Abstract
Overusing blood cultures (BCxs) can lead to false positives, unnecessary antibiotics and increased healthcare costs. Despite studies on inpatient BCx algorithms, none have focused on cardiothoracic surgery (CTS) patients, with complex postoperative care and invasive devices. This study aimed to evaluate the impact of a BCx algorithm on BCx event (BCE) rates in CTS step-down units. The study was conducted in three CTS step-down units at Duke University Hospital. The BCx algorithm, based on Seidelman et al. (2023), was implemented in June 2023. BCE rates, incidence rate ratios (IRRs) and adverse outcome IRRs were compared between pre- and post-intervention periods using ITS and χ2 tests. We analysed 4978 BCE during the study period: 3439 (893 patients) pre-intervention and 1539 (452 patients) post-intervention. BCE rates decreased [IRR = 0.78 (95% confidence interval (CI) 0.74, 0.83, P-value< 0.01)] without significant differences in adverse outcomes such as central line-associated bloodstream infection (CLABSI) rates (IRR = 0.6, 95% CI 0.17, 2.30), readmission rates (IRR = 0.99, 95% CI 0.88, 1.12) or in-hospital mortality (IRR = 3.53, 95% CI 0.32, 38.90). Our study supports the beneficial effects of a BCx algorithm, which reduces unnecessary BCxs in CTS patients without compromising patient safety.