Improving blood culture procurement: a prospective 5-year hospital-wide study.

BACKGROUND: Appropriate procurement of blood cultures (BC) is essential for diagnosis of bacteremia and susceptibility testing. This includes (1) adequate preparation of the venipuncture site to minimize contamination; (2) obtaining ≥ two sets with a time interval before starting antibiotics. Although these recommendations are standard since the 1960s, adherence is far less than expected - which may adversely impact on the management of bacteremic patients. AIMS OF STUDY: This single-center study conducted in Shaare Zedek Medical Center aimed to decrease the proportion of contaminated BCs and to increase the percentage of obtaining two sets of BC/ blood-culture-taking episode. METHODS: Determination of both markers at baseline, then monthly for one year, then subsequently on a quarterly basis; showing data from all departments in real-time to all department directors; and providing short educative lectures during departmental staff meetings, at baseline and after 1-2 years. These markers were adopted as one of the hospital-wide quality measures. RESULTS: In the 20-year period 2000-2019 more than 1 million BCs were obtained, of which 70% were from patients ≤ 72 h in hospital. During the 5-year study (2020-2024), the percent of blood-culture-taking episodes from which two culture sets were obtained increased annually by ± 16% from a baseline of 27% (9010/33306) in 2020, to 46% (18462/40191) in 2024 (Incidence Rate Ratios, IRR 1.16 [95%CI 1.13-1.18], p < 0.001). This improvement was observed in almost all departments and was especially profound in the emergency department (ED), starting at a baseline of 19% (1979/10326) and increasing to 53% (5304/9915)(IRR 1.33 [95%CI 1.27-1.39], p < 0.001). During the same period, the annual proportion of false-positive BCs, from which only contaminants were isolated, decreased annually by 18% from 2.4% (1592/65230) in 2020 to 1.3% (895/68991) in 2024 (IRR 0.82 [95%CI 0.77-0.88], p < 0.001). This improvement was observed in all departments: in the emergency department, this rate decreased from 3.3% (676/20529) to 1.56% (272/17459) (IRR 0.79 [95%CI 0.75-0.83], p < 0.001). CONCLUSION: A simple educational intervention, combined with meticulous data mining and presentation of each department's results, with comparison of all other departments, led to significant and sustained improvement in measurable markers.