Vancomycin Pharmacokinetics in Critically Ill Children for ML-PPK Modeling

Data from 821 critically ill children in two tertiary centers in China (2013–2023) supports a machine-learning/population-pharmacokinetic hybrid model. The dataset includes 1,767 vancomycin concentration measurements and 29 candidate variables for predicting individual clearance and volume of distribution. The research was authored by Jihui Chen and published on figshare in May 2026.

Use Cases

• Developing predictive models for vancomycin clearance based on body weight, renal function, and cardiothoracic surgery status.
• Training algorithms to estimate volume of distribution using routinely available clinical variables.
• Validating machine-learning models against Bayesian posterior pharmacokinetic estimates.
• Identifying minimal predictor sets for patient-specific initial dosing regimens.

Strengths

• Includes 821 patient records from a 10-year period (2013–2023).
• Contains 1,767 vancomycin concentration measurements for model training.
• Model performance was validated on a held-out test set and an external cohort, with reported R² values of 0.85 for clearance and 0.95 for volume of distribution.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect geographic bias inherent to the two Chinese tertiary centers.

Provenance

Source

figshare

Collection Method

Retrospective analysis of clinical data from two tertiary centers.

Time Range

2013–2023

Freshness

Last updated 2026-05-26 04:39:12; freshness should be verified.

Geography

China