# Depression in left-behind adolescents from single-parent families: a nomogram based on multidimensional risk factors

## TL;DR

This study creates a predictive model to identify depression risk in left-behind adolescents from single-parent families using multiple factors.

## Contribution

The study introduces a nomogram model integrating multidimensional risk factors for depression in a specific vulnerable adolescent group.

## Key findings

- Key predictors of depression include gender, age, family satisfaction, and mobile phone use.
- The nomogram model showed good accuracy with AUC values of 0.771 and 0.759 in training and validation sets.
- The model provides a tool for early screening and targeted mental health interventions for at-risk adolescents.

## Abstract

Depression is a significant issue affecting adolescents’ mental health. While depression research is relatively extensive, studies focusing on left-behind adolescents from single-parent families remain limited. Due to their unique family structure, this group is more susceptible to multiple stressors, increasing their risk of depression.

This study aims to construct a predictive model based on a nomogram to identify the multidimensional characteristics of depression risk among left-behind adolescents from single-parent families, providing theoretical and practical evidence for early screening and targeted mental health interventions.

Cross-sectional data from the China Psychological Health Guardian Project (CPHG) were utilized to select samples of left-behind adolescents aged 12–20 years from single-parent families (N = 3731). Key variables were identified using Lasso regression, followed by the optimization of the model through binary logistic regression. A nomogram prediction model was then constructed based on significant variables.

The study identified gender, age, duration of parental separation, family satisfaction, parental education levels, substance dependence, weekday sleep duration, weekend mobile phone use duration, and chronic diseases as key predictors of depression risk. The nomogram model demonstrated good discriminatory and predictive accuracy, with AUC values of 0.771 and 0.759 in the training and validation sets, respectively.

By integrating multidimensional variables, this study developed a predictive model for depression risk among left-behind adolescents from single-parent families, providing scientific evidence for the early identification and intervention of high-risk individuals.