Antimicrobial Resistance (AMR) is not just a health crisis; it’s a ticking time bomb that threatens global health security. In Low-and Middle-Income Countries (LMICs), the situation is particularly dire. Factors such as overuse of antibiotics, lack of awareness, and inadequate healthcare infrastructure contribute to a vicious cycle of non-compliance with Antimicrobial Stewardship Programs (ASPs).

In Brazil and similar regions, the complexity of these ASPs and the scarcity of frontline healthcare support further exacerbate the crisis. This research aims to break this cycle by leveraging cutting-edge AI technologies to make ASPs more accessible and effective. This project is designed in two distinct phases:

Phase One: The first phase focuses on transforming the existing antimicrobial guidelines of the Albert Einstein Hospital into a conversational agent. This agent will be capable of providing real-time, context-aware recommendations to physicians using RAG application architecture.

Phase Two: The second phase will involve using Large Language Models (LLMs) for information retrieval within Electronic Health Records (EHRs) and automating the evaluation of guideline adherence. This talk specifically covers the first phase of the project (methodology, architecture, validations and applied results).