Related Experiment Video
Updated: May 6, 2026
09:10
Determining the Functional Status of the Corticospinal Tract Within One Week of Stroke
Published on: February 22, 2020
8.5K
Prediction Models for Post-Stroke Cognitive Impairment: A Systematic Review and Meta-Analysis
Yifang Yang1, Yajing Chen1, Yiyi Yang1
1School of Nursing, Evidence-Based Nursing Center, Lanzhou University, Lanzhou, China.
Public Health Nursing (Boston, Mass.)
|January 9, 2025
Summary
This review identifies key predictors for post-stroke cognitive impairment (PSCI), such as age and hypertension. Developing accurate PSCI prediction models is crucial for early intervention and improved patient outcomes.
Area of Science:
- Neurology
- Public Health
- Biostatistics
Background:
- Stroke is a leading cause of adult disability worldwide.
- Post-stroke cognitive impairment (PSCI) significantly affects daily life and social function.
- Accurate PSCI risk prediction models are vital for early identification and prevention.
Purpose of the Study:
- Systematically review and analyze existing PSCI prediction models.
- Identify key risk factors associated with PSCI.
- Evaluate the performance and limitations of current models.
Main Methods:
- Comprehensive literature search across major databases (PubMed, Cochrane, Embase).
- Independent data extraction and risk of bias assessment using CHARMS and PROBAST tools.
- Systematic analysis of 20 identified articles on PSCI prediction models.
Main Results:
- Twenty articles reported PSCI prediction models with incidence rates from 8% to 75%.
- Model performance varied, with AUC values ranging from 0.66 to 0.969 (development) and 0.763 to 0.893 (validation).
- Key predictors identified include age, diabetes, hs-CRP, hypertension, and homocysteine.
Conclusions:
- Existing PSCI prediction models show promise but often lack external validation and exhibit heterogeneity.
- Recommended use of comprehensive predictive factors for screening high-risk patients.
- Future research should focus on refining models with novel variables and robust validation.
