Dynamic evolution of readmission risk factors across short-, medium-, and long-term horizons in type 2 diabetes: a machine learning-based predictive modeling study with SHAP interpretability - Takeaways - MDSpire

Dynamic evolution of readmission risk factors across short-, medium-, and long-term horizons in type 2 diabetes: a machine learning-based predictive modeling study with SHAP interpretability

  • By

  • Lei Li

  • Sheng Jiang

  • June 22, 2026

  • 0 min

Share

  • 1

    This study developed nine machine learning models to predict readmission in 12,041 T2DM patients over 30, 60, and 365 days.

  • 2

    Artificial neural networks achieved the highest AUROC for 30-day and 60-day predictions, while random forest excelled for 365-day predictions.

  • 3

    SHAP analysis indicated that age was the primary predictor for 30-day readmissions, while inflammatory markers were key for 60-day readmissions.

  • 4

    Diabetes-specific chronic complications were the dominant risk factors for 365-day readmissions, highlighting the evolution of risk factors over time.

  • 5

    The study emphasizes the need for time-specific model selection and dynamic risk monitoring in T2DM patients to improve clinical outcomes.

Original Source(s)

Dynamic evolution of readmission risk factors across short-, medium-, and long-term horizons in type 2 diabetes: a machine learning-based predictive modeling study with SHAP interpretability

  • by Lei Li, Sheng Jiang

  • June 22, 2026

Related Content

Anti-Diabetic Retinopathy Molecular Mechanism of Dihuang Yinzi: Insights from Network Pharmacology, Metabolomics, and Microbiome Analysis

AI Screening for Diabetic Retinopathy

Editorial: Dynamics of endocrine system dysfunctions and neuro-ophthalmic manifestations: early diagnostic to therapeutic implications