Answers are grounded in this paper and cite their sources.
Background Post-prescription review (PPR) by clinical microbiology/infectious diseases specialists is a proven intervention for optimising antibiotic management in hospitals. However, hospitals lack sufficient staff to conduct PPR at scale. This study investigated the feasibility and diagnostic performance of supervised machine learning to monitor electronic medical records and prioritise PPR in real-time. Methods Over two years, infection specialists categorised PPR recommendations as ‘stop’, ‘change’, or ‘continue’ therapy at an acute hospital in London, UK. These labels (n=2,625) were linked to features generated from electronic patient records. Random forest, XGBoost and C5 classifiers were trained before undergoing prospective evaluation in an unseen representative validation dataset (n=446 PPR decisions). The pre-specified minimum predictive performance target was an area under the curve (AUC) of 0.75, with statistical power to detect an AUC <0.68 or >0.82. We then aggregated the validation dataset at patient level (n=358) and compared the clinical utility of targeted PPR using the classifier against unprioritised (random) PPR. Results In the prospective validation set 145/358 (41%) patients were recommended to change/stop treatment. The best-fitting classifier (random forest, cross-validation AUC=0.74) achieved an AUC of 0.70 [95% confidence interval: 0.65; 0.75]. If just the top 30% of patients receiving antibiotics could be reviewed, the classifier would help stop/change treatment in 68/145 patients requiring a change, compared to 43/145 if patients were selected at random. Conclusions The prospective clinical evaluation demonstrated the approach is feasible, and achieves moderate predictive performance in real-world conditions.
No takes yet. Share an insight, caveat, or question.
Harris et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: