University of Oxford

Predictive AI models to help tackle antibiotic resistance in hospitals and communities (Driver Project)

We will develop new artificial intelligence (AI) tools that help doctors choose the best antibiotics for patients with infections. This is possible because of a new NHS programme that safely brings together data from millions of hospital and community patients in one place, a Secure Data Environment.
Antimicrobial resistance (AMR) can stop antibiotics from working properly. This means patients can take longer to recover, and in serious infections may have a higher chance of dying. We want to use AI to predict which patients with infections are at risk of AMR. Our approach will rapidly scan detailed medical records from hospitals and GPs to help select the best antibiotics for each patient.
We will focus on two common infections: bloodstream infections, which can be life-threatening, and urinary tract infections (UTIs), which lead to millions of GP and pharmacy visits each year.
We will use medical data from hospitals and GPs across England, covering 18 million people. We will also co-design and build a digital platform so doctors can easily see the AI advice while seeing patients.
Our team aims to build AI models that work in real life, focus on patients most at risk, make sure the system is fair, and have working tools ready for real-world clinical testing to help doctors fight infections more safely and effectively.

The project aims to develop new approaches to use antibiotics more effectively, ensuring more patients receive active treatment while reducing future risks of antimicrobial resistance (AMR). Predictive AI models and a clinical platform could increase the number of patients with serious and common infections receiving appropriate initial antibiotic treatment, helping patients recover more quickly, reducing repeat GP consultations, shortening hospital stays, and reducing mortality.
The tools could also support wider use of narrow-spectrum antibiotics for patients at low risk of AMR, helping reduce the spread of resistance and supporting delivery of the National AMR Action Plan. Benefits for healthcare providers include reduced consultation rates, shorter inpatient admissions, lower antibiotic costs, and reduced infection prevention and control requirements.
The project will generate evidence to support implementation within the NHS and contribute to academic outputs through research publications and conference presentations. Public contributors will help oversee the research, while training opportunities will support early career researchers. The project will also strengthen collaborations between clinicians, healthcare data specialists, and AI experts.

15/09/2026

Live - Contracts Signed

SDE156

Infection

WMSDE trusted research environment

Yes

No

Thames Valley and Surrey Secure Data Environment

Not yet published.