Blood culture bottle shortage and diagnostic yield: A 5-year interrupted time-series analysis

Blood culture bottle shortages threaten stewardship and patient safety, yet data remain limited. We evaluated the consequences of the 2024 BD BACTEC™ blood culture bottle shortage in a laboratory network in São Paulo, Brazil. We performed a 5-year interrupted time-series study using aggregated blood culture (BC) data from January 1, 2020, through December 31, 2024. Prespecified periods were baseline (Period A; weeks 1-2020 to 32-2024), shortage (Period B; weeks 33-44/2024), and post-shortage (Period C; weeks 45-52/2024). Outcomes were BC event counts, contamination proportion and counts, positivity proportion, and clinically significant counts. Period comparisons used Mann-Whitney U tests. A seasonal ARIMA(0,1,1)(0,1,1)₅₂ model trained on Period A generated counterfactual forecasts. A total of 936,981 BCE events were analyzed, including 886,632 in Period A, 22,801 in Period B, and 27,548 in Period C. BC utilization declined from a median of 78.55 events/week in Period A to 48.02 in Period B and 48.87 in Period C (both P < 0.01 versus baseline). Contaminated-event counts declined, but contamination proportion was higher during the shortage and lower post-shortage compared with baseline (2.61%, 3.12%, and 1.96%; A versus B, P < 0.01; A versus C, P = 0.05). True-positive counts declined, whereas positivity proportion remained stable (10.88%, 11.22%, and 10.62%; A versus B, P = 0.12; A versus C, P = 0.07). Relative to the ARIMA counterfactual, BC utilization during the shortage was 25% lower than expected and remained depressed post-shortage. The shortage was associated with lower BC utilization, fewer contaminated events, and stable positivity proportion, but also fewer true-positive events, highlighting the trade-off between blood-culture conservation and diagnostic detection of bacteremia.