Antibiotic resistance prediction with an attention-based bi-LSTM clinical decision support system
Abstract
Background and objective: Antimicrobial resistance is recognized by the World Health Organization as a significant global health threat. In clinical practice, the accurate identification of bacterial susceptibility to antibiotics is crucial. However, clinical laboratories often take several days to complete this process and, in the meantime, physicians rely on probabilistic and empirical reasoning, coupled with local hospital guidelines. Methods: In this work, we propose an attention-based bidirectional-Long Short-Term Memory recurrent neural network as a clinical decision support system to predict antibiotic resistance at the patient's bedside prior to the arrival of final antimicrobial testing results from the laboratory. More precisely, the model gives predictions at each stage of the bacterial identification process for a set of 47 single antibiotics and combinations, to support clinicians in their prescribing decision. Results: Great results were achieved, with a mean area under the receiver operating characteristic curve and a mean area under the precision-recall curve reaching up to 0.9. The attention mechanism was used to visualize the importance attributed to each feature and to better interpret the prediction results. Conclusion: The model has been integrated into a user-friendly and responsive web application, accessible on both mobile phones and desktops, to be used as a prototype clinical decision support system.
Related
- Artificial Intelligence in the Fight Against Antimicrobial Resistance: Applications in Diagnosis, Susceptibility Prediction, Precision Dosing, and Antibiotic Discovery.
- Bacteriophage and Antibiotic Resistance Are Positively Associated across a Phylogenetically Diverse Set of Clinical Pseudomonas aeruginosa Isolates
- Response to comment on “Using genomic data and machine learning to predict antibiotic resistance: A tutorial paper”
- Genomic Insights into Multi-Drug-Resistant, Klebsiella pneumoniae Isolated from an Infertile Woman: A Comprehensive Analysis of Antibiotic Resistance and Virulence Factors
- Drainage water management reduces nutrient, Escherichia coli, and antibiotic resistance gene export
BibTeX
@article{vouriot2026antibiotic,
title = {{Antibiotic resistance prediction with an attention-based bi-LSTM clinical decision support system}},
author = {Laurent Vouriot and Stanislas Rebaudet and Sabine Camiade and Melissa Lebsir and Jean Gaudart and Raquel Urena},
journal = {Computer Methods and Programs in Biomedicine},
year = {2026},
month = nov,
doi = {10.1016/j.cmpb.2026.109376},
url = {https://inserm.hal.science/inserm-05848870v1},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Antibiotic resistance prediction with an attention-based bi-LSTM clinical decision support system
AU - Laurent Vouriot
AU - Stanislas Rebaudet
AU - Sabine Camiade
AU - Melissa Lebsir
AU - Jean Gaudart
AU - Raquel Urena
PY - 2026
DA - 2026/11/02
JO - Computer Methods and Programs in Biomedicine
DO - 10.1016/j.cmpb.2026.109376
UR - https://inserm.hal.science/inserm-05848870v1
AB - Background and objective: Antimicrobial resistance is recognized by the World Health Organization as a significant global health threat. In clinical practice, the accurate identification of bacterial susceptibility to antibiotics is crucial. However, clinical laboratories often take several days to complete this process and, in the meantime, physicians rely on probabilistic and empirical reasoning, coupled with local hospital guidelines. Methods: In this work, we propose an attention-based bidirectional-Long Short-Term Memory recurrent neural network as a clinical decision support system to predict antibiotic resistance at the patient's bedside prior to the arrival of final antimicrobial testing results from the laboratory. More precisely, the model gives predictions at each stage of the bacterial identification process for a set of 47 single antibiotics and combinations, to support clinicians in their prescribing decision. Results: Great results were achieved, with a mean area under the receiver operating characteristic curve and a mean area under the precision-recall curve reaching up to 0.9. The attention mechanism was used to visualize the importance attributed to each feature and to better interpret the prediction results. Conclusion: The model has been integrated into a user-friendly and responsive web application, accessible on both mobile phones and desktops, to be used as a prototype clinical decision support system.
SN - 0169-2607
ER -