When a patient is suspected of having a bacterial infection, antibiotics are often given and at the same time blood, urine and other samples might be collected to try to identify the bacteria causing the infection. Different antibiotics work against different bacteria, and test results, which typically arrive a few days after samples are taken, allow treatment to be adjusted to make sure that the antibiotics given work well. Antibiotics given before the test results are known are called empirical treatment. Usually, empirical antibiotic treatments are chosen to have a high chance of working against the most likely bacteria to be causing the infection, but this aim must be balanced against the risks of causing drug resistance by over-using certain antibiotics.
We will focus on understanding the relationship between different empirical antibiotic strategies and patient survival, in particular to find out what, if any, extra deaths are caused by delays in giving patients effective antibiotics. We will investigate what other things affect this relationship and how best to apply our methods across different patient populations.
The findings will help us understand the impact of different antibiotic policies on patient outcomes and in designing empirical antibiotic treatment guidelines in both the UK and other countries.