New research by scientists at the University of Liverpool looks at how artificial intelligence (AI) can help doctors make better choices when prescribing antibiotics for urinary tract infection (UTI), one of the world's most common bacterial infections. By improving the precision of antibiotic prescribing, the research aims to tackle the growing global crisis of antimicrobial resistance (AMR).
In a new paper, this latest work illustrates a pioneering AI-based approach to treatment decision-making that combines human judgement and data. The new algorithm works by blending the experience of real doctors with data-driven predictions.
It uses a mathematical tool, called a utility function, to weigh up the pros and cons of each antibiotic option for an individual patient. By picking the most appropriate antibiotic for each patient, the system hopes to reduce unnecessary use of strong antibiotics and slow down resistance.
Dr Alexander Howard, from the Department of Pharmacology & Therapeutics at the University of Liverpool said: "Antimicrobial resistance (AMR) is one of the top global public health and development threats. It is estimated that bacterial AMR was directly responsible for 1.27 million global deaths in 2019 and contributed to 4.95 million deaths.
"In an era where antimicrobial resistance continues to increase, innovative solutions to facilitate precision use of antimicrobials are required - our utility-based system may present such a solution."
The simulation-study used real healthcare data and found the AI's recommendations were as good as those made by actual doctors but were less likely to lead to antibiotic resistance and more likely to suggest antibiotics that can be taken by mouth (rather than given directly into the bloodstream through a drip).
Importantly, the algorithm also has a built-in, innovative safety feature. If a patient is very ill, it automatically puts more emphasis on picking an effective antibiotic, making sure the treatment is effective when it matters most.
Dr Howard concluded: "Further research is now needed across a range of global settings to ensure the results apply more widely, especially in areas where antibiotic resistance has the biggest impact. However, this study shows that using AI alongside doctors' expertise could improve antibiotic prescribing, help fight resistance, and make treatments safer and more convenient for patients."
This latest work contributes to the University's globally-significant and innovative research in the fields of Therapeutics Innovation and Infection Resilience. These form part of a series of research frontiers that showcase the University's unique research strengths. The Therapeutics Innovation frontier positions the University of Liverpool at the forefront of efforts to revolutionise the drug discovery and medicine development pathway. Meanwhile our Infection Resilience frontier addresses urgent global and national healthcare challenges, delivering scientific breakthroughs and practical solutions to defeat infectious diseases and protect society.
The paper, 'Algorithmic antibiotic decision-making in urinary tract infection using prescriber-informed prediction of treatment utility' was published in npj Digital Medicine (DOI: 10.1038/s41746-026-02369-z).
