Development and Validation of Risk Assessment Models for Anxiety and Depression in Chinese Children with Childhood Household Dysfunction.

BACKGROUND

Children exposed to childhood household dysfunction (CHD) are at high risk for anxiety and depression. However, few tools are available for early identification. This study developed two risk assessment models for anxiety and depression in children with CHD to enable early detection and intervention.

METHODS

A cross-sectional study recruited 1276 children with CHD from Changsha, Hunan, who were randomly divided into training and validation sets (7:3). Bidirectional stepwise multivariate logistic regressions were used to develop models based on socio-demographic, individual, family, and societal levels. Model performance was assessed using receiver operating characteristic (ROC) curves, area under the curve (AUC), sensitivity, specificity, and accuracy (ACC).

RESULTS

Among the participants, 736 (57.8 %) had depression risk, and 892 (70.0 %) had anxiety risk. The Anxiety Risk Model included seven variables: gender, grade, rumination, adolescent life events, number of childhood trauma experiences, positive coping style, and sleep quality. It demonstrated strong performance with an AUC of 0.875 (95 % CI: 0.852-0.897) in the training set and 0.864 (95 % CI: 0.826-0.902) in the validation set, with a sensitivity of 74.3 %, specificity of 79.6 %, and an accuracy of 75.1 %. The Depression Risk Model included five variables: grade, rumination, adolescent life events, positive coping style, and sleep quality, with an AUC of 0.895 (95 % CI: 0.874-0.916) in the training set and 0.869 (95 % CI: 0.829-0.908) in the validation set, with sensitivity of 75.1 %, specificity of 80.9 %, and accuracy of 76.8 %.

CONCLUSION

The two risk assessment models offer valuable insights into the early identification of children at risk of depression and anxiety within similar urban adolescent populations. These findings necessitate further validation using longitudinal designs within larger and more diverse populations to confirm the models' generalizability and robustness before recommendations for wider clinical application can be made.