Prediction Models for Post-Stroke Hospital Readmission: A Systematic Review

Area of Science:

• Medical Informatics
• Clinical Prediction Models
• Health Services Research

Background:

• Stroke readmission poses a significant burden on healthcare systems.
• Accurate prediction of post-stroke readmission is crucial for resource allocation and patient management.
• Existing prediction models require rigorous evaluation for clinical utility.

Purpose of the Study:

• To systematically review and evaluate the predictive performance and methodological quality of post-stroke readmission prediction models.
• To identify key predictors associated with stroke readmission.
• To provide guidance on selecting and adapting risk assessment tools for clinical practice.

Main Methods:

• A comprehensive literature search was performed up to February 1, 2024.
• Two researchers independently screened studies and extracted data using the CHARMS checklist.
• Included 11 studies encompassing 16 distinct prediction models.

Main Results:

• Models showed a wide range in predictive performance (AUC 0.520-0.940).
• Common predictors included length of stay, hypertension, age, and functional disability.
• Methodological quality was often limited, with a high risk of bias in data analysis and insufficient external validation.

Conclusions:

• Existing stroke readmission models demonstrate potential but suffer from generalizability issues due to methodological limitations.
• External validation and iterative adaptation of current models are recommended over developing new ones.
• Future models should be customized for local contexts and potentially enhanced with new predictors and user-friendly interfaces.