Prediction Models for Post-Stroke Delirium: A Systematic Review with an Exploratory Meta-Analysis of Predictors
-
By
-
Ma, Weiya
-
Ma, Sumin
-
Tang, Qiaomin
-
Sun, Yuanyuan
-
Hu, Chen
-
June 5, 2026
-
Clinical Scorecard: Systematic Review and Exploratory Meta-Analysis of Predictors in Post-Stroke Delirium Prediction Models
At a Glance
| Category | Detail |
|---|
| Condition | Post-Stroke Delirium (PSD) |
| Key Mechanisms | Age, NIHSS score, neutrophil-to-lymphocyte ratio, visual impairment, infection |
| Target Population | Patients who have experienced a stroke |
| Care Setting | Clinical settings involving stroke management |
Key Highlights
- Sixteen studies with 24 models included in the analysis
- Pooled AUC for model discrimination was 0.83
- Common significant predictors identified include age, NIHSS, and infection
- High overall risk of bias in all studies assessed
- Calibration performance was acceptable in six studies
Guideline-Based Recommendations
Diagnosis
- Future studies should adhere to unified PSD diagnosis criteria
Management
- Employ robust validation strategies for prediction models
Monitoring & Follow-up
- Explore advanced modeling techniques to improve model reliability
Risks
- High risk of bias due to methodological shortcomings
Patient & Prescribing Data
Patients post-stroke at risk for delirium
Clinical utility of existing models remains uncertain
Clinical Best Practices
- Conduct systematic assessments of predictors in PSD
- Utilize the Prediction Model Risk of Bias Assessment Tool (PROBAST) for evaluation
- Ensure calibration and clinical utility are evaluated in future models
Related Resources & Content
