Scholarlink

Background <p>Most informal caregivers of older people with dementia (PwD) experience a high level of psychological distress, which significantly impacts caregivers’ psychosomatic health. Effective identification and timely intervention for psychological distress are of great significance for improving the psychosomatic health of informal caregivers. However, there is a lack of predictive models for psychological distress in this population. This study aims to develop and interpret nomogram and eXtreme Gradient Boosting (XGBoost) models to predict the psychological distress of informal caregivers of PwD.</p> Methods <p>This cross-sectional study collected 485 pairs of PwD and informal caregivers from October 2023 to September 2024. Participants were randomly divided into training and validation cohorts (7:3 ratio). We constructed nomogram and XGBoost models to predict psychological distress. The nomogram was subsequently developed into an online web calculator, and SHapley Additive exPlanations (SHAP) was employed to enhance the XGBoost model’s interpretability.</p> Results <p>The final models included six predictors: caregivers’ social support, caregivers’ employment situation, caregivers’ average daily care duration, caregivers’ monthly family income, caregivers’ self-efficacy and PwD’s chronic disease. The area under the curve (AUC) of nomogram and XGBoost models in the validation cohort were 0.882 (95%CI: 0.825-0.939) and 0.884 (95%CI: 0.829-0.939) respectively. Two models exhibited good calibration and clinical utility. SHAP analysis identified caregivers’ social support as the most influential predictor of psychological distress in XGBoost model.</p> Conclusion <p>The nomogram and XGBoost models developed in this study offer valuable tools for predicting psychological distress in informal caregivers of PwD. The nomogram-based web calculator could be an effective tool for predicting the risk of psychological distress in practice. Our interpretable XGBoost model provides valuable insights into individualized intervention strategies for psychological distress.</p> Trial registration <p>Not applicable. This study is an observational study without involving any interventions or treatment methods. Therefore, this study does not meet the definition of a clinical trial and does not require registration of a clinical trial number.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Predicting psychological distress in informal caregivers of older people with dementia: development and interpretation of a nomogram and XGBoost model

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月22日

查看全文
  • Wenli Shi,
  • Siwei Tian,
  • Yixin Wang,
  • Chan Cai,
  • Bing Cheng,
  • Chenyang Li,
  • Xuan Liu,
  • Chongqing Shi

摘要

Background

Most informal caregivers of older people with dementia (PwD) experience a high level of psychological distress, which significantly impacts caregivers’ psychosomatic health. Effective identification and timely intervention for psychological distress are of great significance for improving the psychosomatic health of informal caregivers. However, there is a lack of predictive models for psychological distress in this population. This study aims to develop and interpret nomogram and eXtreme Gradient Boosting (XGBoost) models to predict the psychological distress of informal caregivers of PwD.

Methods

This cross-sectional study collected 485 pairs of PwD and informal caregivers from October 2023 to September 2024. Participants were randomly divided into training and validation cohorts (7:3 ratio). We constructed nomogram and XGBoost models to predict psychological distress. The nomogram was subsequently developed into an online web calculator, and SHapley Additive exPlanations (SHAP) was employed to enhance the XGBoost model’s interpretability.

Results

The final models included six predictors: caregivers’ social support, caregivers’ employment situation, caregivers’ average daily care duration, caregivers’ monthly family income, caregivers’ self-efficacy and PwD’s chronic disease. The area under the curve (AUC) of nomogram and XGBoost models in the validation cohort were 0.882 (95%CI: 0.825-0.939) and 0.884 (95%CI: 0.829-0.939) respectively. Two models exhibited good calibration and clinical utility. SHAP analysis identified caregivers’ social support as the most influential predictor of psychological distress in XGBoost model.

Conclusion

The nomogram and XGBoost models developed in this study offer valuable tools for predicting psychological distress in informal caregivers of PwD. The nomogram-based web calculator could be an effective tool for predicting the risk of psychological distress in practice. Our interpretable XGBoost model provides valuable insights into individualized intervention strategies for psychological distress.

Trial registration

Not applicable. This study is an observational study without involving any interventions or treatment methods. Therefore, this study does not meet the definition of a clinical trial and does not require registration of a clinical trial number.