External Validation of Three DVT Risk Models in 1,270 Acute Stroke Patients

Lina Fu's dataset contains results from a single-center cohort study validating three deep vein thrombosis risk prediction models for acute stroke patients. The study enrolled 1,270 patients from the Stroke Center of South China Hospital of Shenzhen University between January 2023 and January 2025. Performance metrics for the Shen Xiaofang, Lu Qiufang, and Xi Pan models are provided, including AUC, Brier scores, and calibration curves.

Use Cases

• Benchmarking predictive model performance based on reported AUC, Brier score, and calibration metrics.
• Supporting clinical decision-making for DVT risk assessment based on the described net benefit analysis.
• Comparing the discrimination and calibration of different risk models for a specific patient cohort.
• Informing model selection for clinical practice based on external validation results.

Strengths

• Dataset reports performance metrics for three distinct models, enabling direct comparison.
• Study includes 1,270 patient records with a clear DVT incidence rate of 17.08%.
• Validation uses multiple assessment methods: AUC, Brier score, calibration curves, and decision curve analysis.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Data reflects a single-center design, which may limit generalizability as noted in the conclusion.

Provenance

Source

Lina Fu via figshare.

Collection Method

Consecutive enrollment of patients diagnosed with acute stroke at a single hospital center.

Time Range

January 2023 to January 2025.

Freshness

Last updated 2026-04-10 05:29:10; freshness should be verified.

Geography

Stroke Center of South China Hospital of Shenzhen University (single-center, Shenzhen, China).