Use of a large language model for surveillance of central line-associated bloodstream infections at a pediatric medical center

We assessed the accuracy of a large language model (LLM) for clinical decision support for central line-associated bloodstream infection (CLABSI) identification. Comparing LLM-assisted to manual review, the LLM could efficiently identify CLABSI and secondary BSIs with high sensitivity and specificity. Infection preventionists reported high satisfaction with the tool.