Impact of modifiable healthcare factors on outcome after bloodstream infection

There has been a range of research exploring on non-modifiable risk factors, such as patient comorbidities, for mortality in patients with a bloodstream infection (BSI). Factors of patient care which can be modified have been found to be associated with survival in other disease areas, but have not been explored in patients with a BSI to date. In this thesis, I explore the impact of modifiable factors on mortality in patients with a BSI using data from the BSI-FOO research programme (RP-PG-0707-10043). I begin by looking at healthcare setting related factors, before exploring the effect of duration of therapy. I then use trial emulation methods to emulate the MERINO trial, a recent randomised controlled trial comparing piperacillin-tazobactam and meropenem for the treatment of E coli or Klebsiella species BSI. Finally, I explore the association between the minimum inhibitory concentration (MIC) and mortality. Healthcare setting related risk factors associated with 28-day mortality were ward speciality, ward activity, ward movements, and time to receipt of appropriate antimicrobial therapy in the first seven days. In terms of duration of therapy, the hazard of all-cause mortality for short therapy vs long therapy was 1.74 (95% CI 1.36, 2.24) and for intermediate vs long therapy was 1.09 (95 % CI 0.98, 1.22). In the emulated trial, the odds for mortality was 1.31 times higher (95% CI 0.40 to 4.26) in patients in receipt of piperacillin-tazobactam compared to meropenem, after adjustment for propensity score. This was lower than the odds ratio observed in the MERINO trial, 3.7 (95% CI 1.5 to 10.4). Finally, there was no evidence to suggest a relationship between MIC/EUCAST breakpoint ratio and 28-day mortality in patients with a Gram-negative BSI. This thesis underlines the importance of appropriate antimicrobials within the first seven days, and the potential for ward activity, ward movements and duration of therapy to impact on survival in patients wit