Data Challenge

Surveillance of antimicrobial resistance in UK hospitals relies on new methods that use data analytic tools. The researchers will make use of the Pfizer ATLAS data which contains some of the most detailed information on AMR and usage of antibiotics, to build sophisticated machine learning algorithms that can forecast both resistance trends and patterns in antibiotic use at the hospital and community level. Linking these datasets with other EHR data is meant to transform the way the UK monitors and acts on AMR.

Developed models must be able to track new AMR patterns and drug use as they happen, helping detect danger zones quickly and guide suitable actions by healthcare experts. These tools will look at patient information, existing health conditions, previous exposure to antibiotics and the patterns of antibiotic use in the hospital to highlight high-risk areas, catch inappropriate antibiotic use and advise on suitable new treatments—which will help patients by speeding up proper treatment and lowering the use of unnecessary antibiotics.

Having detailed antibiotic data and working with resistance data, the research will direct attention to where stewardship is necessary and generate useful alerts for highly used antibiotics. Intelligence from predictive insights will benefit public health by feeding targeted interventions to deal with multidrug resistance and showing how ML forecasts can support the UK Health Security Agency, using a One Health, integrated approach.

It will also look at the structural reasons for slow ML usage by streamlining data flows and including ethical safeguards, following advice from recent government assessments. Converting passive monitoring into useful, forward-looking knowledge, this work will boost the UK’s ability to handle AMR and create a workable model for the world.