AMR Driver: Predictive AI models to help tackle antibiotic resistance in hospitals and communities

Lay Description

Antibiotics are a key part of treating infections. However, the use and overuse of antibiotics lead to bacteria developing antibiotic resistance, where antibiotics no longer work as expected. When antibiotics are given to patients with infections, doctors and other professionals try to balance giving treatments that have the best possible chance of working while also avoiding over-treating patients.

We aim to use artificial intelligence approaches together with the information that is routinely collected when patients come to hospital or see a GP to develop tools to predict the best antibiotic to start. We will compare the performance of the tools to decisions that hospital and primary care clinicians make. We will also compare the predictions made on which antibiotics to start to the results of laboratory tests looking for antibiotic resistance that are available later.

Public Benefit Statement

Ultimately, we are aiming to develop tools that will help care for patients and supporting giving them the best possible antibiotic treatment for infections. This could reduce deaths from antibiotic resistance, shorten the time patients spend in hospital, and decrease the need for repeat hospital admissions and attendances in primary care. Models could also assist healthcare workers in prescribing antibiotics and reduce the risks faced by healthcare organisations due to patients not receiving effective antibiotics.