An AI tool and AMR score can help the NHS choose and manage antibiotics more effectively

Antimicrobial resistance (AMR) is a growing health challenge. It makes infections harder to treat, leads to longer hospital stays, and reduces the number of effective medicines.

Globally, AMR is linked to almost five million deaths a year. In UK hospitals, prescribing the wrong antibiotic, dose or duration is a major contributor to this. Clinicians often rely on complex guidelines that can be difficult to use in high-pressure settings. And healthcare systems lack a clear and consistent way to measure and track AMR.

In two studies led by Dr William Waldock and published in 2026, researchers at Imperial College London developed practical tools to address these challenges.

The first – the REMAPS AMR Burden Score – provides a structured way to measure and monitor AMR, combining data on resistance rates, clinical outcomes and prescribing practices.

The second, an AI decision-support tool called Ask Eolas, was evaluated in a structured simulation across 45 prescribing scenarios. Clinicians using Ask Eolas achieved 100 per cent prescribing accuracy, compared with 60 per cent for an existing clinical app and 47 per cent for standard guidelines. Clinicians using the tool reported reduced workload and were more confident in their decision-making.

Together, these tools provide a new approach to measuring and managing AMR, with the potential to make prescribing safer and more consistent, particularly in clinical settings where high volumes of antibiotics are used.

Both tools are being tested further in clinical settings to support wider NHS adoption and to reduce AMR. 

Read more about this research and its implications in our blog post.