Changes in observed bloodstream infection epidemiology during a single-set blood culture restriction: an interrupted time series analysis

During the 2024 US national blood culture bottle shortage, many hospitals implemented restrictions limiting blood culture collection to conserve supplies. The impact of these constraints on the microbiologic epidemiology of bloodstream infections (BSIs) remains incompletely characterized. We conducted an interrupted time series analysis of BSI episodes identified at two quaternary hospitals between June 2023 and June 2025, encompassing a systemwide restriction limiting clinicians to one blood culture set per patient every 24 h from June through December 2024. We grouped blood cultures into discrete infection episodes and categorized these by organism group. We normalized monthly incidence rates to acute-care patient-days and used segmented Poisson regression models to estimate changes in incidence associated with restriction onset and withdrawal. We identified 5,391 BSI episodes among 4,484 patients. During the restriction period, the relative proportion of episodes due to enterococci, fungi, gram-negative rods, and Staphylococcus aureus / lugdunensis increased, whereas episodes due to anaerobes, streptococci, coagulase-negative staphylococci/common commensals, and polymicrobial infections decreased. Interrupted time series analyses demonstrated an overall decline in detected BSI incidence, with statistically significant decreases observed for streptococci and gram-negative rods. Estimates for other groups were limited by small numbers and imprecision. Blood culture restrictions during a national supply shortage were associated with measurable shifts in the detected distribution of bloodstream pathogens and reduced overall BSI detection. These findings suggest that diagnostic constraints may selectively impair detection of certain organism groups, with important implications for laboratory surveillance, antimicrobial stewardship, and interpretation of microbiologic trends during periods of constrained testing.