Systematic Review of Venous Thromboembolism Risk Prediction Models for Stroke Patients

Seven risk prediction models for venous thromboembolism (VTE) in stroke patients were evaluated in this systematic review and meta-analysis. The pooled VTE incidence across studies was 20.8%, and the pooled area under the curve (AUC) for model performance was 0.87. The review, authored by Lina Fu and last updated in April 2026, is available as a 15.6 KB document under a CC-BY-4.0 license.

Use Cases

• Benchmarking new VTE prediction models against a reported pooled AUC of 0.87.
• Informing clinical study design based on a reported pooled VTE incidence of 20.8%.
• Assessing methodological rigor in predictive modeling using the PROBAST checklist framework mentioned in the description.
• Identifying research gaps for models requiring external validation, as noted for all seven included models.

Strengths

• Includes a meta-analysis providing a pooled VTE incidence of 20.8% and a pooled AUC of 0.87.
• Systematically reviewed 2,726 records from multiple databases including PubMed, Embase, and CNKI.
• Explicitly assesses model risk of bias using the PROBAST checklist.
• Registered with PROSPERO (CRD42024603132), indicating a structured review protocol.

Limitations

• The dataset is a 15.6 KB document, indicating a very limited scope of raw data or results.
• Column-level documentation is absent; field semantics must be inferred after download.
• The description notes all included models were at high risk of bias and lacked independent external validation.

Provenance

Source

Systematic review and meta-analysis by author Lina Fu, published on figshare.

Collection Method

Systematic search across eight databases (e.g., PubMed, Embase, CNKI) from inception to September 1, 2025, with meta-analysis performed using Stata 17.0.

Time Range

Literature search covered from database inception to September 1, 2025.

Freshness

Last updated 2026-04-23 05:34:51.