This sub-theme will focus on the data and machine learning aspect of our work.
- Developing an algorithm for predicting antimicrobial resistance (AMR) in patients with bloodstream infections. Looking at patients’ medical records, we will aim to predict whether they are going to present with infections that are resistant to drugs.
- Devising another prediction model that will indicate whether we should be reducing or stopping antimicrobial or antibiotic medications and switching to another other kind of medication.
- Machine learning model for predicting AMR in TB where genetic analysis cannot be used. In most circumstances we can use genetics to predict which strains are resistant to treatments, but this is not possible with some mutations.
- Automated system for predicting a patient’s stay in hospital, how long they are going to be in intensive care and when they might be expected to move to home treatment. This has far-reaching potential for the NHS.