Risk Prediction Model for Pediatric Catheter-Related Thrombosis

A study of 188 critically ill children with central venous catheters, where 31 developed catheter-related thrombosis. The dataset supports a nomogram prediction model developed from clinical risk factors, achieving an area under the ROC curve of 0.952. The research was authored by Xiaolan Zhou.

Use Cases

• Develop a risk prediction nomogram using independent predictors like age, parenteral nutrition status, D-dimer, and fibrinogen levels.
• Analyze the incidence rate of catheter-related thrombosis, which was 16.5% in the cohort of 188 pediatric patients.
• Validate the discriminative ability of a clinical model using the reported area under the ROC curve of 0.952.
• Compare clinical characteristics, such as catheter type and D-dimer levels, between thrombosis and non-thrombosis patient groups.

Strengths

• Model validation shows strong predictive performance with an area under the ROC curve of 0.952.
• The dataset is derived from a specific clinical study of 188 critically ill pediatric patients.
• Clear identification of independent risk factors: age, parenteral nutrition, D-dimer, and fibrinogen levels.

Limitations

• The dataset is very small at 15.3 KB, likely containing summary statistics rather than raw patient-level data.
• Limited to a single study's patient cohort of 188 children, which may affect generalizability.
• The primary data format is a DOCX document, which is not structured for direct computational analysis.

Provenance

Source

Figshare, authored by Xiaolan Zhou.

Collection Method

Clinical data collection from 188 critically ill children with central venous catheters, categorized into thrombosis and non-thrombosis groups.

Freshness

Last updated on 2026-03-24.