Clinical Prediction Model for 1-Year Ischemic Stroke Recurrence in 738 Patients

738 patients with first-ever ischemic stroke admitted to a tertiary hospital in China in 2021. This dataset supports a clinical prediction model developed using six admission variables to estimate the 1-year recurrence risk, where 96 patients (13.0%) experienced recurrence. The model was developed and internally validated by DaYing Fan, with results published on figshare under a CC-BY-4.0 license.

Use Cases

• Validate stroke recurrence prediction models based on the six predictors (NIHSS score, age, admission systolic blood pressure, uric acid, apolipoprotein A1, neutrophil percentage).
• Benchmark machine learning model performance against the reported logistic regression model's AUC of 0.764.
• Study associations between clinical biomarkers and stroke outcomes based on the described predictor-outcome relationships.

Strengths

• Model performance metrics are provided, including an optimism-corrected AUC of 0.750 and a Brier score of 0.0981.
• The dataset is derived from a clearly defined cohort of 738 eligible patients with a specific outcome measure (13.0% recurrence rate).
• Predictor handling is documented, with missing data addressed via multiple imputation.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the underlying data is unknown, which may limit suitability assessment.
• Data reflects a single-center, one-year cohort from China, which may limit generalizability.

Provenance

Source

Retrospective cohort study at a tertiary hospital in China.

Collection Method

Clinical data collection from consecutive patients admitted between January and December 2021.

Time Range

2021

Freshness

Last updated 2026-04 23 04:21:56; freshness should be verified.

Geography

China