AI and Machine Learning in Antibiotic Dose Prediction
Author Information
Author(s): Iria Varela-Rey, Enrique Bandín-Vilar, Francisco José Toja-Camba, Antonio Cañizo-Outeiriño, Francisco Cajade-Pascual, Marcos Ortega-Hortas, Víctor Mangas-Sanjuan, Miguel González-Barcia, Irene Zarra-Ferro, Cristina Mondelo-García, Anxo Fernández-Ferreiro, Yukihiro Hamada
Primary Institution: Health Research Institute of Santiago de Compostela (IDIS)
Hypothesis
The application of AI and machine learning techniques can improve pharmacokinetic modeling and dose prediction of antibiotics.
Conclusion
AI techniques show promising results in pharmacokinetic modeling, but their effectiveness compared to classical methods is still uncertain.
Supporting Evidence
- AI techniques can process large datasets quickly, which is beneficial for predicting antibiotic doses.
- Most studies used AI techniques like XGBoost and neural networks for dose prediction.
- AI models showed acceptable prediction error values compared to conventional methods.
Takeaway
This study looks at how computers can help doctors figure out the right amount of antibiotics to give patients, which can be really tricky.
Methodology
A systematic search was conducted in EMBASE, OVID, and PubMed databases to analyze studies using AI for antibiotic dose prediction.
Potential Biases
The retrospective design of many studies may introduce selection biases.
Limitations
The studies analyzed are mostly retrospective and may suffer from biases, and there is a lack of standard metrics for comparison.
Participant Demographics
The studies included a mix of adult and pediatric patients, with some focusing on critically ill patients.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website