are associated with modifiable psychosocial and clinical factors. The validated nursing-oriented prediction models offer practical tools to support early risk stratification and targeted psychological care planning in nephrology nursing practice.

Figures