Reducing unnecessary blood cultures in oncology through algorithmic stewardship in a tertiary medical center

Unnecessary blood cultures (BCx) increase healthcare costs, BCx contamination, and antibiotic overuse, particularly in cancer patients. Recent shortages in BCx bottles have amplified the need for more judicious BCx practices, especially in high-risk populations like oncology patients. This study evaluates the impact of a BCx algorithm on BCx rates and clinical outcomes in a solid-tumor oncology unit at Duke University Hospital.

We implemented a BCx algorithm in a solid-tumor oncology unit at Duke University Hospital. The primary objective was to compare BCx events in the pre- (2/2021-1/2023) and post-intervention (2/2023-7/2024) periods, which we defined as the collection of one or more BCx sets ordered by a clinician for a specific clinical indication per 100 inpatient days. Secondary outcomes were analyzed during the same periods and included length of stay, in-hospital mortality, 30-day readmission rates, and antibiotic days of therapy (DOT). BCx appropriateness was also assessed. An interrupted time-series analysis using Poisson regression assessed changes in BCx rates over time.

Following the implementation of the algorithm, there was a 38% reduction in the rate of BCx events per 100 patient days. Additionally, BCx appropriateness improved, and we observed no increase in adverse outcomes including DOT, in-hospital mortality, and 30-day readmission rates.

A BCx algorithm implementation in a solid-tumor oncology unit led to a substantial decrease in BCx event rates without an increase in adverse outcomes. This suggests that judicious BCx stewardship through algorithmic guidance is a viable strategy for reducing unnecessary testing in oncology patients.